Variational Inequalities and

Network Equilibrium Problems

Variational Inequalities and
Network Equilibrium Problems

Edited by

F. Giannessi
University of Pis a
Pisa, Italy

and

A. Maugeri
University of Catania
Catania, Italy

Springer Science+Business Media, LLC

 Library of Congress Cataloging-in-Publication Data
On file

Proceedings of the International School of Mathematics "G Stampacchia" 19th Course on Variational
Inequalities and Network Equilibrium Problems, held June 19—25, 1994, in Erice, Italy

ISBN 978-1-4899-1360-9 ISBN 978-1-4899-1358-6 (eBook)

DOI 10.1007/978-1-4899-1358-6

© Springer Science+Business Media New York 1995

Originally published by Plenum Press, New York in 1995
Softcover reprint of the hardcover 1st edition 1995

109 8 7 6 5 4 3 2 1

All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher
PREFACE

This volume brings forth a set of papers presented at the conference on "Varia-
tional Inequalities and network equilibrium problems", held in Erice at the "G. Stam-
pacchia" School of the "E. Majorana" Centre for Scientific Culture in the period 19~25
June 1994.
The meeting was conceived to contribute to the exchange between Variational
Analysis and equilibrium problems, especially those related to network design. Most
of the approaches and viewpoints of these fields are present in the volume, both as
concerns the theory and the applications of equilibrium problems to transportation,
computer and electric networks, to market behavior, and to bi~level programming.
Being convinced of the great importance of equilibrium problems as well as of their
complexity, the organizers hope that the merging of points of view coming from differ-
ent fields will stimulate theoretical research and applications.
In this context Variational and Quasi~Variational Inequalities have shown them-
selves to be very important models for equilibrium problems. As a consequence in the
last two decades they have received a lot of attention both as to mathematical inves-
tigation and applications. The proof that the above mentioned equilibrium problems
can be expressed, in terms of Variational or Quasi~Variational Inequalities also in the
non~standard and non~symmetric cases, has been a crucial improvement. Alternatives
and complements to this approach are offered by: the use of the normal map; the
belonging of zero to a multifunction; the use of gap functions; the use of models with
unilateral constraints; the separation theory; the duality theory. All these approaches
deserve attention and may lead to enlarging the class of equilibrium problems which we
can deal with in the applications. To be able to widen the applicability of the models
obviously has an important feedback in the theoretical investigation: Calculus of Vari-
ations, Stability Theory, Separation Theory, Complementarity, Duality, Generalized
Monotonicity, Normal Manifold Theory are instances of fields whose growth has been
strongly influenced by the applications. The editors hope this volume will contribute
to further exchanges between researchers in theory and those in applications.
We express our sincere thanks to all those who took part in the conference. Their
invaluable discussions have made this book possible. Special mention should once
more be made of the" E. Majorana" Centre, which offered its facilities and stimulating
environment for the meeting. We are all indebted to the Mathematical Committee and
to the group for Functional Analysis and Applications (GNAFA) of the Italian National
Research Council (CNR), and to the Italian Ministry for University (MURST) for their

v
VI

finantial support. We thank Giuliana Cai (Pisa) for the fine typing. Finallly, we want
to express our special thanks to PLENUM Publishing Co. for having offered to publish
this volume and for their constant encouragement and cooperation.

F. Giannessi (Pisa) A. Maugeri (Catania)

CONTENTS

1. On a separation approach to variational inequalities (C. Antoni)

1. Introduction ............................................................ 1

2. Separation .............................................................. 3

3. Gap functions .......................................................... 4

4. Connections between gap functions and duality for extremum problems. . . 5

5. Connections with complementarity problems ............................ 6

2. Traffic scheduling in telecommunication systems and network flow

(M. Bonuccelli)
1. Introduction 9

2. Basic problem .......................................................... 10

3. Variable bandwidth systems ............................................ 13

4. General switching modes and hierarchical systems ....................... 15

5. Conclusions ............................................................. 18

3. On the duality theory for finite dimensional variational inequalities

(M. Castellani and G. Mastroeni)

1. Introduction ............................................................ 21

2. Duality for variational inequalities ...................................... 22

3. The dual variational inequality of Mosco ................................ 24

4. Applications to classical Lagrangean duality ............................. 27

5. Connections with complementarity problems ............................ 30

4. Some properties of periodic solutions of linear control systems

via quasi-variational inequalities (P. Cubiotti)

1. Introduction ............................................................ 33

vii
Vlll

2. Preliminaries and notations ............................................. 34

3. The main result ........................................................ 35

4. An application to linear control systems ................................. 39

5. Generalized quasi-variational inequalities and traffic

equilibrium problem (M. De Luca)

1. Introduction ............................................................ 45

2. The direct method ...................................................... 46

3. An existence theorem for discontinuous multivalued costs. ....... ........ 51

6. Vector variational inequality and geometric vector optimization

(K.-H. Elster and R. Elster)

1. Introduction ............................................................ 55

2. Notations and preliminaries............................................. 56

3. Relationships between the vector variational inequality and the

geometric vector inequality ............................................. 61

4. Geometric vector optimization on the viewpoint of VVI ................. 63

7. Testing a new class of algorithms for nonlinear complementarity problems

(F. Facchinei and J. Soares)

1. Introduction ............................................................ 69

2. Basic definitions ........................................................ 70

3. Merit functions for NCP ................................................ 72

4. An algorithm for NCP .................................................. 74

5. Numerical results ....................................................... 78

8. Equilibrium in transport networks with capacity constraints (P. Ferrari)

1. Introduction ............................................................ 85

2. The concept of equilibrium ....... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 86

3. The equilibrium solution ................................................ 89

4. A modification of cost functions that assures the existence of equilibrium. 93

5. A computational example ............................................... 96

 ix

9. Separation of sets and gap functions for quasi-variational inequalities

(F. Gia.nnessi)

1. Introduction 101

2. Gap functions for quasi-variational inequalities. The weak case .......... 102

3. The strong case ......................................................... , 108

4. Finite dimensional VI with polyhedral domain .......................... III

5. Connections with duality ............................................... 115

6. Some extensions ........................................................ 118

10. Stability of monotone variational inequalities with various applications

(J. Gwinner)

1. Introduction 123

2. The variational inequality problem ...................................... 124

3. An abstract stability result ............................................. 127

4. An application to distributed market equilibria with bounds 131

5. An application to elliptic variational inequalities modelling

stationary contact problems ............................................ 134

11. A primal-dual proximal point algorithm for variational

inequality problems (K. Iwaoka, M. Fukushima and T. Ibaraki)

1. Introduction ............................................................ 143

2. Algorithm .............................................................. 145

3. Convergence............................................................ 147

4. An application to traffic assignment problems ........................... 151

5. Conclusion .............................................................. 152

12. Relations between t-, S-, z-domain descriptions of periodically-

switched networks (M. Koksal)

1. Introduction ............................................................. 155

2. Solution ................................................................. 158

 x

3. z-domain transfer function ............................................... 165

4. Conclusions .............................................................. 166

13. On side constrained models of traffic equilibria

(T. Larsson and M. Patriksson)
1. Introduction and motivation ............................................. 169

2. A side constrained assignment model ..................................... 172

3. Solving the side constrained model ....................................... 174

4. Conclusions and further research ......................................... 176

14. Advantages and drawbacks of variational inequalities formulations

(P. Marcotte)

1. Introduction ............................................................. 179

2. Notation and basic definitions ............................................ 180

3. Oligopoly models ........................................................ 181

4. The traffic assignment problem .......................................... 184

5. The generalized bilevel programming problem ............................ 187

6. A bicriterion problem .................................................... 190

15. Variational and quasi-variational inequalities in network flow models.

Recent developments in theory and algorithms (A. Maugeri)

1. Introduction ............................................................. 195

2. Models .................................................................. 196

3. Gap functions and equivalent differentiable optimization problems.

An empirical approach .................................................. 198

4. Gap functions. An image space formulation .............................. 201

5. On the convergence rate of projection methods .......................... 203

6. Sensitivity analysis. Paradoxes .......................................... 206

16. A variational inequality associated to a scalar conservation law

with unilateral constraints (M.K.V. Murthy)

1. Introduction ............................................................ 213

 xi

2. The Cauchy problem for first order quasi linear equations

on the whole space ....................................................... 216

3. The Cauchy-Dirichlet problem for first order quasi linear

equations on a bounded cylindrical domain and viscosity solutions ......... 219

4. First order quasi linear equations on a bounded cylindrical

domain with unilateral constraints ........................................ 221

5. Idea of the proof .......................................... . . . . . . . . . . . . . . .. 222

17. Continuity of the solution set of homogeneous equilibrium problems

and linear complementarity problems (W. Oettli and N.D. Yen)

1. Introduction and preliminaries ............................................ 225

2. Homogeneous equilibrium and complementarity problems ................... 226

3. Linear complementarity problems ......................................... 231

8. Tensor approximations of smooth nonlinear complementarity systems

(T. Rapcsak)

1. Introduction ............................................................... 235

2. Smooth nonlinear complementarity systems ................................. 236

3. Tensor field complementarity systems ...................................... 237

4. Tensor approximations of smooth functions on Riemannian manifolds ....... 239

5. An optimization approach ................................................. 244

6. Concluding remarks ......................................................... 247

19. Basic existence theorems for generalized v<u'iational and quasi-

variational inequalities (B. Ricceri) ........................................... 251

20. Sensitivity analysis of variational inequalities by normal-map

techniques (S. Robinson)

1. Introduction .............................................................. 257

2. Variational inequalities and normal maps .................................. 258

3. The normal manifold ..................................................... 260

4. Implicit functions and sensitivity analysis ................................. 264

xii

5. Conclusion ............................................................. 268

21. Monotone relations and network equilibrium (R.T. Rockafellar)

1. Introduction ............................................................ 271

2. Variational inequalities and optimization ................................ 272

3. Network equilibrium .................................................... 275

4. Equilibrium as a variational inequality .................................. 278

5. Optimal flows and potentials ............................................ 281

6. Equilibrium as finding a zero of a mapping .............................. 285

22. Generalized monotonicity-concepts and uses (S. Schaible)

1. Introduction ............................................................ 289

2. Concepts of generalized monotonicity ................................... 290

3. Generalized monotone variational inequality problems ................... 294

Contributors .............................................................. 301

Index ..................................................................... 303

ON A SEPARATION APPROACH TO VARIATIONAL INEQUALITIES

Antoni Carla

Istituto Nazionale di Alta Matematica " F. Severi",

Citta Universitaria, Roma, Italy
and
Accademia Navale, Viale Italia, Livorno, Italy
e-mail antoni@bianchi.dm.unipi.it

Abstract: A class of Variational Inequalities in a Hilbert space is considered and the

existence of a solution is expressed as the impossibility of a system. Then, separation
arguments are applied to obtain results about gap functions and duality.

Key words: Variational Inequality, separation of sets, Lagrangean function, gap func-
tion, duality.

1. INTRODUCTION

A class of Variational Inequalities in a real Hilbert space is considered. The

Variational Inequality is reduced to the impossibility of a suitable system, and then
it is associated with a general separation scheme. This allows to analyze, in general
terms, some topics which have received a lot of attention recently. In this paper a
general class of gap functions is derived from the separation scheme and a duality
scheme is defined; however it is concevable to exploit the same approach to investigate
other topics, such as, for instance, existence, uniqueness, saddle point, penalty, solving
algoritms. Here, the possibility of investigating Lagrangean functions associated with
Variational Inequality, gap functions and Dual Variational Inequalities are discussed.
The present separation scheme contains, as a particular case, that introduced in [3]
and hence the class of the gap functions considered here contains that of [3]. The same
scheme allows to recover the Lagrangean functions exploited in [2] to construct Dual
Variational Inequalities, in particular that introduced by Mosco [5].

Let 2 be a real Hilbert space, F : 2 -+ 2, f :2 -+ R U {+oo}. Consider the

following Variational Inequality (in short VI): find y E Dom f such that

(F(y), x - y) "2. f(y) - f(x), 'v'x E 2,

1
2

where Dom f is in the sense of [6J.

Let y E Dom f be fixed; consider the function .p(y; . ) : 3 --. R U {-oo}, with

.p(y;x):= (F(y),y - x) + fey) - f(x).

Then, obviously, y is a solution of VI if and only if the system (in the
unknown x):

.p(y; x) > 0, x E3
is impossible.
Now, another system is introduced to characterize a solution y of VI. More pre-
cisely, fix a vector space V and a function cp(y; ., . ) : 3 x V --. RU {-oo}, and assume
the function cp(y; ., . ) satisfies the following condition:

cp(y;x,O) = .p(y;x), Vx E 3. (Ll)

Finally, let

V:= {v E V: cp(y;x,v) = .p(y;x), Vx E Domcp(y; .,v)},

and let Vo be any subset of V, such that

OEVo~V. (1.2)

Lemma 1.1 Let y E Domf be fixed. The system (Sy) is impossible if and only if the
following system (in the unknowns x, v);

cp(y;x,v) > 0, x E 3, v E Va
is impossible.

Therefore, we are interested in analyzing the disjunction of the sets

qy):= ((u,v) E R x V: u = cp(y;x,v), x E 3},

1i:={(u,v)ERxV:u>O, vEVo}.

In fact, in terms of separation arguments, Lemma 1.1 becomes the following proposi-
tion, whose obvious proof shows Lemma 1.1 too.

Lemma 1.2 Let V be a vecto? space, cp(y; ., . ) : 3 X V --. R U {-oo} a function such
that (1.1) holds, and let Vo be a set which fulfils (1.2). Then y E Domf is a solution
of VI if and only if
qy)n1i=0. (1.3)

Proof. Assume qy) n 1i = 0. Then cp(y;x,O) = .p(y;x) :::; 0, Vx E 3.

Therefore y is a solution for VI. Vice versa, suppose that there exists
(il, v) E K(y) n 1i, that is there exists x E 3 such that cp(y; x, v) > 0. Thus y cannot
be a solution of VI. o
 3

2. SEPARATION

The disjunction of K(y) and 1i is studied by means of their inclusion in level sets
of suitable functionals defined on R xV, depending on a parameter w which belongs
to a fixed class fl (wallows to describe the class of the functionals w(·, . ;w) ). The
class fulfils the following conditions:

VwEfl,
i) w( . , '; w) : R x V -+ R, w( u, V; w) = u + "Y( V; w)
ii) lev~o w(·, . ;w) ;21i.

Remark 2.1. When Vo = R+ the class of functionals considered above becomes that
introduced in [4J.

Remark 2.2. If the type of inclusion in ii) is reversed, a different kind of separation is
obtained. We call it " strong separation". By symmetry we call " weak " that defined
by i), ii); it is conceivable and interesting to investigate the strong case too.

Remark 2.3. Condition ii) is equivalent to the following one

From ii) it follows that (1.3) is true if there exists w E fl such that

w(u,v;w) ~ 0, V(u,v) E K(y). (2.1)

In order to find such an w E fl, the following weaker relation is studied:

inf sup w(u,v;w) ~ 0,

",EO (u,v)EK:(y)

that is
inf sup [cp(y; x, v) + "Y(v;w)] ~ O. (2.2)
",EO (z,v)EB><V

Definition 2.1. Let y E Dom! be fixed. The function

A(y; ., . ) : =: x n -+ R U {+oo, -oo}

defined by
A(y;x,w):= inf [-cp(y;x,v)-"y(v;w)],
vEV

will be called the Lagrangean function associated with (P Sy) .

Definition 2.2. Let y E Dom! be fixed. The problem

sup inf A(y; x,w),

"'EO xE:
4

will be called dual problem usociated with (PSy).

Proposition 2.1. Let Y E Domf be fixed. If there exists w E 0 which fulfils condition
(2.1), then (y,w), is a saddle point for for the Lagrangean A(Yi ., .) on B x n .

Proof. The assumption implies that

A(Yix,w) ~ 0, "Ix E B. (2.3)

Moreover, since It'(YiY,O) = <fJ(YiY) = 0, and ')'(OiW) ~ °'Vw E 0, we have

A(YiY,W) ~ -1t'(YiY,O) -')'(Oi W) ~ 0, Vw EO. (2.4)

From (2.3) and (2.4) it follows that

A(YiY,W) ~ A(YiX,W), "Ix E B. (2.5)

Because of (2.3) and (2.4) it holds

A(YiY,W) = 0i
hence, we achieve the inequalities

A(YiY,W) ~ A(YiY,W), Vw E O.

These, together with (2.5), complete the proof. 0

Remark 2.4. Let V = B and let It'(Yi ., .) : B x ::: -+ R U { -oo} be the function
defined by
It'(Yi x,v) = (F(y),y - x) + f(y) - f(x + v).
Moreover set Va = {OJ, 0 = B, ,),(ViW) = -(v,w). Then, the Lagrangean function
A(Yi ., . ) coincides with that considered in [2] to obtain the Dual Variational Inequality
proposed by Mosco [5].

Remark 2.5. The present approach can be applied to any problem which may be
reduced to the impossibility of a system like (Sy). This happens, in particular, for
extremum problems: using this approach it is possible to derive Lagrange and Fenchel
duality [1].

3. GAP FUNCTIONS

Recently, a new method has been introduced to solve a Variational

Inequality: a non-negative function g, such that g(y) = 0 if and only if Y is a so-
lution for VI is constructedi therefore, we look for a minimun of g on the feasible set
of VI. Such a g is known as gap function.
 5

Using the previous approach we are able to consider a general class of gap func-
tions.

Proposition 3.1 The function g: Domf -+ Ru {+oo},

g(y) := inf sup [ - A(y;x,w)] , (3.1)

",EO xES

is non-negative. Moreover, if g(y) = 0, then y solves VI.

Proof. The definition of A(y; " . ) implies that

sup [-A(y;x,w)] ~-A(y;y,w)~O, 'IwEn,

xES

therefore 9 is non-negative. Let g(y) = 0. Ab absurdo, assume y is not a solution for

VI; then there exists x E :=: such that

'f'(y;x,O) > 0;

then
sup [- A(y;x,w)] ~ 'f'(y;x,O) +,(O;w) ~ 'f'(y;x,O) > 0.
xES

Hence g(y) > 0. o

Remark 3.1. The existence of an wEn, such that lev~o w(·, . ;w) ;2 K(y), is a
sufficient condition to prove that

y solves VI => g(y) = 0.

Then, to prove the function 9 defined in (3.1) is a gap function, we need some assump-
tions about K(y), in particular about the function 'f'(y; " . ).

Remark 3.2. The function 9 defined in (3.1) collapses in those introduced in [3] when
the functions 'f'(y; . , . ) and ,( . ; w) are specialized.

4. CONNECTIONS BETWEEN GAP FUNCTIONS AND DUALITY

FOR EXTREMUM PROBLEMS

Consider the following Variational Inequality: find y E K such that

(F(y),x - y) ~ 0, 'Ix E K,

where K := {x E R n : Bx ~ b}, B is an m x n real matrix and bERm. Assume

there exists a convex function h such that F = Vh, that is VI has the integrability
6
property. We will show that the previous scheme gives a gap function 9 such that g(y)
is the difference between the primal value h(y) and the "Wolfe dual value" given by

sup [h(y)-{A,By-b)] .
• e1R~.
Vh(.)=.B

Proposition 4.1 The function g: K -+ R,

g(y):= infm [(F(y),y)-{>.,b)]. (4.1)

.e1R+.
F(.)=.B

is a gap function for VI.

Proof. g is obtained from (3.1) setting V = Rm,

={
(F(y),y-x),ifBX~b+V,
<P ( yjX,V ) .
-00, otherwise,

choosing Vo = R~, n = R~ and -Y(Vjw) = (w,v). Then, recalling Proposition 3.1, it

only needs to prove that y solution of VI implies g(y) = O. Indeed, y is a solution if
and only if K(y) n 11. = 0. Since K(y) and 11. are convex sets, there exists an w E R~
such that
sup [<p(YjX,V) + (w,v)] $0.
(z,v)ER x1Rm R

Remark 4.1. Let y E K be fixed. The value g(y) of the gap function g defined in
(4.1), can be written as follows:

g(y)=h(y)- sup [h(y)-{>.,By-b)] ,

'2: 0 •
Vh(.)-.B=O

where the supremum is with respect to >.. This expression for 9 suggests to use Wolfe's
duality theory for extremum problems to find again 9 is a gap function.

Remark 4.2. The function g is also considered in [3].

5. CONNECTIONS WITH COMPLEMENTARITY PROBLEMS

Consider the following Complementarity Problem (in short, CP): find y E C such
that
(F(y), y) = 0, F(y) E C* ,

where C ~ R n is a closed convex cone, and C· its (positive) polar cone.

 7

In the literature, among the gap functions associated with CP, there is the follow-
ing g : C -+ R with
g(y) = (F(y), y) .
The problem
(P) inf{g(y), yEC, F(y)EC*},
turns out to be equivalent to CPo

The function g and the problem (P) can be derived from the previous scheme. In
fact, CP coincides with the following VI: find y E C , such that
(F(y),x-y)~O, VxEC.

Remark 5.1. This VI is obtained from the one introduced in Sect. 1 setting f = oe,
the indicator function of C.

Vy E C, let 'f'(y; ., . ) : R n x R n -+ R U {-oo} be the function defined by

(F(y),y-x), ifxEC+v,
'f' ( y;x,v ) ={
-00 , otherwise.
Moreover let Vo be the cone C (such a Vo satisfies (1.2)); finally set n = C* and
,(v;w) = (v,w) (such a , satisfies ii)). In this case, the Lagrangean function
A(y; . , . ) : R n x C* -+ R, is given by
A(y;x,w) = -(F(y),y) + (x,F(y) -w).
and then the function 9 defined in (3.1) turns out to be the following
g(y) = {(F(Y),y), if F(y) E C* ,
+00, otherwise.
The function 9 coincides with the gap function g on Dom gj then the infimum of 9 is
found:
inf {g(y), y E C, F(y) E C*}.
This is precisely the problem (P).

REFERENCES

[1] C. Antoni, "Dual problems and separation of sets. Lagrange and Fenchel duality". To appear.
[2] M. Castellani and G. Mastroeni, "On the duality theory for finite Variational Inequalities", this
volume.
[3] F. Giannessi, "Separation of sets and Gap Functions for Quasi-Variational Inequalities", this
volume.
[4] F. Giannessi, "Theorems of the alternative and optimality conditions". lou. Optim. Theory
Appl., 42 N. 3, pp. 331-365, 1984.
[5] U. Mosco, "Dual Variational Inequalities", Journal of Mathematical Analysis and Applications,
40, 202-206, 1972.
[6] R.T. Rockafellar, "Convex Analysis", Princeton University Press, Princeton, 1970.
TRAFFIC SCHEDULING IN TELECOMMUNICATION SYSTEMS AND
NETWORK FLOW

Bonuccelli Maurizio A.

Dipartimento di Scienze dell'Informazione

Universita' di Roma "La Sapienza"
Via Salaria, 113, 00198 Roma, Italy
e-mail: algor@vm.cnuce.cnr.it

Abstract. We present several problems related to the traffic scheduling in computer

networks and telecommunication systems, and show how network flow can be used to
efficiently solve these problems. The problems consist in minimizing the time needed
to transmit a set of information units (packets) from the sources to their destinations,
in systems made of entities communicating by means of the so-called "interconnection
networks". The aim is to schedule the packets so to avoid communication conflicts
(packet collisions), and minimize the schedule length. Several side-constraints arise
according to the technical features of the specific systems under consideration, and
accordingly several different problems arise. Some of these problems can be represented
(and solved) by a sequence of max flows in proper bipartite networks. Others are
computationally intractable (i.e. NP-complete).

Key Words: Telecommunication traffic scheduling, max flow, circulation, polyno-

mial time algorithms, NP-completeness

1. INTRODUCTION

An interconnection network is a physical device capable of connecting a set of inputs

with a set of outputs in a variety of ways, and is used to convey messages from the
inputs to the outputs. Specifically, an interconnection network with n inputs and n
outputs can connect each input with an output, according to the values of some control
signals. In order to assure a correct reception of the messages, each input cannot be
simultaneously connected with more than one output, and each output can receive at
most one message at a time. This constraint is due to the nature of interconnection
networks, which are made of electronic devices and wires. If two or more signals are
simultaneously put into a wire, they are mixed, and it is impossible to distinguish one
to the other at the receiver. Interconnection networks can be classified according to
the number of different input-output connections they can support. An input-output
connection can be represented by a permutation of the inputs or of the outputs.

9
10

Among the several classes of interconnection networks, the so-called permutation

networks, namely interconnection networks capable of supporting any input permuta-
tion, are the most powerful and most important ones. In this paper, we shall consider
permutation networks only. They are the core of several communication and processing
systems, like telephone systems, the so called SS/TDMA communication satellites, and
distributed memory parallel processing systems. Examples of permutation networks
are the crossbar switch, Clos network, Omega network, Benes network, and so on.
All these networks share the above input-output contention problem. Several pro-
tocols have been proposed in order to overcome such a problem. One of the most pop-
ular and effective is the time division multiple access (TDMA for short). This protocol
consists in setting up a collision-free schedule of the messages, and then transmitting
according to such a schedule. A schedule is called frame, and is composed of several
sub-frames, named time slots. Time slots are the time units of the frames, and their
length is equal to the time needed to transmit a packet. Thus, we assume packetized
communication. This technique is emerging as the most suitable one for transmitting
not only data, but also voice and video [10,20,21].
Under TDMA protocol, the system operates in the following way: first, each input
issues output connection requests. Then, a schedule (also called time slot assignment,
or TSA for short) of the above requested communications is produced and broadcast to
all the system users (inputs and outputs). Finally, transmission takes place, according
to the schedule. When the frame is over, the above procedure is repeated. Thus, cyclic
transmission in a synchronous system in performed.
There are many ways of sc4eduling a given requests set, each with different features.
A very important property of such schedules is their time length. Shorter schedules
better utilize the system, since they allow the starting of new frames earlier, thus
improving system performance. In the following, we shall look for minimum length
schedules, namely those that maximize the performance. According to the system
under investigation, several side constraints to the scheduling problem may arise, and
so several different minimum length TSA problems originate [2-11,13-21].
This paper is organized as. follows: in the next section, the basic TSA problem is
introduced, and a fast, max flow based algorithm for optimally solve it is presented.
In Sect. 3, the results of Sect. 2 are generalized, in order to solve the TSA problem for
variable bandwidth systems. Sect. 4 deals with other TSA problems, solved by means
of circulation theory. Finally, conclusions and open problems terminate the paper.

2. BASIC PROBLEM

In the following, we assume that traffic in an n input, n output interconnection

network must be scheduled in a minimum length frame. The i-th input is denoted by
Ii and the j-th output by OJ. The transmission requests to be scheduled are represented
by an nxn nonnegative integer matrix D, with entry dj,j representing the number of
packets that Ii wants to send to OJ (and so the number of time slots required by such a
transmission). Thus, the i-th row of D represents the number of packets that Ii wants
to transmit, and the j-th column the number of packets bound for OJ. The physical
constraint described in the introduction imposes that at most one packet from Ii, for
each i, 1 :5 i :5 n, and at most one packet to OJ, for each j, 1 :5 j :5 n, is scheduled in
each time slot.
The packet scheduling of a time slot can be represented by a matrix, called switching
matrix, and denoted by S. Switching matrices are matrices with 0-1 entries, and the
above physical constraint force them to have at most one entry equal to 1 in each
 11

line (row or column). Observe that when S has exactly one entry equal to 1 in each
line, it is a permutation matrix . Given this representation, a TSA can be seen as
a decomposition of D into switching matrices. So, we have the following minimum
length scheduling problem (also called ML TSA):
Given a nonnegative integer (traffic) matrix D, decompose it into the smallest
number of switching matrices, namely find 51, 52, ... , 5 L such that D = 2:r=o 5 i , and
L is minimum.
We shall first recall a classical linear algebra result. A doubly stochastic matrix is
a nonnegative matrix in which each line (i.e. row or column) has sum equal to 1.
The classical Birkhof f - vonN eumann theorem states that any doubly stochastic
matrix is a convex combination of permutation matrices (or, equivalently, the doubly
stochastic matrices are the convex envelope of the set of permutation matrices).
Birkhoff-von Neumann theorem could be the theoretical basis for our TSA prob-
lems. Unfortunately, it deals with non-integer matrices, while we are interested in
matrices with integer entries only. Thus, we need to generalize in some way that theo-
rem. Let us define pseudo doubly stochastic matrices as nonnegative integer matrices
with identical line sums. Notice that a doubly stochastic matrix can always be ob-
tained by a proper pseudo doubly stochastic one by dividing each entry by the line
sum (and viceversa, by multiplying each entry by the inverse of the smallest nonzero
entry).
Let us go back to our TSA problem. A lower bound L on the length of any schedule
for it can be readily obtained as follows:

L = max { 2:7=1 di,i, for each i;2:i=1 di,i, for each j }

A critical line (critical entry) is a line (entry) with sum equal to L.

It is often useful to represent a matrix A as a bipartite graph. In order to do so,
we associate the rows of A to a set U of nodes, and the columns to another set V.
Then, we put ai,j parallel edges joining Uj E U with Vj E V whenever entry aj,j 2: O.
In this setting, a switching matrix corresponds to a matching (a set of edges no two of
which are incident in the same node) in the associated bipartite graph. So, a TSA is a
partition of the graph edges into matchings, namely an edge coloring, and a minimum
length TSA corresponds to a minimum edge coloring.
We shall prove the next theorem by means of network flow. Notice that the theorem
is a restatement of the Birkhoff-von Neumann theorem. We prove it here to introduce
a proof technique that will be used extensively in the rest of the paper.

Theorem 1. Let us assume that traffic matrix D is pseudo doubly stochastic. Then,
we can extract a permutation matrix from it, or equivalently the associated bipartite
graph has a perfect matching (each node is incident in one edge of the matching).

Proof. Let us consider the bipartite graph associated to D, with one edge joining
Ui E U with Vi E V whenever entry di,i 2: 0, instead of di,i parallel edges, and let us
assign a capacity equal to di,j to such an edge. Then, we add a source node s, and an
edge between s and each node Ui E U, as well as a sink node t, and an edge between
each node Vi E V and t. All these edges have capacity 1. It is easy to see that a
feasible flow in the above network corresponds to a matching (just pick up the (Ui, Vj)
edges with non zero flow). That matching is perfect if and only if all the nodes are
involved in positive flows. This is true if the cut separating s from all the other nodes
12

in the network, is a minimum one.

Let (0', r) be a generic cut, with capacity C(O', r). Then,
C(O', r) = I{R; E r}1 + LR;EO' LCJEr dj,j + I{Cj E O'}I=
=Pr+ LR;EO' LCJEr dj,j+ 'YO'=
=Pr+LR;EO' dj,j+ LR;EO' LCjEO' di,j+ 'YO'=
=Pr+I{Rj E O'}I L -LR;EO' LCjEO' dj,j+ 'YO'=
=N + I{R; E O'}I (L -1)+'Y0'- LR;EO' LCjEO' dj,j
There are two cases: 'YO' ~ I{Ri E O'}I, and 'YO' < I{Rj E O'}I. In the former case, we
have that
N + I{R; E O'}I (L -1)+'Y0'- LR;EO' LCiEO' di,j ~ N+I{Ri E O'}I L-
LR;EO' LCiEO' di,j ~ N
while in the latter, we have that
N + I{R; E O'}I (L -1)+'Y0'- LR;EO' LCiEO' di,j ~ N+'YO' (L -1) + 'YO' -
LR;EO' LCiEO' di,j ~ N .
In both cases, this generic cut cannot have a capacity smaller than the cut sepa-
rating s from all the other nodes in the network, and the theorem is proved.
Q.E.D.

The above theorem can be used to build an optimal algorithm for our basic TSA
problem. In fact, if D is pseudo doubly stochastic, and S is a permutation matrix
extracted from it, then D'=D-S is again pseudo doubly stochastic. Besides, if D had
a lower bound equal to L, then D' has a lower bound L'=L-l. If we repeat the above
procedure untill we are left with a zero matrix, we have a decomposition of D into L
permutation matrices, an optimal TSA.

Optimal Basic Algorithm

step 1 Set up the Bipartite Graph B associated to D;
step 2 Find a Perfect Matching, P, of B;
step 3 Derive the Permutation Matrix from P;
step 4 Subtract P from B;
step 5 If the edge set of B is empty, then halt; else goto step 2.

The time complexity of the above optimal algorithm is O(L*MA), where MA is

the time complexity of the chosen matching algorithm (if the algorithm presented
in [1] is applied, then MA is O(n 2 .5 )), since a perfect matching is found L times.
Unfortunately, the above complexity is not a polynomial one, but pseudopolynomial,
and so the algorithm can be very slow, when the numbers in the input are large.
We can dramatically lower the above complexity by observing that the same perfect
matching can be found several times by the optimal algorithm. Is it possible to find it
once, and then put it in the schedule the proper number of times? How large is such
a number? Let us call that number the multiplicity factor of the matching, and let
us denote it by Jl. It should be the largest number allowing the remaining matrix to
be pseudo doubly stochastic again, namely
/-t = min {dj,j such that edge (i,j) is in P}
Then, we replace step 4 with the following:
step 4' Subtract /-t P from B;
Notice that, now, each subtraction reduces the lower bound by /-t (and not one),
and eliminates one edge from B (i.e., we have one less positive entry in D). Since we
have at most O(n 2 ) edges in B, we must find at most O(n 2 ) matchings, and the overall
 13

time complexity now is O(n 4 .S ), which is fully polynomial, independent of the size of
the input numbers [4].
The traffic matrix is rarely in pseudo doubly stochastic shape. What can we do
when the traffic matrix has a generic shape, and is not pseudo doubly stochastic?
Fortunately, it is possible to add some dummy traffic to the entries of the matrix
so that it is transformed into a pseudo doubly stochastic one. The resulting matrix
has the same lower bound on the schedule length of the original matrix, and the
transformation algorithm takes only O(n 2 ) time [4}. Essentially, the algorithm (called
filling) is as follows:

Filling Algorithm
for i:=l to N do
for j:=l to N do
di,j f- di,j + L - max { E7=1 di,j; E:'=1 di,j }

Obviously, the dummy traffic must be deleted from the schedule as soon as it is
produced.
When the traffic matrix D has a special structure, faster algorithms that take into
account such structure, can be used.
As we shall see, the above approach can be used to solve in polynomial time other,
more complex problems. For instance, the above algorithm can be used to optimally
solve the TSA problem in which the traffic matrix is not a square one, but has a num-
ber of rows different from that of the columns, and when there is a further constraint
on the total number of nonzero entries that can be present in the switching matrices.
This constraint takes into account the presence of the so-called transponders (fre-
quency shifters) on board of the satellite, thus limiting the number of simultaneous
communications [5J.

3. VARIABLE BANDWIDTH SYSTEMS

In this section, we shall consider satellite systems with unequal uplink/downlink

(namely, satellite input and output) radio frequency beams. More specifically, input i
is the frequency multiplex of at most ai packets, while output j is the multiplex of at
most f3j packets. These systems have been devised to allow earth stations with higher
traffic to have more bandwidth for transmitting/receving than the lower traffic ones
[2,9,13J.
The above arrangement implies a physical constraint diferent from the one of the
basic system. Specifically, in this case at most ai simultaneous signals can be emitted
from input i, and at most f3j simultaneous signals can be directed to output j. Notice
that from a mathematical point of view, this is a generalization of the basic problem
constraint, where ai and f3j are always equal to one.
A further constraint can be added, namely that a total of at most K signals can be
handled by the system at any given time. This constraint is meaningful in a satellite
setting, and takes into account of the number of transponders (frequency shifters) on
board of the satellite.
It is natural to extend the definition of switching matrix in order to adapt to this
case. Thus, we define af3-switching matrix as a matrix S such that:

E7=1 Si,j ::; ai, for each i;

14

Ei=t Si,j ~ (3;, for each jj

Ei=t Ei=t Si,j ~ Kj (*)
Si,j 2: 0 and integer.

Observe that the same constraints, with ai=1 and (3j=l, hold for the basic problem.
An a(3 -switching matrix S is called complete whenever the above inequalities are tight.
A traffic matrix D is called normalized pseudo doubly stochastic (NPDS for short) if
there exists some integer number L such that:

Ei=t di,j = ai L, for each ij

Ei=t di,j = (3j L, for each jj
A time slot assignment in variable bandwidth systems is a decomposition of a traffic
matrix D into a(3 -switching matrices.
An obvious lower bound, similar to the one derived for the basic problem, holds in
this case too. Specifically, let L denote such a bound. Then,

L = max { r -k Ei=t di,; 1, for each i;f t Ei=t di,j 1, for each j j

r k2:;=1 2:7=t di,; 1 }

If D is NPDS, then it is possible to add an extra row, an extra column, and extra
entries such that the resulting matrix is NPDS with the same lower bound of the
original matrix, and the constraint (*) can be neglected. A polynomial time algorithm
(called filling algorithm) for the above transformation has been presented in [13].
Thus, we can assume that D is NPDS, and we can neglect the constraint (*). Notice
that, unlike the basic problem, in this case an a(3 -switching matrix is not a matching,
and thus matching algorithms cannot be used to derive a TSA. Fortunately, it is
possible to use here the network flow approach already proposed for the basic problem
[2,9].
Traffic matrix D always contains a complete a(3 -switching matrix (provided it is
a NPDS matrix). A proof of this statement can be readily derived from the proof of
Theorem 1. In fact, we can use the same network representation of the bipartite graph
associated to D, with the only change in the capacity of the (s, Ui) and (v;, t) edges,
which now become ai and (3j, respectively. Again, the above statement is true if and
only if the cut separating s from the other nodes is a minimum capacity one. In order
to prove this, we assume (with no loss of generality) that 2:;=1 ai = 2:7=1 (3j .
Let (0', T) be a generic cut, with capacity C( 0', T). Then,

C(O', T) = 2:RiET aj + 2:R,EO' 2:C,ET dj ,;+ 2:C,EO' (3; 2: 2:']=1 (3;=

=2:C,EU (3; + 2:C,ET (3; .

Let us multiply both sides of the inequality by L. Then, we get

Thus,
 15

and so,

2:RiE'T 2:C ECT

J d i,; + (L-l) 2: R ,ECT 2:C E'T
J d i ,; ~ 0,

which is always true, since L ~ 1, and di ,; ~ 0, by assumption.

Following the approach used in the previous section, the above proposition together
with the filling algorithm can be used to provide an optimal algorithm for the TSA
problem in variable bandwidth systems. Indeed, the same scheme used in the previous
section, with the obvious change in step 2 (find a maximum flow instead of a perfect
matching), suffices. Again, such an optimal algorithm is pseudo-polynomial time, and
again there exists a set of multiplicity factors whose use in the algorithm can make
polynomial its time complexity. However, this time the multiplicity factors are more
tricky to find, and it is not at all obvious that their use lead to polynomial time
complexity [2].
Let F= { Ii,; : i f= s, j f= t and A; greater than 0 } be the set of positive maximum
flows in the (i,j) edges. Then, the multiplicity factor for F is:
Jl = min { Ld i,; -;- Ii,; J : Ii,; E F }
The time complexity of this algorithm, with the above multiplicity factors, is 0(1)
(n2+~)), where 1> is the time complexity of the best maximum flow algorithm (which
depends on the special structure that the bipartite graph has, and can change from
matrix to matrix) [1], and ~ is a lower bound to the size of the problem input. More
specifically,
~ = 2:7=1 2:';=1 log bi,; 1 r
and bi,; = min { di,; , ai , f3; } .
Let us call bottleneck entry an entry di ,; such that Jl = di ,; -;- A; and Ii,; is greater
than 0 . We divide the af3-switching matrices into two groups: group 1 contains those
af3-switching matrices having at least one bottleneck entry di ,; such that di ,; is greater
than bi,;. All the other af3-switching matrices belong to group 2. It is possible to
show [2] that there are at most ( af3-switching matrices of type 1 (where ( is less
than n 2 , and is the number of nonzero entries in D), and that after each type 2 af3-
switching matrix, at least one entry is at least halved. From these observations the
time complexity follows.

4. GENERAL SWITCHING MODES AND HIERARCHICAL SYSTEMS

We consider now the case of general switching modes [7]. In this case, we have the
same constraints of the variable bandwidth systems, with the additional constraint that
simultaneous packet transmission between any I/O pair is not allowed. Mathematically,
this means that each af3-switching matrix must have only 0 or 1 entries: summarizing,
we have the constraints presented in the previous section, with the additional constraint
Si,; E {O, I}, for each i andj. An af3-switching matrix meeting this additional constraint
is called general af3-switching matrix, or Gaf3SM. Some of the properties already
established for variable bandwidth systems, carryover to the general switching modes
problem studied in this section, subject to some changes. For instance, the lower bound
to the minimum schedule length, L, is the same as before, with the following additional
inequality: L ~ { di ,;, for each i and j }. Thus, in this case we can have critical lines,
and critical entries, also (these being those entries di,;=L). The following tightness
parameters must be defined, in order to adapt the optimal algorithm shown above to
the present problem [7].
16

Pi = max { OJ :Lj'=1 di,i - O'i (L-1) }, for each ij

8j = max { OJ :Li=1 di,j - flj (L-1) }, for each jj
t/J = max { OJ :Li=1 :Lj'=1 di,i - K(L-1) }, for each j j
6i ,j = 0 if di,i is less than L, and {ii,j = 1 if di,j = L, for each i and j.

Let S be a GO'flSM. With an argument similar to those already used, it is possible

to show that if the sum of the i-th row (j-th column) of S is not smaller than Pi (8 j ,
respectively), each Si,j is not smaller than the corresponding {ii,j, and the sum of all its
entries is at least t/J, then D'=D-L has a lower bound equal to L-1, and the approach
presented above can be used again to generate an optimal TSA. Notice that in this case
D is not required to be in a pseudo doubly stochastic form, since such a form cannot
be defined. The existence of a GO'flSM with the above features can be established
by means of circulation in the bipartite network associated to D. Such a network is
identical to that presented in the previous section, with the addition of the (t,s) edge
directed from t to s. This edge has a capacity equal to K, and a lower bound equal to
t/J. The other edges are directed from left to right. For instance, all the other edges
incident in s, are outgoing. The generic edge (s, Ui) has capacity O'i and lower bound
Pij the edge (Ui,Vj) has capacity 1 and lower bound {ii,jj finally, the edge (vj,t) has
capacity flj and lower bound 8j •
We recall a classical result in circulation theory, Hoffman's Theorem: a circulation
exists in a network if and only if, for each subset w of the node set N, we have:

:L(i,j)E(N-w,w) Ai,i :$ :L(i,j)E(w,N-w) Xi,j ,

where Ai,j is the lower bound and Xi,i the capacity of the edge (ij).
Let w be such that s is in it, while t is not, in our bipartite network. Then, Hoffman's
inequality becomes the following:

where ei,j is 0 if di,j is 0, and is 1 otherwise. Observe that

:L iEw L,jEN-w ei,j 2: t L,iEw L,jEN-w di,j=

=i; ( K(L-1) + t/J - L,iEN-w L,j'=1 di,j - L,:'=1 L,jEw di,j + L,iEN-w L,jEw di,j) 2:
2: t ( K(L-1) + t/J - L L,iEN-w O'i - L L,jEw flj + L,iEN-w L,jEw di,j) 2:
2: t ( K(L-1) + t/J ) - L,iEN-w O'i - L,jEw flj + L,iEN-w L,jEW 6i,j .

If we substitute this expression for L,iEw L,jEN-w ei,j in (**), we get:

t
t/J :$ (K(L-1) + t/J),
and so t/J :$ K, which is always true by assumption.
Q.E.D.

A similar argument applies for the other three cases, namely when both s and tare
in Wj when t is in w, while s is notj nd finally when both sand t are not in w.
Like in the previous cases, an optimal algorithm can be constructed from this
property of the GO'flSM. It suffices to find a feasible circulation instead of a maximum
flow in the bipartite network [7J. Again, a multiplicity factor is in order, for the
algorithm to be polynomial time. Now, such a factor, J.L must meet the following
 17

constraints [7]:

J1 l .) )j , £or each'1 suchth at"n

:'S (La.-2:;-ld"
a"- 2:"
J
L-j:l
. sma11er th an
Si,j IS 0i;

l
I J~l I,)

J1:'S (Li3J
i3 _- "n_ . )d"J)j , lOr each J. such th at
£ "n . aller t h an f3i;
L-i:l Si,i IS sm
J 1=1 I,)

J1 _
< l(KL-2:~=l2:;=l d"J)j , wh en "n
K-2::~,2:;=18'.J) L-i:l
"n
L-j:l
.
Si,j IS
11
sma er
th
an
K;
J1 :'S L - di,j if Si,j = OJ
jl :'S di,j if Si,j =1.

Such a multiplicity factor allows the algorithm to have a time complexity O( <P
n 2 ), where <p is the time complexity of your favourite algorithm for finding a feasible
circulation [1]. In fact, at most O(n 2 ) Gof3SM must be in the TSA since each time
we produce one such matrix, either jl = di,j (and so in the remaining traffic matrix,
D', we have at least one more zero entry), or jl is smaller than di,j, in which case D'
will have at least one more critical item (row, column, or entry). The details of the
algorithm and its properties for general switching modes can be found in [7].
A similar approach can be used to give a polynomial time optimal algorithm for
the so called hierarchical switching systems (HSS) [6]. Both the inputs and the
outputs are grouped. Input group Uh is the multiplex of Ph inputs, which can produce
up to qh parallel signals, while up to Xk parallel signals are demultiplexed into Yk
streams, corresponding to the Yk outputs in group Vk. Such an arrangement have been
proposed to better utilize the system when there is low traffic between the inputs and
the outputs. Now, we have a physical constraint, additional to the basic ones: at most
qh (Xk' respectively) entries can be selected out of the Ph rows (Yk columns, resp.) of
group Uh (Vk ).
The corresponding switching matrix, called hierarchical switching matrix, or HSM,
must satisfy the basic constraints, and the following two:

2:iEUh 2:j:l Si,j :'S qh;

2:?:1 2:iE V. Si,i :'S Xk;

for each group.

Similarly to the general switching modes case, it is possible to show that a circu-
lation based optimal algorithm exists, with time complexity O( <p n 2 ). The details of
this result are left to the interested reader, and can be found in [6].
All the TSA problems that we have seen up to now can be optimally solved by
polynomial time algorithms. Unfortunately, it is not always so. In fact, there are con-
straint sets leading to NP-completeness [3,8,11,14,16]. In these cases, no polynomial
time optimal algorithms exist, and we are left with the choice between optimal but
slow (exponential time) algorithms, often of the branch-and-bound type, or fast but
suboptimal heuristics. However, network flow is of help even in this case, as many
heuristics are based on network flow. Such heuristics generally attain good perfor-
mances, namely produce schedules not far from the optimal, on the average. Let us
consider, for example, the incremental time slot assignment problem [8]. This problem
arises in a multimedia setting, when several types of traffic, each with specific features,
and different requirements, must be scheduled in the same system. The problem is
modelled by two traffic matrices, V (which can contain, for instance, voice traffic),
and D (data traffic). Traffic in V have a higher priority than that in D (voice has a
18

more strict delay constraint). So, we first derive an optimal schedule for V. In such
a schedule, some switching matrices will not contain all the possible non-zero entries,
and some D traffic can be scheduled in it. The problem consists in placing as much
D traffic as possible into the schedule for V, without violating the basic constraints
(at most one nonzero entry in each line of every switching matrix). This problem is
NP-complete even in very restricted cases. The NP-completeness proof is based on a
reduction of the well known timetable design problem to the TSA one [8]. An effective
heuristic for the incremental TSA problem have been proposed. Such an heuristic is
based on the Lagrangean Relaxation of the integer linear programming formulation of
the problem, consisting in a series of independent Hitchcock-Koopmans transportation
(namely, maximum profit flow) problems [8].

5. CONCLUSIONS

In this paper, we have presented several problems, called time slot assignment,
which are related to traffic scheduling in computer networks and telecommunication
systems based on interconnection networks. We have shown how network flow can be
used to efficiently solve such problems, whenever possible, namely when the problems
are not NP-complete.
Time slot assignment problems have simple expressions for the lower bound on the
schedule length. An interesting phenomenon has shown up to now in this setting:
whenever the lower bound can be achieved for any traffic matrix, the problem can be
solved by polynomial time network flow based algorithms. When it is possible to find a
traffic matrix for which any schedule is longer than the lower bound, NP-completeness
has always been established. An interesting question arises: is it possible to show that
the above phenomenon always holds, with no counter-examples?
Several problems are left open in this area. Two outstanding open problems are
the TSA of on line requests, and the scheduling of multicast communication. In the
first, instead of the traffic matrix gathering, packets must be scheduled as soon as they
are generated. This could speed up the overall system operation, and is strictly related
to the successfull ATM communication protocol. The second problem arises when the
same message must be sent from one source to multiple destinations, and is typical
of emerging applications, like teleconferencing. In this case, we cannot use a matrix
or a bipartite graph to represent the traffic to be scheduled: a hypergraph is a more
appropriate model for the problem.

REFERENCES
[I] R.K. Ahuja, T.L. Magnanti, and J.B. Orlin: "Network flows: theory, algorithms and applications",
Prentice Hall, New York, 1992.
[2] P. Barcaccia and M.A. Bonuccelli: "A polynomial time optimal algorithm for time slot assignment
of variable bandwidth systems", ACM-IEEE Trans. Networking., to appear.
[3] A.A. Bertossi, G. Bongiovanni, and M.A. Bonuccelli: "Time slot assignment in SS/TDMA systems
with intersatellite links", IEEE Trans. Commun., 35, pp. 602-608, 1987.
[4] G. Bongiovanni, D. Coppersmith, and C.K. Wong: "An optimal time slot assignment for an
SS/TDMA system with variable number of transponders", IEEE Trans. Commun., 29, pp.721-
726, 1981.
[5] G. Bongiovanni, D.T. Tang, and C.K. Wong: "A general multibeam satellite switching algorithm",
IEEE Trans. Commun., 29, pp.l025-1036, 1981.
 19

[6] M.A. Bonuccelli: "A fast time slot assignment algorithm for TDM hierarchical switching systems" ,
IEEE Trans. Commun., 37, pp. 870-874, 1989.
[7] M.A. Bonuccelli: "A polynomial time optimal algorithm for satellite-switched time- division mul-
tiple access satellite communications with general switching modes", SIAM Journal on Discrete
Mathematics, 4, pp. 28-35, 1991.
[8) M.A. Bonuccelli, I.S. Gopal, and C.K.Wong: "Incremental time slot assignment in SS/TDMA
satellite systems", IEEE Trans. Commun., 39, pp. 1147-1156, 1991.
[9) S. Chalasani and A. Varma: "Efficient time slot assignment algorithms for SS/TDMA systems with
variable bandwidth beams", Proc. IEEE INFOCOM '91, Miami, pp. 658-667, April 1991.
[10) K.Y. Eng and A.S. Acampora: "Fundamental conditions governing TDM switching assignments in
terrestrial and satellite networks", IEEE Trans. Commun., 35, pp. 755-761, 1987.
[ll) A. Ganz and Y. Gao: "SS/TDMA scheduling for satellite clusters", IEEE Trans. Commun., 40,
pp. 597-603, 1992.
[12) M.R. Garey and D.S. Johnson: "Computers and intractability: a guide to the theory of NP-
completeness", Freeman, S. Francisco, 1979.
[13] I.S. Gopal, G. Bongiovanni, M.A. Bonuccelli, D.T. Tang, and C.K. Wong: "An optimal switching
algorithm for multibeam satellite systems with variable bandwidth" ,IEEE Trans. Commun., 30,
pp. 2475-2481, 1982.
[14) I.S. Gopal, M.A. Bonuccelli, and C.K.Wong, "Scheduling in multibeam satellites with interfering
zones" , IEEE Trans. Commun., 31, pp. 941-951, 1983.
[15) I.S. Gopal , D. Coppersmith and C.K. Wong: "Minimizing packet waiting time in a multibeam
satellite system", IEEE Trans. Commun., 30, pp. 305-316, 1982.
[16) I.S. Gopal and C.K. Wong, "Minimizing the number of switching in an SS/TDMA system", IEEE
Trans. Common., 33, pp. 497-501, 1985.
[17) T. Inukai: "An efficient SS/TDMA time slot assignment algorithm", IEEE Trans. Commun., 27,
pp. 1449-1455, 1979.
[18] J.L. Lewandowski, J.W.S. Liu, and C.L. Liu: "SS/TDMA time slot assignment with restricted
switching modes", IEEE Trans. Commun., 31, pp.149-154, 1983.
[19) Y.K. Tham, "Burst assignment for satellite-switched and earth-station frequency-hopping TDMA
networks", lEE Proceedings, 137, pp. 247-255, 1990.
[20) W.W. Wu, "On the efficacy of traffic assignment in on Digital Satellite Communication, Montreal,
P.Q., Canada, October 1978.
[21] W.W. Wu, "Elements of digital satellite communication", Computer Science Press, Rockville MD,
1985.
ON THE DUALITY THEORY FOR FINITE
DIMENSIONAL VARIATIONAL INEQUALITIES

Castellani Marco 1 and Mastroeni Giandomenico2

1 Dipartimento di Matematica
Via Buonarroti, 2 -56127 Pisa
e-mail castella@gauss.dm.unipi.it

2Dipartimento di Matematica
Via Buonarroti, 2 -56127 Pisa
e-mail giando@karush.dm.unipi.it

Abstract. By means of a separation scheme, considered in a parametric image space,

it is possible to associate to a Variational Inequality a parametric dual problem from
which dual variational inequalities can be derived.

Key Words: Variational Inequality, separation, Lagrangean function, duality.

1. INTRODUCTION

Duality theory for optimization problems has been widely developed in the liter-
ature (see [2,5,6]). In [2] it has been proven that the dual problem can be obtained by
searching for a suitable separating function defined in the image space associated to
the optimization problem. Following this scheme, an optimality condition for a Vari-
ational Inequality (VI) has been stated [1] by means of separation arguments, in a
suitable parametric image space and a parametric dual problem has been associated to
the Variational Inequality. In this work we will show how it is possible to recover the
dual Variational Inequality introduced by Mosco [4], from the general parametric dual
problem and we will investigate the relations between the dual Variational Inequality
and a saddle point condition of a generalized lagrangean function associated to (VI).

In the sequel we will study Variational Inequalities defined on R n where (.,.) is

the Euclidean scalar product and we will consider the following notations:
let f : R,n ---+ R, be a convex function and 9 : R,n ---+ R be a concave function;
then
r(x*) = sup {(x·,x) - f(x)},
xER n

21
22

g*(x*)= inf {(X*,X)-g(X)}

",ERn

are the Fenchel conjugates, in convex and concave sense respectively, of the functions
f and 9 and are defined on the dual space (RR)* ~ RR;
the subdifferential of the convex function f at Xo ERR is the set

if C is a closed convex cone, in R m , the positive polar of C is defined by the set

C*={z*ERm: (z*,z)~O VzEC};

a function G : R R -+ R m is said C-concave iff

G(ax+(l-a)y)-aG(x)-(l-a)G(Y)EC, Vx,yERR, VaE[O,l].

2. DUALITY FOR VARIATIONAL INEQUALITIES

In this section we summarize the duality scheme for Variational Inequalities in-
troduced in [1]. Consider the following Variational Inequality

Find y E Domf such that

{ (GVI)
(F(y),x - y) ~ fey) - f(x), "Ix ERn,

where F : R n -+ R n and f : RR -+ R. For each y E Domf, let

</ley; x) = (F(y), y - x) + fey) - f(x);

then y is a solution of (GVI) if and only if the system

</l(y;x) > 0, x ERR

is impossible.
Fix a vector space V and a function <t'(y;.,.) : RR xV -+ R that satisfies the
following property:

o E V- de!
= {V E V: <t'(y;x,v) = </l(y;x), "Ix E Dom<t'(y; ·,v)}. (2.1)

Then, y E Domf is a solution for (GVI) iff the system (PS)y in the unknowns x and
V:
<t'(y;x,v) > 0,
{
V E Yo
is impossible, where Yo is any subset such that

o E Vo ~ V.
 23

Consider the following subsets of R xV:

K(y):={(u,v)eRxV: u=cp(y;x,v), xeR n }

'H:={(u,v)eRxV: u>o,veVo}.

It is easy to see that the system (PS}y is impossible iff the previous subsets are disjoint.

Proposition 2.1. Let Va vector space and

a function such that (2.1) holds. Then y is a solution for (GVI) iff

K(y) n'H = 0. (2.2)

The disjunction of K(y) and'H is studied introducing the class of separation func-
tions depending on a parameter w E Q

w(·,·;w): RxV -+ R

that fulfil the following conditions Vw E Q:

i) w(·,·;w):RxV-+R, w(u,v;w) = u+'Y(v;w);
ii) lev>o'Y(' ;w) 2 Vo
where 'Y : V x Q -+ R.

We observe that the condition ii) is equivalent to the following one

'H ~ lev>ow(', ·;w), Vw E Q;

Remark 2.1. If there exists an wE Q such that

sup (cp(y;x,v) + "Y(v;w») ~ 0,

X,v

condition (2.2) is true.

The previous remark leads to introduce a dual problem associated to the system
(PS)y; let A(y;·,·) : R n xQ ~ R

A(Yjx,w):= inf (-cp(y;x,v) -"Y(Vjw»

vEV

the Lagrangean function associated to the system (PS)y.

Definition 2.1. Let y E Dom! be fixed. The dual problem associated to the system
(PS)y is defined by
sup inf A(y; x, w) (PD)y
"'EO xERn
24

3. THE DUAL VARIATIONAL INEQUALITY OF MOSCO

In this section we will see how, considering the dual problem introduced by means
of the separation scheme, it is possible to recover the dual Variational Inequality in-
troduced by Mosco [4].
Consider the following generalized Variational Inequality:

Find fj ERn such that

{ (GVI)
(F(fj), x - fj} ;:: f(fj) - f(x), Vx E R n ,

where F: R n ---+ R n and f: R n ---+ R is convex and lower semicontinuous.

Under the assumption that F has an inverse, in [4] it has been introduced the
dual Variational Inequality:

Find u* E R n such that

{ (DGVI)
(_F-l( -u*), u* - u*} ;:: f*(u*) - f*(u*), Vu* E RR.

and it has been proven the following result:

Theorem 3.1. [4] If F has an inverse and f is a proper, convex and l.s.c. function,
then:

x is a solution of (GVI) {::? u* = -F(x) is a solution of (DGVI). 0

Consider the dual problem introduced in Definition 2.1: if we choose

rp(y; x,w) = (F(y), y - x} + f(y) - f(x + w)

where n = R n , ,(v,w) = -(v,w) and V = R n , Vo = {OJ, we obtain the following

lagrangean function:

AF(y;X,W) = (F(y),x - y) - f(y) - (w,x) - f*(-w).

The next lemma characterizes a saddle point of AF.

Lemma 3.1. (xo,w) ERR X RR is a saddle point for AFCii;x,w) if and only if Xo E
8f*( -F(fj)) and w = F(f}).
Proof. (xo,w) E R n x R n is a saddle point for AF(fj;x,w) {::? the following
inequalities are fulfilled:
a) (w,xo) + f*(-w)::; (w,xo) + f*(-w),
b) (F(y),xo) - (w,xo) ::; (F(fj),x) - (w,x),
From a), by the definition of subgradient, we obtain Xo E 8f*(-w); from b) we
get that it must be necessarily w = F(y). 0
 25

By means of Lemma 3.1 we obtain the main result of this section.

Theorem 3.2. If P possesses an inverse and f is a proper convex and l.s.c. function,
then
(1) y E R n is a solution for (GVI) {::} (y, P(y» E R n x R n is a saddle point for
AF(Yj x, w).
(2) it* = -P(jj) E R n is a solution for (DGVI) {::} (fi,P(fi» E R n x R n is a saddle
point for AF(Y; x, w).
Proof.
(1) *) Y is a solution for (GVI) {::} -P(y) E af(y) {::} [5] Y E af*( -P(y» and for
Lemma 3.1 we have the thesis.
{=:) From Lemma 3.1 we know that y E af*( -P(y» {::} [5] - P(j}) E af(Y)·

(2) *) We observe that u* is a solution of (DGVI) if and only if

p-l(_it*) E aJ*(it*).

Since P-l( -it*) = P- 1 (P(y) = Y we obtain

Y E aJ*( -P(fi».

{=: ) It follows immediately from Lemma 3.1. o

Remark 3.1. Theorem 3.1 follows immediately from Theorem 3.2; moreover Theorem
3.1 can be proven, more generally, in a Hausdorff locally convex space [4].
Remark 3.2. It is possible to obtain directly Theorem 3.1, without considering a sad-
dle point condition for A F , starting from the definition of the dual problem associated
to the system (PS)y
sup inf AF(Y;X,w).
wEn rERR

We have that

inf AF(Y; x,w) = {-( (P(y), y) + f(y) + f*(w»), if w = P(y),

xERR -00, otherwise.

Therefore the solution of problem (PD)y is w = P(y). Moreover if y is a solution for

(GVI), that is
-P(y) E af(y),
the following relations hold

-w E af(y) {::} [6] Y E aJ*(-w) {::} P-1(w) E aJ*(-w).

Putting w = -u we obtain P-l(_u) E af*(u) which is equivalent to (DGVI).

It is interesting to consider the case in which the Variational Inequality has the
integrability property that is 3h E C1(Rn;R) convex such that P = "ilh.
26

Proposition 3.1. If h and f are convex functions and h E Cl(Rn; R) is such that
F = Vh then y is a solution of (GVI) if and only if y is a solution of the following
extremum problem (P)
min {h(x)
xER n
+ f(x)}.
Proof. We observe that:

y is a solution for (GVI) {:} -F{y) E 8f(y) {:}

{:} 0 E F(y) + 8f(y) = Vh(y) + 8f(y) = 8(h + f)(y) {:}

y is an optimal solution for (P). o

Let us write problem (P) in the following way

min {hex) - (-f(x»}

xERn

and consider the Fenchel dual (D) associated to (P)

max {( - f).(x·) - h·{x·)}.

x·ER n

It is possible to show that the dual Variational Inequality (DGVI) is equivalent (up
to a symmetry) to the classical first order optimality condition for the dual problem
(D).

Theorem 3.3. Suppose that F possesses an inverse and that there exists a convex
function h E C1(Rn; R) such that F = Vh; then u* is a solution for (DGVI) {:}
-u· is an optimal solution for (D).
Proof. From the relation (- f)*(x*) = -1*( -x*), putting 'Ij;*(x*) = 1*( -x*) we
obtain that (D) is equivalent to the problem:

The following relations hold:

x* is an optimal solution for (D) {:} 0 E 8('Ij;* + h*)(x*) = 8'1j;*(x*) + 8h(x*) {:}

{:} 3v E 8h*(x*) such that - v E 8'1j;*(x*).

We observe that v E 8h*(x*) {:} x* E 8h(v) = Vh(v) = F(v) {:} v

Therefore:

x* is an optimal solution for (D) {:} _F-l(x*) E 8'1j;*(x*) {:}

{:} (F(x·), x· - x·) ? 'Ij;*(x*) - 'Ij;*(x*)

Putting u* = -x· and u· = -x· we obtain (DGVI). o
 27

Remark 3.3. We note that, given a saddle point (y, F(y» of AF, a solution of
(DGVI) can be obtained considering the vector -F(Y). If we want the saddle point
condition to be equivalent to the problems (GVI) and (DGVI), we need that F(y) is
a solution of (DGVI). This can be obtained putting z' = -u' in (DGVI), therefore
we have the following dual Variational Inequality:

Find z· ERn such that

{ (DGVI)
(-F- 1 (z'),z' - z·) ~ (-<p).(z·) - (-<p).(z·), Vz' ERn.

4. APPLICATIONS TO CLASSICAL LAGRANGEAN DUALITY

In the previous section we have obtained Mosco dual Variational Inequality by

means of Fenchel duality. Our aim, now, is to develop the study considering the
classical Lagrangean duality.
Unfortunately we will see that it will not be possible, in general, to define a dual
Variational Inequality in the dual space n: similarly to Wolfe duality we shall have to
consider a dual Variational Inequality defined on the product space nn xn, given by
the primal space by the dual space.
Consider the Variational Inequality:

Find y E Q such that

{ (VI)
(F(y), x - y) ~ 0, Vx E Q.

where Q = {x E R n : g(x) E C}, g(x) := (gl(X), ... ,gm(x», with g;(x) nn-+
R, i = 1, .. , m and C E R m is a closed convex cone with vertex in the origin of nm.
Following the scheme introduced in Sect.2, consider the Lagrangean function
AL(y;x,w) obtained choosing l'(v,w) = (v,w), V = n = R m , Vo = C and

. ) _ {(F(Y),y - x), if g(x) E C +v,

<p ( y, x, v - h .
-00, ot erWlse.

Proposition 4.1.

AL(y;x,w) = {~J;,)'x - y) - (w,g(x»), ifw E C',

otherwise.

Proof. By definition

AL(y;x,w) = inf «(F(y),x - y) - (v,w).

g(x)EG+v

Since

inf (-(v,w) = inf(-(g(x) _ c,w) = {-(w,g(x»), if wE C',

g(xlEG+v cEG -00, otherwise,

we achieve the thesis. o

28

In order to associate a dual Variational Inequality to the problem (VI) we will

consider a saddle point condition for the function AL, that can be characterized intro-
ducing another well-known equivalent formulation of the problem (VI) given by the
following system:
F(x) - AVg(X) = 0
{ (A,g(X») =0 (S)
A E C· g( x) E C.
The next result states the connection between (VI) and (S).

Lemma 4.1. Suppose that 9 is a differentiable C-concave function and that 3xo E RR
such that g(xo) E intC. Then y E Q is a solution of (VI) <=> 3X E R m such that
(y, X) is a solution of (S).

Proof. We observe that y is a solution for (VI) iff Y is a global optimal solution of
the problem (Ply
min(F(y), x - y).
zEQ

=» For the regularity condition, 3A E R m Kuhn-Tucker multiplier such that

~z((F(y),x - y) - (X,g(x»))(y) = 0
{ (A,g(y») =0 (4.1)
XE C·, g(y) E C

Developing (4.1) we obtain the thesis.

*') We observe that if y solves system (S) then it fulfills the Kuhn-Tucker conditions
of (Ply that, in the hypotheses of the lemma are sufficient for optimality. 0

Remark 4.1. If there exists hE C1(Rn, R) such that F = Vh, the system (S) is the
Kuhn-Tucker necessary condition of the problem (P), defined in section 3.

Lemma 4.2. Suppose that 9 is a differentiable C-concave functionj then (y, X) is a

solution of the system (S) <=> (y, X) is a saddle point of AL(yj x, A) on RR X R m .
Proof. (y, X) is a saddle point of AL(yj x, A) <=> the following inequalities hold
'Vx E RR and 'VA E C* :

(F(y), y - y) - (A, g(y» 5 (F(y), y - y) - (X, g(y») 5

5 (F(y),x - y) - (X,g(x»)
that is
-(A, g(y») 5 -(X, g(y») 5 (F(y), x - y) - (X, g(x »). (4.2)

The first inequality in (4.2) is equivalent to the following relations:

(X,g(y» = 0 and g(y) E C.

 29

In fact, since 0 E C· we have

-(X, g(y») 2: o. (4.3)

Ab absurdo if g(y) ¢ C = (C·)· there exists >. E C· such that (>., g(y») < o. Since
VOl > 0 we have 01>. E C· we obtain

(-OIA, g(y») $ -(A, g(y», VOl> 0

which is absurd since

-(OIA, g(y») -+ +00, if 01 -+ +00.

Therefore g(y) E C and we have that (X, g(y») 2: 0 which implies, with (4.3)

(X, g(y») = O.
The second inequality in (4.2) is equivalent to

W(x) = (F(y),x - y) - (X,g(x») 2: 0, Vx E Rn. (4.4)

Since W is convex and differentiable and jj is a global minimum point of W, (4.4) is

equivalent to
o = VW(jj) = F(jj) - Xv g(jj).
Therefore (jj, X) is a saddle point of AL(jji x, >.) if and only if (jj, X) is a solution of (S)
and the lemma is proven. 0

We are able to state the main result of this section.

Theorem 4.1. Suppose that 9 is a differentiable C-concave function and that 3xo E
R n such that g(xo) E intC. Then jj is a solution of (VI) {::} 3X E C· such that
(jj,X) is a saddle point of AL(jjiX,>') on R n xC·.

Proof. It follows from Lemma 4.1 and Lemma 4.2. o

Remark 4.2. From theorem 4.1, we can interpret Xas the solution of a (Lagrangean)
dual Variational Inequality (DVI) and the system (S) represents the relation between
the solutions jj and X.

The dual Variational Inequality (DVI) is defined in order to guarantee that Xis
a solution of (DVI) if and only if (jj, X) is a saddle point of AL(jji x, A).
Consider the set

Q = ((x,>.) E R n xC·: F(x) - >.Vg(x) =0 and A E C*}.

 30

The dual Variational Inequality (DVI) can be defined in the following way:

{
Find (y, X)_ E Q such that
(4.5)
(g(fl), >. - >.) 2: 0, V>' E C*.

Another way of formulating the system (S) by means of a Variational Inequality is the
following:
Find (y, X) E R n xC· such that
{ (4.6)
(r(fl, X), (x - y, >. - X») 2: 0, Vex, >.) E R n xC*.
where

Remark 4.3. In the hypotheses of Lemma 4.1 it is immediate to see that (fl,X) is a
solution of (4.5) or (4.6) if and only if it is a solution of the system (S).

5. CONNECTIONS WITH COMPLEMENTARITY PROBLEMS

Consider the Variational Inequality (VI) introduced in the previous section in the
hypothesys that the feasible set Q is a closed convex cone in R n containing the origin.
To this aim we define g(x) = x and C = Q. It is well known [3] that the problem (VI)
collapses into the generalized complementarity problem defined as follows:

Find y E Q such that

{ (GCP)
(F(y), y) = 0, F(y) E Q.

Applying the scheme introduced in Sect. 2 it is possible to associate to (GCP) a dual

problem. The Lagrangean function AL : R n xQ* -+ R becomes

AL(YjX,W) = (F(y),x -Y) - (w,x).

From Theorem 4.1 we have that y is a solution for (GCP) iff there exists wE Q* such
that (y,w) is a saddle point for AL(y,·,·) on R n xQ* that is
i) (y,w -w) 2: 0, Vw E Q*
ii) (F(y),x - fj) - (w,x) 2: -(w,fj), 'Ix.
From i) it follows that (w,y) = 0, since (w,y) 2: °
and 0 E Q.j therefore ii)
becomes
(F(y) - w, x) 2: (F(fi), y), 'Ix E Rn.

The previous inequality implies the following relations:

w = F(fl), (F(y), y) = 0.

If we suppose that F possesses an inverse we can write y = F- 1 (w) and, substituting

in i), we obtain:
(F-l(W),W-W) 2:0, VwEQ*.
 31

Therefore, since Q* is a convex cone, we can define the following dual of the problem
(GCP):
Find w E Q* such that
{ (DGCP)
(F-l(W),w) = 0, F-1(w) E Q

Remark 5.1. The same result would be obtained if we considered the Lagrangean
function AF defined in section 3, since the constraints of the feasible set Q are given
by the identity function g(x) = x. In fact (DGCP) has been obtained in [4] as dual of
the Variational Inequality (GVI) in which ¢>( x) is the indicatorfunction of the closed
convex cone Q.

REFERENCES

[1) C. A.ntoni , "On a Separation Approach to Variational Inequalities", this Volume.

[2) F. Giannessi, "Theorems of the Alternative and Optimality Conditions", Jou. Optimization
Theory Appl., vol.42, n.3, pp. 331-365, 1984.
[3) S. Karamardian, "Generalized Complementarity Problem", lou. Optimization Theory Appl.,
vol.8, pp. 161-167, 1971.
[4) U. Mosco, "Dual Variational Inequalities", Jou. of Mathematical Analysis and Applications,
n.40, pp. 202-206, 1972.
[5) R.T. Rockafellar, "Conjugate Duality and Optimization", Society for Industrial and Applied
Mathematics, Regional Monograph Series, 1974.
[6) R.T. Rockafellar, "Convex Analysis", Princeton University Press, Princeton, 1970.
SOME PROPERTIES OF PERIODIC SOLUTIONS OF LINEAR
CONTROL SYSTEMS VIA QUASI-VARIATIONAL
INEQU ALITIES

Cubiotti Paolo

Department of Mathematics
University of Messina
98166 Sant'Agata-Messina, Italy
e-mail: pcub@imeuniv.unime.it

Abstract. In this paper we present an application of Quasi-Variational Inequalities

to periodic solutions of linear control systems. Starting from a recent existence result
on generalized Quasi-Variational Inequalities, we derive an alternative theorem con-
cerning qualitative properties of the fixed-point set of certain multifunctions. Then
we consider a linear control system and we obtain, in particular, sufficient conditions
for the existence of a periodic solution with the property that the final value lies on
the relative boundary of the corresponding attainable set.

Key Words: Generalized Quasi-Variational Inequalities, fixed points, relative bound-

ary, linear control systems, attainable set.

1. INTRODUCTION

Let X be a nonempty subset of the n-dimensional space R n, and let {3 : X ~ 2x

and ¢> : X ~ 2Rn be two multifunctions. Then, the generalized Quasi-Variational
Inequality problem associated with X,{3 and ¢> (GQVI (X,{3,¢» in short) introduced
by Chan and Pang in [1] is to find (x, z) E X x R n such that
xE{3(x),ZE¢>(x) and sup (z,x-y)::;O.
yEP(x)
When ¢> is single-valued, the prefix "generalized" is usually omitted. One of the clas-
sical existence results for GQVI (X, {3, ¢» is the celebrated Chan and Pang's theorem,
which we now state (see [1], Corollary 3.1).

Theorem 1. Let X be compact and convex, {3 continuoU3 with nonempty compact

convex value3, ¢> upper 3emicontinuou3 with nonempty compact contractible value3.
Then GQVI (X, {3, ¢» admit3 at lea3t one 30lution.

33
34
Even though in the original formulation of Theorem 1 the values of the multifunction f3
were not supposed to be closed, the reader can verify that such requirement is necessary
in order to make the proof correct. Also, it is easily seen that Theorem 1 does not work
if f3 is not closed-valued. To this aim, it suffices to consider X = [0,1), f3(x) = )0, 1[
and <jJ(x) = {x}. Very recently, the following result was proved [2).

Theorem 2. Let X be a clo.!ed convex 8ub.!et of R n , f3 : X -t 2x and <jJ : X -t 2Rn

two multifunction.!, K ~ X a nonempty compact 8et. Assume that:

(i) the set <jJ(x) is nonempty and compact for each x E X, and convex for each x E K,
with x E f3(x);

(ii) for each y E X - X, the set {x EX: inf (z, y) ~ O} is closed;

zE4>(z)

(iii) f3 is a lower semicontinuous multifunction with closed graph, and, for every x E
X, f3( x) is a convex set which meets K;

(iv) for each x E X\K, with x E f3(x), one has

sup inf (z,x - y) > O.

yE.8(z)nK zE,p(z)

Then GQVI (X, f3, <jJ) has at least one solution belonging to K X R n •

Theorem 2 above is a nice improvement of Theorem 1. In particular, the original upper

semicontinuity assumption on the compact-valued multifunction <jJ is replaced by a
more general condition (assumption (ii» and the set X is not necessarily bounded.
We point out that if the assumptions (i) and (iii) are satisfied for some nonempty
compact set K ~ X, the (nonempty and closed) set E = {x EX: x E f3(x)} is
bounded, and each ¢(x) is convex for x E E, then assumption (iv) is also satisfied. To
see this, it suffices to consider, instead of K, the compact set K' = K U E.
In this paper, we first derive a more general version of Theorem 2, by relaxing the
coercivity conditions assumed in (iii) and (iv). Then we derive an alternative theorem
concerning qualitative properties of the fixed-point set of certain multifunctionsj in
particular, sufficient conditions for the existence of a fixed point which lies on the
relative boundary of the corresponding value are obtained. Finally, we apply our
result in order to obtain some properties of the set of the periodic solutions of a linear
control system. In some cases, our result guarantees the existence of a periodic solution
whose final value lies on the relative boundary of the corresponding attainable set.

2. PRELIMINARIES AND NOTATIONS

Given two metric spaces (5, p), (V, d) and a multifunction t/J : 5 - t 2 v, we say that
t/J is lower semicontinuous in 5 if the set t/J-(n) = {s E 5: 1jJ(s) n n:f:. 0} is open in 5
n
for each open ~ V. We recall that a nonempty-valued 1jJ is lower semicontinuous in
 35

S if and only if for any sequence {sd in S converging to s and any v E .,p(s), there is
a sequence {vd converging to v such that Vk E .,p(Sk) for all kEN. We say that the
multifunction .,p is upper semicontinuous in S if .,p-(n) is closed in S for each closed
n ~ V. We say that .,p is continuous in S if it is both lower and upper semicontinuous.
The graph of .,p is the set {( s, v) E S x V : v E .,p( s)}. We recall that if the graph of
the multifunction .,p is closed, then .,p has closed values; if .,p is upper semi continuous
with nonempty closed values, then the graph of .,p is closed; if Y is compact and .,p has
closed graph, then .,p is upper semicontinuous. We say that .,p has open lower sections
if the set .,p-({ v}) is open in S for any v E V. It is clear from the definition that if.,p
has open lower sections, then it is lower semicontinuous. Given A, B ~ V, A, B # 0,
the Hausdorff distance dH(A, B) is defined by

dH(A, B) = max{sup d(v, B), supd(z, A)},

vEA zEB

where d(v,B) = inf d(v,z). The multifunction.,p is said to be Lipschitzean if there

zEB
exists a constant L ;:::: 0 such that

for all s, u E S. If L < 1, the multifunction .,p is said to be a contraction. The reader
is referred to [3,4] for an excellent treatment of the basic facts about multifunctions.
If So E S, r > 0, we put B(so,r) = {s E S: peso,s) < r} and R(so,r) = {s E S:

p( so, s) ::; r}.

Henceforth, the space R n will be considered with the Euclidean norm induced
by the usual inner product C·). If A ~ R n , we shall denote by int(A), span (A),
aff(A), and conv(A), the interior, the linear hull, the affine hull and the convex hull
of A, respectively. Also, we shall denote by riCA) and or(A) the relative interior of
A (that is, the interior of A in aff(A)) and the relative boundary of A (that is, the
boundary of A in aff(A)), respectively. When A ~ C ~ R n , we shall denote by
intc(A) the interior of A in C. If, A,B ~ R n , we put A- B = {x -y: x E A, y E B},
A.L = {x E R n : (x,y) = 0 for all YEA}.

3. THE MAIN RESULT

The following result is the announced generalization of Theorem 2.

Theorem 3. Let X be a cloud convex subset of R n , K ~ X a nonempty compact

set, 1> : X ---t 2Rn and f3 : X ---t 2x two multifunctions. Assume that:

(i) the set 1>(x) is convex for each x E K, with x E f3(x);

(ii) the set 1>( x) is nonempty and compact for each x EX;

(iii) for each y E X - X, the set {x EX: inf (z,y)::; O} is closed;

zE4>(x)
36
(iv) f3 is a lower semicontinuous multifunction with closed graph and convex values.
Moreover, assume that there exists an increasing sequence {f,,} of positive real numbers,
with X n B(O,f!) '" 0 and lim
"-+00
f" = +00, such that, if one puts D" = B(O,f,,), for
each kEN one has:

(v) f3(x) n D" '" 0 for all x E X n D,,;

(vi) for each x E (X n D,,)\K, with x E f3(x), one has

sup inf (z,x - y) > O.

yE/i(X)nDk zE</>(x)

Then there exists at least one solution to GQVI (X, f3, <p) belonging to K x RR.

Proof. Fix kEN. By assumptions (iv) and (v) and Lemma 3.1 of [5], the multi-
function x -> f3( x) n D" is continuous on X n D" with nonempty closed convex values.
Since X n D" is convex and compact, the original proof of Theorem 2 (see [2], Proof
of Theorem 1) shows that there exists a vector x" EX n D" such that X" E f3(x,,) and

(3.1)

To see this, it suffices to observe that in the proof of Theorem 2 the convexity as-
sumption on the set <p( x) is needed only in the last step, in order to apply a minimax
result. Now, assumption (vi) implies x" E K. Since K is compact, the sequence {xA;}
admits a subsequence, still denoted by {xA;}, converging to a point x E K. Of course,
assumption (iv) implies x E f3(x). We claim that

inf (z, x - y) ::; 0 for all y E f3(x). (3.2)

zE</>(Z)

To see this, we argue as in the last part of the proof of Theorem 1 of [2]. Assume that
there exists fj E f3( x) such that

inf (z,
zE</>(i:)
x- fj) > O.

Choose kEN in such a way that K ~ Die and y E Die. Since the multifunction
x -> f3(x) n Die is continuous on X n D k , there exists a sequence {vd converging to y
such that

Now, consider the multifunction e:X -> 2 span (X -X) defined by

e(x) = {w E span(X - X): inf (z,w)

zE</>(x)
> O}.

Since assumption (iii) is actually satisfied for each y E span(X - X) (see [2]), the
multifunction ehas open lower sections and thus it is lower semicontinuous. Moreover,
 37

concavity of y -+ inf (z, y) implies that each set ~(x) is convex and open in span(X -
zEq,(x)
X). By Proposition 3 of [2], there exists 0' > 0 such that

inf (z, w) > 0 for all x E B(x, 0') n X, w E B(x - g, 0') n span(X - X).
zEq,( x)

Choosing kEN such that k 2:: k and max{lIxk - xII, IIvk - gil} < 0'/2, we obtain

inf (z, Xk - Vk) > 0.

ZEq,(Xk)

Since Vk E (3(Xk)nDk, this contradicts (3.1). Hence, (3.2) holds. Since by assumptions
(i) and (ii) the set ¢>(x) is compact and convex, Theorem 5 at p. 216 of [6] implies that
there is some z E ¢>( x) such that

sup (z,x - y) = sup inf (z,x - y) :::; 0,

yE{J(x) yE{J(x) zEq,(x)

that is our claim. o

Proposition 1. Let S be a closed convex subset of R", z E R" and x E riC S) such
that
(z, x - y) :::; 0, Ir/y E S. (3.3)

Then z belongs to (span( S - S)).L.

Proof. Let v,w E S. We have aff(S) = v + F, with F = span(S - v). Since

x E int(v+F)(S), hence x - v E intF(S - v), there exists ..\ > 0 such that

(x - v) + ..\(w - v) E S - v.

Thus, x + ..\(w - v) E S. By (3.3), we have (z, w - v) :::; 0. The arbitrariness of v and

w implies that (z, y) = 0 for all yES - S. This completes the proof. 0

Now we state our main result.

Theorem 4. Let X be a closed convex subset of R", (3 : X -+ 2 X a lower semi-

continuous multifunction with closed graph and convex values. Assume that there ex-
ists an increasing sequence {€k} of positive real numbers, with X n B(O, €1) =1= 0 and
lim €k = +00, such that for each kEN one has
k-oo

Then, at least one of the following assertions holds:

(A) The set {x EX: x E (3(x)} is unbounded.
(B) There exists x E X such that x E o,«(3(x)).
 38
(C) For each wE R n there exists x E X, with x E f3(x), such that (w,x - y) = 0 for
all y E f3(x).

Remark 1. It can be easily checked that Theorem 4 can fail if any of the assertions (A),
(B) and (C) is removed from the statement. To see this, one can take n = 1, X = R,
f3(x) = R as regards assertion (A)j n = 2, X = [-1,1] x R, f3(u, v) = {u} x [-1,1]
as regards assertion (B)j X = Rn,f3(x) = {O} as regards assertion (C). Also, it should
be noticed that assertion (C) is contradicted, in particular, if int(f3( x)) =1= 0 for every
x E X, with x E f3(x). To see this, choose any w E R n \{O}, x E {x EX: x E f3(x)},
v E int(f3(x)). If we pick any A > 0 such that VA = v + Aw E f3(x), we get

(w, vA) = (w, v) + Allwll 2 =1= (w, v).

It is easy to realize that the last fact contradicts (C). However, as we shall see in the next
section, there may arise other situations in which assertion (C) above is contradicted.

Proof of Theorem 4. By the classical Fan-Kakutani fixed point theorem (see [7],
Theorem 1), it is easily seen that the set E = {x EX: x E f3(x)} is nonempty. Now,
assume that assertions (A) and (C) do not hold. Thus, there is some w E R n such
that for every x E E one has

sup (w, z) > O. (3.4)

zE(:J(z;)-(:J(z;)

Consider the multifunction 1jJ : E --+ 2R n defined by putting

1jJ(x) = (f3(x) - f3(x)) n {y E R n : (w,y) > O}

for each x E E. By (3.4), each 1jJ(x) is nonempty. Moreover, since the multifunction
f3(x) - f3(x) is convex valued and lower semicontinuous (see [4], Theorems 7.3.11 and
7.3.15) and the set {y E R n : (w, y) > O} is open and convex, the multifunction 1jJ is
convex valued and lower semicontinuous. Of course, the closedness of the graph of f3
implies that the set E is closed. Using standard selection arguments, it is not difficult
to realize that there exists a continuous f : X --+ R n such that f( x) E 1jJ( x) for all
x E E. To this aim, it suffices to consider the lower semi continuous (see, for instance,
Lemma 2.3 of [8]) and convex valued multifunction 'Y : X --+ 2R n defined by

( ) _{1jJ(x), ifxEE,
'Y x - Rn, if x E X\E,
and then to apply Theorem 3.1111 of [91. Now, if we apply Theorem 3 with K = E and
</>(x) = {f(x)}, we get that there exists a point x E E such that

sup (f(x),x-y) $0.

yE(:J(i)

Suppose that x E ri(f3( x)). Then, Proposition 1 implies that f( x) E (f3( x) - f3( x)).L,
and this is a contradiction since f( x) E 1jJ( x). In fact, the inclusion f( x) E (f3( x) -
f3(x )).L n (f3(x) - f3(x)) implies f(x) = 0, while the inclusion f(x) E {y E R n : (w, y) >
O} implies f( x) =1= O. Thus, the point x lies in 8r (f3( x)), as desired. 0
 39

4. AN APPLICATION TO LINEAR CONTROL SYSTEMS

Given T > 0, two matrix functions A : [0, T] -+ R nxn and B : [0, T] -+ R nxm , a
nonempty compact set n ~ R m, we say that an absolutely continuous x(·) : [0, T] -+
R n is a solution of the linear control system
X' = A(t)x + B(t)u, u(t) E n (4.1)

if there exists a measurable u : [0, T] -+ n such that

x'(t) = A(t)x(t) + B(t)u(t) a.e. in [0, T].

For Xo ERn, we shall denote by AT(XO) the attainable set for the system (4.1) at time
T starting from Xo, namely, the set

AT(XO) = {x(T): x(·) is a solution of (4.1) in [O,T] and x(O) = xo} .

Of course, each fixed point of the multifunction AT gives a periodic solution of problem
(4.1) in [O,T]. We shall denote by AT(xo) the attainable set at time T starting from
Xo for the relaxed problem

X' = A(t)x + B(t)u, u(t) E conv(n). (4.2)

Before giving the main result of this section, we need some preliminaries.

Definition 1 (see [10], Definition 4.1.1 and Proposition 4.1.1). Let D be a nonempty
subset of R n , xED. The set

TD(X) = {v E R,n : liminf d(x

h-O+
+ hv, D)jh = O}
is called the Bouligand's contingent cone to D at x. When D is convex, TD(X) is simply
called the tangent cone to D at x.

We recall that if x E int(D), then TD(X) = R,n. Moreover, when D is convex, we have
the following nice characterization (see 10, Proposition 5.1.1).

Proposition 2. Let D be a nonempty convex subset of Rn, xED. Then

TD(X) = (U (l/h)(D - x)),

h>O

and thus it is a closed convex cone.

In the sequel we shall consider the tangent cone TD(X) when D is a closed ball. In this
occurrence, the next proposition provides a useful characterization.

Proposition 3. Let r be a positive real number, D the closed ball of radius r centered
at the origin in Rn, x E R n with IIxll = r.
Then we have TD(X) = {v E R n : (v, x) ::; O}.

Proof. If r = I, our claim is proved at p. 221 of [10]. The general case follows by an
easy application of Proposition 2. 0
 40
The following result is well-known. Its proof follows at once from Sections 10 and 11
of [11] and from Theorem lA at p. 164 of [12].

Theorem 5. Let A : [0, T] -+ RRXR, B : [0, T] -+ RRxm be two matrix functions whose
entries are Lebesgue summable in [0, T], and let n ~ R m be a nonempty compact set.
Then, for each Xo E RR, the set AT( xo) is nonempty, compact, convex, and

AT(xo) = AT(XO) = H(T)xo + AT(O), (4.3)

where H(t) is the principal matrix solution of the matrix differential equation

H' = A(t)H.
The main result of this section is the following.

Theorem 6. Let A: [O,T] -+ RRXR,B : [O,T] -+ RRxm be two matrix functions

whose entries are in the space LOO([O, T]), and let n
~ R m be a nonempty compact
set. Assume that there exists an increasing sequence {flo} of positive real numbers, with
lim flo = +00, such that, if one puts Dk = B(O, fk), for each kEN one has
k-+oo

C,,(t,x) := {u E conv(n): A(t)x + B(t)u E TDk(X)} f:. 0 (4.4)

for all (t, x) E [0, T] X D". Then, at least one of the following assertions holds:

(A) The set {x E RR : x E AT(X)} is unbounded.

(B) There exists x E RR such that xE8 r (AT(X)).

(e) There exists x ERR such that AT(X) = {x}.

Proof. We first assume, in addition, that the set n is convex. Let V be a subset of
[0, T] with vanishing Lebesgue measure such that each entry of the matrices A and B is
bounded in [0, T]\ V, and let to E [0, T]. Let A: [0, T]-+ RRXR and B : [0, T]-+ RRxm
be defined by

A(t) = {A(t), ift E [0, T]\ V, B(t) _ {B(t), if t E [0, T]\ V,

A(to), iftEV, - B(to), if t E V.

Of course, A and B have measurable and bounded entries in [0, T] and also satisfy
assumption (4.4). Moreover, the problem

x' = A(t)x + B(t)u, u(t) E n

admits the same solutions of the original problem. Thus, it is not restrictive to suppose
each entry of A and B to be, in addition, bounded in [0, T]. By Filippov's theorem
(see, for instance, Theorem 8.2.10 of [3]), problem (4.1) is equivalent to the differential
inclusion
x' E F(t, x), (4.5)
 41

where
F(t, x) = U{A(t)x+B(t)u}
uEIl

for each (t, x) E [0, T] x Rn. We recall that a solution of problem (4.5) is an absolutely
continuous x(·) : [0, T]-+ R n such that

x'(t) E F(t, x(t» a.e. In [0, T].

Now, fix kEN. Assumption (4.4) implies that

F(t, x) n TD.(X) :I 0, 'V(t, x) E [0, T] X Dk. (4.6)

At this point, it is not difficult to realize that all the assumptions of Theorem 7.1 of
[13] are satisfied. In particular, the measurability of the multifunction F(·,x) follows
from Theorem 8.2.8 of [3]; moreover, since for each fixed t E [0, T] the multifunction
F(t,.) is Lipschitzean with constant IIA(t)11 (the norm of A(t) as a linear operator from
R n into itself), we have that

dH(F(t,x),F(t,y» ;:; Lllx - yll

for all (t,x,y) E [O,T] x R n xRn , where

L = sup IIA(t)1I < +00;

tE[O,Tj

x
finally, if (t,x,u) E [O,T] R n xn,
K is chosen in such a way that n~ B(O,K) (of
course, the last inclusion is in the space Rm) and

c=max{L,K sup IIB(t)II},

tE[O,Tj

we easily obtain

IIA(t)x + B(t)ull :::; IIA(t)lIllxll + KIIB(t)1I :::; (1 + IIxlDc,

hence F(t,x) ~ (1 + IIxII)B(O, c) for all (t,x) E [O,T] X Rn. Thus, for each Xo E Dk
there exists a solution x(·) of problem (4.5) in [0, T] satisfying

x(O) = Xo, x(t) E Dk, 'Vt E [0, T].

Consequently, we have

By Theorem 5, each set AT(XO) is nonempty, compact, convex, and (4.3) holds. Thus,
by Proposition 1.4.14 of [3], the multifunction AT : R n -+ 2Rn is continuous. In
particular, since each AT( xo) is closed, the graph of AT is closed. Thus, all the
assumptions of Theorem 4 are satisfied with X = R n and f3(x) = AT(X). In particular,
the set {x E R n : x E AT(X)} is nonempty. Now, assume that assertion (C) above
42

does not hold. By (4.3), this implies that there exist two points z, v E AT(O) such that
z =1= v. If we pick w = z - v and choose any Xo E {x E R n : x E AT(X)}, since the
vectors
Zo = H(T)xo + z, Vo = H(T)xo + v
are in AT(XO), we have

(w, Zo - vo) = (w, z - v) = liz - vll 2 > 0.

It is easy to realize that the last fact contradicts assertion (C) of Theorem 4. Thus,
when n is convex, our claim follows directly from Theorem 4. For the general case, it
suffices to consider the relaxed problem (4.2) and then to apply (4.3). 0

We now give some comments and counter-examples to possible improvements of The-

orem 6. First, we show that Theorem 6 can fail if assumption (4.4) is not satisfied
(henceforth, we shall use subscripts to denote component of vectors).

Example 1. Consider the linear control system

For such system the assumption (4.4) of Theorem 6 cannot be satisfied. In fact, if we
°
take any t E [O,T], f" > 0, X = (Xl,X2) E R2 with Xl > 0, X2 > and IIxll = fk, by
Proposition 3 we have

C,,(t, x) = ([1,2] x [1,2]) n TD. (x) = 0

(as usual, Die = 11(0, fie». All the other assumptions of Theorem 6 are trivially satis-
fied. Using standard arguments (see, for instance, [11] and [12]) we get

AT(X) = x + ([T, 2T] x [T,2T]).

Of course, such a multifunction admits no fixed points.

The next example shows that assertion (A) can not be removed from the statement of
Theorem 6.

Example 2. Consider the linear control system

°
All the assumptions of Theorem 6 are satisfied. In particular, one can take fie = k.
In fact, given kEN, t E [O,T], x E R2 with IIxll = k, we have E C,,(t,x) (see
Proposition 3). We have
AT(X) = x + [-T,T]2.
 43

Since x E int(AT(x)) for all x E R2, assertions (B) and (C) are contradicted, as
claimed.

The next example shows that assertion (B) can not be removed from the statement of
Theorem 6.

Example 3. Consider the linear control system

{ xi = -Xl + U
x~ = -X2 + u, lul:5 1.
Again, all the assumptions of Theorem 6 are satisfied, with n = [-1,1). Again, one
can take fk = k. In fact, if t E [0, T), x E R n , IIxll = k, by Proposition 3 we have that
A(t)x = -x E TD.(X), hence 0 E Ck(t,X). For such system, we have

AT(X) = x exp( -T) + iCy, y) : Iyl :5 1 - exp( -Tn·

We easily get
{x E R2 : x E AT(Xn ~ B(O, v'2),
hence assertion (A) of Theorem 6 does not hold. Moreover, since the attainable sets
are line segments never reducing to a single point, assertion (e) is also contradicted,
as desired.

Now, let us observe that assertion (C) of Theorem 6 is very restrictive, and it is
contradicted in many significant situations. However, it can not be dropped from the
statement, as the following very simple example shows.

Example 4. Consider the linear control system

Xi = -Xl +U
{
x~ = -x2 + u, U E {O}.

All the assumptions of Theorem 6 are satisfied, with n = {O}. Once more, one can
take fk = k. In fact, if t E [0, T), x E R2, IIxll = k, by Proposition 3 we have that
A(t)x = -x E TD.(X), hence 0 E Ck(t,X). In this case we have

AT(X) = x {exp(-T)}
for all x ERn. Hence, the only fixed point of AT is the origin, we have AT(O) = {O}
and assertions (A) and (B) of Theorem 6 are contradicted.

Remark 2. We point out that Theorem 6 is significant in the case IIGII ~ 1, where
G : R n -+ R n is the linear operator defined by G(x) = H(T)x. In fact, for
IIGII < 1, Theorem 6 automatically works without assumption (4.4), giving a more
precise information; that is, at least one of assertions (B) and (e) holds. To see this,
observe that, by (4.3), the fact IIGII < 1 implies that the multifunction AT is a mul-
tivalued contraction with nonempty compact values, hence, by eovitz and Nadler's
 44

theorem (see [14]), it admits at least one fixed point. Now, assume that assertion
(C) is false. As we have already noticed in the proof of Theorem 6, this implies that
the nonempty compact convex set AT(O) contains at least two points. Hence, the set
Or(AT(O» is nonempty. Since by (4.3) we have

for all x ERn, the multifunction x -+ Or( AT( x» also is a multivalued contraction with
nonempty compact values. Hence, it admits at least one fixed point and (B) holds,
as desired. Moreover, we point out that when IIGII < 1, by Theorem 1 of [15], the
set {x E R n : x E AT(X)} is compact, hence bounded. We have already showed (see
Example 1) how in the case IIGII ;::: 1 assumption (4.4) is necessary to make Theorem
6 true. Also, Example 2 shows that if IIGII ;::: 1 then assertion (A) of Theorem 6 is not
necessarily contradicted, even if assumption (4.4) is satisfied.

REFERENCES

[1] D. Cha.n and J.S. Pang, "The generalized quasi-variational inequality problem" , Math. Oper.
Res. 7, pp. 211-222, 1982.
[2] P. Cubiotti, "An existence theorem for generalized quasi-variational inequalities", Set-Valued
Anal. 1, pp. 81-87, 1993.
[3] J.P. Aubin and H. F'rankowska, "Set-Valued Analysis", Birkhauser, Basel, 1990.
[4] E. Klein and A.C. Thompson, "Theory of Correspondences", John Wiley & Sons, New York,
1984.
[5] N.D. Yen, "On a class of discontinuous vector-valued functions and the associated quasi-
variational inequalities", Optimization, to appear.
[6] J.P. Aubin, "Mathematical Methods of Game and Economic Theory", North-Holland, Amster-
dam, 1979.
[7] K. Fan, "Fixed-point a.nd minimax theorems in locally convex topological linear spaces", Proc.
Nat. Acad. Sci. U.S.A. 38, pp. 121-126, 1952.
[8] O. Naselli Ricceri, "On the covering dimension of the fixed point set of certain multifunctions",
Comment. Math. Unit!. Carolin. 32, pp. 281-286, 1991.
[9] E. Michael, "Continuous selections I", Ann. of Math. 63, pp. 361-382, 1956.
[10] J.P. Aubin and A. Cellina, "Differential Inclusions", Springer-Verlag, Berlin, 1984.
[11] H. Hermes and J.P. Lasalle, "Functional Analysis and Time Optimal Control", Academic Press,
New York, 1969.
[12] E.B. Lee and L. Markus, "Foundations of Optimal Control Theory", John Wiley & Sons, New
York,1967.
[13] C. Castaing, Moussaoui S.B. and A. Syam, "Multivalued differential equations on closed convex
sets in Banach spaces", Set- Valued Anal. 1, pp. 329-353, 1994.
[14] H. Covitz and S.B. Nadler, "Multivalued contraction mappings in generalized metric spaces",
Israel J. Math. 8, pp. 5-11, 1970.
[15] J. Saint Raymond, "Multivalued contractions", Set- Valued Anal., to appear.
GENERALIZED QUASI-VARIATIONAL INEQUALITIES
AND TRAFFIC EQUILIBRIUM PROBLEM

De Luca Marino

Dipartimento di Matematica
Universita di Palermo,
Via Archirafi, 34
90123 Palermo, Italy
e-mail: deluca@csiins.csii.unirc.it

Abstract The model that expresses the traffic equilibrium problem in terms of Quasi-
Variational Inequalities is improved taking into account that: i) the cost function may
be discontinuous; ii) the cost function may be considered as a multifunction. Existence
theorems in such directions are given with examples and considerations, based on a
direct computational method, that justify this approach.

Key Words: Quasi-Variational Inequalities, generalized Quasi-Variational Inequal-

ities, traffic networks.

1. INTRODUCTION

In this paper we present some topics related to the Quasi-Variational Inequality

that models a traffic network when the costs are asymmetric and the demands are
elastic. The aim is to obtain suggestions and results for a possible improvement of
this model, based on the assumption that it is more natural to consider the costs as
multifunctions, possibly discontinuous.
In Sect. 2, we explain a direct computational method to find the solutions of the
Quasi-Variational Inequality, different from those based on "projection" or "descent"
techniques, (see [1], and the references therein), and we use this method in order to
point out, by means of an example, qualitative aspects and motivations for a more
general formulation.
In Sect. 3, we prove an existence theorem for the generalized Quasi-Variational
Inequality which comes from this improved approach.
Shortly, we recall that one of the possible model for a traffic network with fixed

45
46

demand leads to the following problem: to find a vector H E R m + such that

HE K, C(H)(F - H) ~ 0, "IF E K, (Ll)

with
K:= {F E R m + : ~F = pl.
Here, m is the number of paths for the I % pairs, CO : R m + ---+ R m + is the
path cost function, p E RI+ is the fixed demand and ~ is a 1X m matrix whose elements
are:
"I.. _
'l'JT -
{I, E
if Rr 'Rj
0, if Rr f. 'Rj r = 1,2, ... ,m, j = 1,2, ... ,1, (1.2)

where Rr is the r-th path and 'Rj is the set of those paths which link the % pair j.
A solution H of the Variational Inequality (1.1) is an equilibrium pattern flow in
the sense of J.G.Wardrop (1952), that is:

In this case, we have equilibrium costs Cj for every % pair j, obtained con-
sidering those paths on which the equilibri.um flows are greater than zero. Hence, we
can require that the demand p depends on this equilibrium costs or, more generally,
directly on the equilibrium pattern flow H.
Thus, if we put, for each H E R m +

K(H) := {F E R m + : ~F = p(H)} (1.3)

the Variational Inequality (1.1) becomes the following Quasi-Variational Inequality: to

find
HE K(H), C(H)(F - H) 2: 0, "IF E K(H). (1.4)

2. THE DIRECT METHOD

It is more convenient for our purpose to rewrite the Q.V.I. (1.4) in a different
form. To this end, let us observe that from (1.3) we can derive the values of 1 variables
Fi (which we may suppose, without loss of generality, be the first 1 variables), because
the matrix ~ in (1.2) is such that in each column there is an unique entry which is 1.
Thus, we have:
m

Fi = Pi(H)- L
r=1+l
~irFr, i = 1,2, ... ,1 (2.1)

and, in particular:
m

Hi = Pi(H)- L
r=l+l
~irHr, i = 1,2, ... ,1. (2.2)
 47

Now, we make the following assumption: there exist

E~Rm-l, O"i:E-+R, i=I,2, ... ,1,

such that
O"i(iI) ~ 0, viI = (H'+l, ... , Hm) E E
and (2.2) is fulfilled when Hi = O"i(iI).
We put:

i) F := (F'+b ... , Fm) E R(m-I)+;

ii) ~:= (<Pir), i = 1,2, ... ,1 r = 1 + 1, ... ,m;

iii) p(iI) := (h(iI), ... ,PI(H))

with Pie H) := PI (0"1 (iI), ... , O"I( iI), H);
. - - - - - - (m-I)+
IV) r(H,F):= (rI+ 1 (H,F), ... ,rm (H,F)) E R
_____ I __ _
with rr(H,F):= Cr(H,F) - 2:: <pirCi(H,F),
;=1
being Cr(H,F):= Cr (F1 (H), ... ,F,(H),F), r = 1+ 1, ... ,m;

v) k(H) := {F E R m - I : JF ~ p(H)}.
Then, if we replace H with H, k with K, and so on, and put r(H) for r(H, H),
the Q.V.1. (1.4) is transformed in the following one: to find

H E K(H), r(H)(F - H) ~ 0, "IF E K(H). (2.3)

In this section we are interested in the numerical evaluation of the solution for
Q.V.1. (2.3). Since the problem cannot be reduced to a convex minimization problem,
without making further assumptions, we cannot use the well-known algorithms of the
convex programming theory.
Some algorithms for the construction of the solution have been established, using
various techniques. We present here a direct method proposed in [2) for Variational
Inequalities, also generalized for Q.V.I. (see for more details [3)).
Let us start observing that every solution of the system

{ H E K(H) (2.4)
r(H) = 0

is a solution of (2.3), whereas any other solution H must belong to the boundary
BK(H) of K(H): instead, if H were an interior point (we observe that the interior of
K(H) is not empty), we should have r(H) = O.
To obtain a face of (m -I)-dimensional polyhedron K(H) it needs to equate some
constraint in v). So, let us describe a face « m - 1) - (h + k ))-dimensional in the
following way: setting Jk = {it,h, ... ,jd the set of indices such that

1 :::; jp :::; 1, p = 1,2, ... , k, k<

- 1·,
 48
Sh = {Sl,S2, ... ,Sh} the set of indices such that

Fs. = 0, 1+ 1 :S Sq :S m, q = 1,2, ... , h, h:S m - I;

I={1+1,1+2, ... ,m}\Sh, J={1,2, ... ,l}\Jk,wehavetheface

K(h,kl(H) = {F E R m - I : Fs. = 0, Sq E Sh, q = 1,2, .. .,h,

L ¢>jprFr = Pjp (H), jp E Jk, P = 1,2, ... , k, Fr 2:: 0, rEI, (2.5)

rEI

L¢>irFr = pj(H), j E J}.

rEI

Consider the restriction of the Q.V.1. to this face:

find H(h,k) E K(h,k)(H(h,k») such that

r(H(h,k»)(F(h,k) _ H(h,k») 2:: 0, VF(h,k) E K(h,k)(H(h,k»). (2.6)

As before, using the equality constraints in (2.5), we can express k components Ft, of

.
F,li E I, i = 1,2, ... ,k. Denoting by L = I\{1},12, ... ,ld and, for r E L,
r(h,k) = {rr - rl, if 3i : ¢>j,r = 1
r rr otherwhise
we obtain the Q.V.I., equivalent to (2.6):

find H(h,k) E Km_I_(hH)(H(h,k») such that

r(H(h,k»)(H(h,k»)(F(h,k) _ H(h,k») 2:: 0, VF(h,k) E Km_l-{h+k)(H(h,k») (2.7)

where Km-I-(h+k) is K(h,k) when the FI, are expressed in terms of the remaining
components of F and H(h,k). As before, we consider the system

(2.8)

Obviously, a solution of (2.8) solves Q.V.1. (2.7) and the following theorem give neces-
sary and sufficient conditions to this solution also solve the original Q.V.1. (2.3).

Theorem 2.1. Let us suppose that

¢>jp •• = 0, Vjp E Jk, VS q E Sh. (2.9)

If there exists H~h,k) E Km_I_(hH)(H~h,k») such that

r(h,k)(H!h,k») = 0,
then H~h,k) is a solution of Q. V.l. (2.9) if and only if

r E Sh
(2.10)
i = 1,2, ... , k .

Proof: see [3J. o

 49

A similar theorem, with slight modifications, gives necessary and sufficient condi-
tions when (2.9) is not fulfilled. IT the above conditions are not verified for each solution
of the system (2.8), or when (2.8) does not admit solutions, Q.V.1. (2.7) cannot have
solutions in the interior of Km-I-(h+A;). Then, we iterate the procedure, passing to a
face of less dimension, until to reach eventually the vertexes of K(H).

Remarks.
- from a computational point of view, the efficience of this direct method is con-
nected with the value of m - 1;
- the direct method can be very useful to study sensitivity analysis and paradoxes,
specially when costs and/or demands depend on parameter, as we shall show in
the next example.

Example (see [4]). Let us consider a network with four nodes PI. P2, Pa, P4 and
five links (Pl,P2),(PI,Pa),(P2,Pa),(P2,P4),(Pa,P4). We have only a travel demand
related to the OlD pair (PI. P4) that is connected by the following paths:

Rl = (P1,Pa) U (Pa,P4)
R2 = (PI. P2) U (P2, P4)
Ra = (P1,P2) U (P2,Pa) U (Pa,P4).
Assuming that the path costs are, respectively:

C1(F) = 11FI + 10Fa + 50

C2 (F) = 11F2 + 10Fa + 50
Ca(F) = lOFI + 10F2 + 21Fa + I'
and that the travel demand is

p(H) = 6 + (1 - )")Hl + (1 - )")H2 + 2)"Ha,

where the parameters I' and ).. are such that I' ~ 0 and 0 ~ ).. ~ t, the equilibrium
distribution H is given by the solution of the Q.V.I.:
a
HE K(H), C(H)(F - H) =L Ci(H)(Fi - Hi) ~ 0, \IF E K(H) (2.11)
i=l

with
K(H) = {F E RH : FI + F2 + Fa = p(H)}.
Following the above procedure, we get

Fa = p(H) - Fl - F2 • (2.12)

By (2.12), setting F = H, we have Ha ~ 0 <=> HE E with

 50

The set K(H), by (2.12), becomes:

(2.13)

Let us observe that, for a fixed A E 0, }), it is possible to choose HI + H2 so

small than

So, there exists H E E such that ECK( H), and this is remarkable for the next existence
results. The Q.V.1. (2.11) becomes:

L fi(H)(Fi -
3
H E K(H), Hi) ~ 0, TlF E K(H) (2.14)
i=1

with K(H) given by (2.13) and

f (H F) = 12F F _ 111 - 3).(H H ) _ 16 + J.L - (2J.L - 100».

1, 1+ 2 1 _ 2), 1 + 2 1 _ 2A

f (H F) =F 111 - 3).(H 16 + J.L - (2J.L -100».

2, 1 + 12F2 _ 1 _ 2), 1 + H)
2
_
1 _ 2). .

Now, we may apply the direct method: the system

does not admit solutions; then we consider the face of 8K(H) corresponding to

(and hence F3 = 0, by (2.12». Check if the system

(2.15)

has solutions such that f 2 (H) < 0, observing that in this case Th.(2.1) is automatically
satisfied because Sh = 0.
The vector H = (f, f' 0) satisfies (2.15) and the above condition. Hence H is a
solution of the Q.V.I.'s (2.14) and (2.11).
Now, if we increase the cost C 3 , considering a positive parameter J.L* > Jl the solution
H does not change but it results that the total cost

also does not change. Thus, we have the following paradox: the C08t on a path increa8e8
but the total cost is constant.
 51

3. AN EXISTENCE THEOREM FOR DISCONTINUOUS

MULTIVALUED COSTS

In this section we improve the above model based on a Quasi-Variational Inequal-

ity. We start recalling a well-known result (see, e.g., [1]): let E be a non empty subset
of R m and let us denote by rand K the functions

r: E -+ R,m
K: E -+ 2E;

let us consider the Q.V.1. (r, K)

HE K(H), r(H)(F - H) ~ 0, VF E K(H). (3.1)

We have:

Theorem 3.1. If E is compact and convex, if both rand K are continuous and if,
for every H E E, the set K(H) is non empty, closed and convex, then Q. V.I. (r, K)
(3.1) has at least one solution.

A recent result [5] improves Th.(3.1) by replacing the continuity of r with a more
general condition:

Theorem 3.2. Let E be compact and convex, and assume that : i) for every F E
E - E, the set
GF = {H E E: r(H)F S; O} (3.2)

is closed; ii) K is a continuous function from E into 2E.

Then, Q. V.I. (r, K) (3.1) has at least one solution.

Example (see [6]). Let us show by means of an example the utility of the assumption
i): let E be the set

p > 0, and let us consider the function r : E -+ R,2 so defined

It results:

It is easily to prove that the set GF defined by (3.2) is closed for ever FEE - E.
So, Th.(3.2) ensures existence of solutions for the Q.V.1. (3.1), nevertheless r is
discontinuous on E.
52

Another improvement is to consider also r as a multifunction:

The assumption i) becomes this one:

i') for each FEE - E the set GF = {H E E: Zer(H)
inf ZF ~ O} is closed.

Also in this case, a very recent result [7] ensures the existence of solutions for the
generalized Quasi-Variational Inequality which replaces the Q.V.I.(r,K) (3.1).
In traffic network model, also this generalization can be useful. It seems more
convenient to associate at each flow distribution F not a single-valued cost distribution
but a closed convex subset of Rm.
The results above are the starting point for our result, but we must overcome the
difficulty that, in our case, the assumption

can be not true. As alredy pointed out, K(H) is not a subset of E for each H E E.
Then, we have the following theorem where K is our multifunction, that is

K(H) = {F E R m : tfoF ~ p(H)}.

Theorem 3.3.

Let us suppose that E is a non empty convex, compact subset of R m and that
i) Pi : E -+ R+, i = 1,2, ... , I, are positive continuous functions;
ii) for each H E E, it results

{ K(H) n riCE) ::/: 0 (3.3)

K(H) ~ aff(E)

iii) for each H E E,r(H) is non empty, convex, compact, and i') holds; then there
exist
HE K(H) n E, Z E r(H) : Z(F - H) ~ 0, \IF E K(H) n E.

Moreover, if
ivy either H is an interior relative point of K(H) n E or Z= 0, there exist

HE K(H) n E,Z E r(H): Z(F - H) ~ 0, \IF E K(H).

By riCE) we denote the "relative interior" of E, that is the interior of E in its

affine hull aff(E), and Z is the projection of Z on aff(E).

Proof: We split the proof in the following steps:

1. K is lower semicontinuous and with closed graph.
2. The multifunction k(H) = K(H) n E is l.s.c. and with closed graph.
 53

3. The assumption iv) implies existence for the generalized Quasi-Variational Inequal-
ity.

Step 1. Let us recall that K is l.s.c. if the following condition holds:

a) VHn -+ H with Hn, H E E, '<In E N and '<IF E K(H), there exists a sequence
Fn, Fn E K(Hn) such that Fn -+ Fj K has closed graph if
b) VHn -+ H with Hn,H E E andVFn -+ F with Fn E K(Hn ), 'In EN it follows
FE K(H). The condition a) is equivalent to the following one:
a') for every open {} ~ Rm, the set {H E E : K(H) n {} =F 0} is a relatively open
subset of E.
By i), condition b) directly follows. To prove a), let us set, for a fixed F with
components Fr, r = 1,2, ... ,m, and for i = 1,2, ... ,I,n E N:
Ai= {r : 1 :5 r :5 m, cf>ir = I}
Br = {r E Ai : Fr > 0, Pi(H) :5 Pi(Hn)}
er = {r E Ai : Fr > 0, Pi(H) > Pi(Hn)}
and observe that, for r E er, the inequality

can be true for only a finite number of terms in the sequence H n , because of the
continuity of p. Let us call No this finite subset of N. Then, we set
Fr, ifr E Br,
F.rn = { F.r - Pi(H)-Pi(Hn)
#(c!,>
'f en d
' 1 r E i an n E
N\ .!YO,
l\T

0, otherwhise.
Thus, we have a sequence Fn whose components Frn are such that lim Frn = Fr and
n .... oo

Step 2. Also for K : H -+ K(H) n E the condition b) follows directly by i). For a),
let us observe that, thanks to ii), we get
K(H) n E = cl(K(H) n ri(E», '<IH E E.
On the other hand, the multifunction H -+ K(H) n ri(E) is l.s.c. by virtue of ii) and
a') .
Then we may apply the result of [7] to obtain a solution of the G.Q.V.1. (f, K).

Step 3.
- If HE ri(K(H) n E) then Z = 0 (it is enough to take F = H ± f), and our claim
is showedj
- If 2 = 0 it results
Z(F - H) = (2 + Z.L)(F - H) = 0
where Z.L is the normal component of Z to aff(E), and then our claim is ensured
also in this case. o
54

Example (see [6]). Let us show an application of Th. (3.3) to a discontinuos Q.V.I.
Let it be:
E = {H E R3 : HI = PI, 0 ~ H2 ~ ~2, H3 = P3}
K(H)={FER3 :FI =PI, O~F2~P2-H2, F3 =P3}
where Pi are positive numbers such that

PI - P3 + 2P2 > O.
Then, let it be :
rI(H) = PI + P2 - 2H2
r 2(H)= :F(H2)
r3(H) = P3 + P2 - 2H2
where :F(H2) is the function
if Pf ~ H2 ~ I!f
if H2 = Pf
if 0 ~ H2 ~ Pf.
The conditions of Th.(3.3) are fulfilled. By the direct method, the vector
P2
H = (PI, 4,P3)
is the solution of the Q.V.1. (r, K). Since H belongs to ri(K(H) n E), this vector is
also the solution of Q.V.1. (r, K).

REFERENCES

[1) P.T. Harker and J.S. Pang, "Finite-Dimensional Variational Inequality and Nonlinear Comple-
mentarity Problems: A Survey of Theory, Algorithms and Applications", Mathematical Pro-
gramming 48, 161-270, 1990.
[2) A. Maugeri, "Convex programming, Variational Inequalities and applications to the traffic equi-
librium problem", Appl. Math. Optim. 16, pp. 169-185, 1987.
[3) M. De Luca and A. Maugeri, "Quasi-Variational Inequalities and applications to equilibrium
problems with elastic demand", in "Nonsmooth Optimization and Related Topics", F.M. Clarke,
V.F. Dem'yanov and F.Giannessi (eds.), Ettore Majorana, International Science Series, Plenum
Press, New York, pp. 61-67, 1989.
[4) M. De Luca and A. Maugeri, "Quasi-Variational Inequalities and applications to the traffic
equilibrium problem: discussion of a paradox", J. Compo Appl. Math., 28, 163-171, 1989.
[5) P. Cubiotti, "Finite-dimensional Quasi-Variational Inequalities associated with discontinuous
functions", J. Optim. Theory Appl., 72, 577-582, 1992.
[6) M. De Luca and A. Maugeri, "Discontinuous quasi-variational inequalities and applications to
equilibrium problems", in "Nonsmooth Optimization. Methods and Applications", F. Giannessi
(ed.) Gordon £3 Breach Sc. Publ., 70-74, 1992.
[7) P. Cubiotti, "An existence theorem for generalized Quasi-Variational Inequalities", Set- Valued
Analysis, VoU, No.1, 81-87, 1993.
VECTOR VARIATIONAL INEQUALITY AND
GEOMETRIC VECTOR OPTIMIZATION

Elster Karl-Heinz 1 and Rosalind Elster2

1 European Laboratory for

Intelligent Techniques Engineering
52076 Aachen, Germany
2 Overseas Publishers Association
10405 Berlin, Germany

Abstract: The paper presents a Vector Variational Inequality which is closely con-
nected with the geometric vector inequality, introduced in [9]. This Vector Variational
Inequality can be obtained as a specialized vector Variational Inequality treated in
[1]. Relationships between the vector Variational Inequality and the geometric vector
optimization ~e shown, concerning a necessary condition and a sufficient condition for
properly efficient solutions of such special structured vector optimization problems.

Key Words: Vector Variational Inequality, geometric vector inequality, vector opti-
mization.

Mathematics Subject Classification 1991: Primary: 49J40 ; Secondary: 90C29 .

1. INTRODUCTION

Variational Inequality theory is a powerful tool for handling optimization prob-

lems. The development in this field shows the importance not only from the theoretical
point of view [11], [12],[14] but also for a wide spectrum of applied problems, for ex-
ample problems in Physical Oceanography [16], in traffic equilibrium [15], in free and
moving boundary problems [5],[13], in Operations research and in Economics [6].
A similar trend can be observed for vector Variational Inequalities. The introduc-
tion of a vector Variational Inequality has been initiated by Giannessi 1980 [10). Since
that time, in several papers interrelations were shown between a vector Variational
Inequality and a vector optimization problem [1], [3],[19] or a vector complementarity
problem [4].

55
56

Motivated by investigations of Giannessi [10] and Chen [1],[2], we consider a vec-

tor Variational Inequality obtained as a specialized vector Variational Inequality of
that one introduced in [1]. Furthermore, a close connection to the geometric vector
inequality introduced in [9] can be shown, which is a base for the treatment of geomet-
ric vector optimization problems [7],[8]. Moreover, this vector Variational Inequality
yields a necessary condition and a sufficient condition for properly efficient solutions
of geometric vector optimization problems.

2. NOTATIONS AND PRELIMINARIES

For proving assertions concerning relationships between the geometric vector in-
equality (briefly: GVI), introduced in [9], and the following vector Variational In-
equality (briefly:VVI), we use the same notations as in [7] and modify those given in
[1],[10],[20] for spaces, mappings, sets, and cones.
We denote by ~(R~) the set of nonnegative (positive) real numbers and by R n
the n-dimensional Euclidean space. Let be given the index sets

J2+ p := {O, 1, ... , s, s + 1, .. , s + p}, (2.1a)

(2.1b)

[k] := {mk, mk + 1, ... , nd, I [k]I:= card[k], k E J2+ p' (2.1c)

where
mo := 1, ml := no + 1, ... , m.+p := n.+p-l + 1, n.+p := n.

Moreover, let be
Sk ~ RI[kJl, k E J2+ p , open convex sets,
'+p
S := II Sk ~ Rn, S =10, (2.2)
k=O

Tk ~ RI[kJl, k E J2+ p , nonempty cones with the apex at the origin,

'+p
T := Sk II Tk ~ Rn, (2.3)
k=O

G~ : Sk - R, k E J2+ p, differentiable functions, (2.4a)

x:= (x(o], ... ,xr.+p]f E S, Gk(x):= G~(X[k])' (2.4b)

G: S _ R·+I+p with G(x) = (Go(x), ... , G.+p(x)f, (2.4c)

V~ : Tk - R, k E J2+ p, arbitrary functions, (2.5a)

Y := (yfo], ... , yr.+p])T E T, Vk(Y):= V~ (Y[k])' (2.5b)

V: T _ R·+I+p with V(y) = (Vo(y), ... , V.+p(y)f. (2.5c)

 57

We introduce the matrices:

Z := _ 1 _[ :
ZO Zo
:
1E R(S+I)Xp, E R~ Vk E J~ (2.6)
\I Z 11 2 • .
Zk

Zs .. , Zs

(evaluation matrix of the decision maker, \I. II Euclidean norm);

Ez:= (EI I ZE2),EI,E2 identity matrices in R(s+I)x(s+I) (2.7)

and in RPxp respectively;

r:=[~l ;2] where (2.8a)

rl := diag(-n hEJ~' r2 := diag( 'Yk hEJO, 'Yk E R+ Vk E J~+p (2.8b)

A(y) := [Alo(Y) 0]
A 2 (y)
where (2.9a)

AI(y):= diag('xk(y)hEJ~' A2 (y):= diag('xk(y»kEJo (2.9b)

and
'xk : T -+ ~, k E J~+p, non-negative functions such that (2.9c)

'xk(y) := 'x~(Y[kl)' where (2.9d)

Y:= [~l ~2]' where (2.10a)

yl := diag(Y~l hEJ~' y2:= diag(Y~l )kEJo. (2.10b)

By the matrices introduced above we obtain

(2.11a)

Az(y) := EzA(y) = (AI(y) I ZA 2(y», (2.11b)

yz := EzY = (yl I Zy2). (2.11c)

Finally, we introduce the following cones and spaces

(2.12)

LA := {A I A according to (2.8) or (2.9) or (2.10)} vectorspace of

(s + 1 + p) X (-) matrices A, (2.13a)

LAz := {Az I Az according to (2.11)} vectorspace of

(s + 1) x (-) matrices Az (2.13b)

Let R s +1 be partially ordered by the cone K : rl 1:. r2 if r2 - rl E K; we write

rl :S r2 (rl{:r2) if r2 - rl ~ K(r 2 - rl ~ [{O).
58

(Yz,·) means the product of Yz and an element of R" .

In the following we consider a Vector Variational Inequality problem in a finite

dimensional Euclidean space. To describe relationships of such a problem with prob-
lems in geometric vector optimization, we give the definition of the geometric vector
inequality as well as some of its properties which are important to state these relation-
ships.

Definition 2.1. A Vector Variational Inequality problem (VVI) consists in finding a

vector xES and a matrix Yz E Lyz such that

(-Yz,r - x){:F(x) - :F(r) Vr E S, (2.14)

:F:= EzF, F according to (2.4) .

Definition 2.2. The vector inequality

Yzx ::; Az(y)G(x) - EzV(y), xES, yET, (2.15)

is called Geometric Vector Inequality (GVI), if

3D = diag(d[~](x)hEJo.+p
with 0;01: d[k](X) E Tk 'Ilk E J2+ p such that
xES -+ { Yzx = Az(y)G(x) - EzV(y)
Vy = (I'odfol x), ... , I'8+P1.+p] ( x»T E T, I'k E ~ 'Ilk E J2+,-
The matrix Yz has the representation

Yz = EzY W = rD = diag(f'k£.'[k](x)hEJO•+p .
Furthermore, we recall to the well-known cone-convexity.

Definition 2.3. Let C C R 8H be a convex cone. The vector function G : S -+ R 8+l+P

is said to be C-convex, if for any xl, x 2 E S and for each cx:E R, 0 ::;cx:::; 1, holds:

For the vector function G : S -+ R 8 +l+p involved in the Geometric Vector Inequality
(2.15), the following properties are proved in [7].

Theorem 2.1. Let the geometric vector inequality (2.15) be given. Then, for each
xES, any 'Yk E ~,k E J2+ p , and

(2.16)
 59

holds:

(i) y E Tj

(ii) Yzx = Az(y)G(x) - EzV(y) 'v'Yz = Ezr Ja(x)j

(iii) Az(y) = Ezr = rZj

(iv) V is positively homogeneous of degree 1:

EzV(rV'G(x)) = EzrV(V'G(x)) 'v'r
according to (2.8).

Theorem 2.2. Any vector function G : S -+ Rs+t+p occuring in (2.15) is

R++1-convex.
In order to obtain a necessary condition and a sufficient condition for solvability
ofthe Vector Variational Inequality (2.14), we need the following results derived in [71·
For vector functions
F: S -+ R~, (2.17)

where

the order relation" <" in Rq will be extended to R~ by

-00 < h < +00 'v'h E R q •

Here +00 (-00) means the imaginary point whose every" coordinate" is +00(-00).
Especially for the elements of R~ we define

For functions (2.16) the definition of cone-convexity is given as follows:

Definition 2.4. Let C ~ Rs+I be a convex cone. The vector function G : S -+ R.~~+p
is said to be C-convex, if

(2.18)

Now we take into consideration appropriate properties of the geometric vector

inequality which are useful for proving relationships we are interested in.

Definition 2.5. Let M ~ R~ be an arbitrary subset. An element in E R~ (m E R~)

is said to be an upper (lower) bound of M in R~, if in ;::: m (m ~ m) '1m E M.
M(M) denotes the set of all upper (lower) bounds of M.
60

Definition 2.6. Let M ~ R~ .An element m* := supM (m* := inf M) in R~ , is

called a vector supremum (vector infimum) of M, if m* E M and m* :5 m for each
mE M, (m* E M and m* ~ m for each m E M) .
In the case M = 0, let sup M = -00 (inf M = +00).
If m* E M (m* EM), then m* := maxM(m* := minM) is called to be the vector
maximum (vector minimum) of M.

Remark 2.1. Note that for any subset M ~ R~

sup M exists and is a unique element of M ~ R~ .
If sup M E R9, then (supM}A; = sup{mk 1m E M},k = 1, ... ,q, that means SUP M
is the coordinatewise supremum of M.

Now we will introduce the Fenchel vector conjugate in the space R~~ .

Definition 2.7. Let G: S -+ R~~+P be R++I-convex. The vector function EzG* :

Lw -+ R~~, where

EzG*(W):= EzSUP{Wx - G(x) I xES} (2.19)

with Lw := {W I W E R(·+1+p)xn according to (2.1O)} is said to be the Fenchel

vector conjugate of G .

Remark 2.2. In the case where G*(W) E R s +1+p we have:

(2.20)

where each G k, k E J2+ p , depends only on the k-th row vector W(k), k E J2+ p , of the
matrix W, W according to (2.10).

The following lemma contains an arithmetic rule for Fenchel vector conjugates.

Lemma 2.1. Let G : S -+ R"+1+p be R++I-convex and let EzG* be the Fenchel
vector conjugate of G. Then, for the function

F: S -+ R s +1+P, where F(x) := A(y)G(x) + Hx+ ex

with
A(y) according to (2.9), A,,(y) > 0 Vk E J~+p,

H according to (2.10), ex:= (ex, ... ,exs+pf E R"+1+P,

the Fenchel vector conjugate is

EzF*(W) = Ez [A(y)G*(A-l(y)(W - H»- ex] VW E A(y)D(G*) + H, (2.21)

where

D(G*) := {A -l(y)(W - H) I SUP {A -l(y)(W - H)x - G(x) I xES} E R~~+P,

 61

A-ley) inverse matrix to A(y), W, H according to (2.10)}. (2.22)

Finally, in view of the mentioned relationships we give the

Definition 2.8. Let the vector function F : S -+ R~~+P be R++l-convex. Then

Wz = EzW, W according to (2.10) is said to be a matriz·subgradient of :F:= EzF :
S -+ R~~ at xES , if

:F(x) ~ :F(x) + Wz(x - x) Vz E S. (2.23)

The set of all matrix-subgradients of :F at x is defined as matriz.subdifferential of :F

at x and denoted by o:F(x).
IT o:F(x) =F 0, then :F is said to be matriz-subdifferentiable at x.

Definition 2.9. Let the vector function F : S -+ R~~+P be R++1-convex and let
:F: S -+ R+~, where :F(x) = EzF(z). Then the set

dom:F:= {z E S I :F(x) E RB+1}

is said to be the domain of:F. F is called properly R++1-convez if dom :F =F 0.

Generalizations of the subdifferential given above can be used for treating Vector
Variational Inequality problems related to certain vector optimization problems in
linear topological spaces [4],[19].

Lemma 2.2. Let:F: S -+ R~~ with :F(x) := EzF(x) be given, where the vector
function F: S -+ R~~+P is properly R++1-convex . Then

(i) Wzx~:F(x)+:F*(Wz) VxES, VWz according to (2.10), (2.24)

(ii) Wz E o:F(x) <:} Wzx = :F(x) + :F*(Wz) Vx E S. (2.25)

3. RELATIONSHIPS BETWEEN THE VECTOR VARIATIONAL

INEQUALITY AND THE GEOMETRIC VECTOR INEQUALITY

In the following we prove a necessary condition and a sufficient condition for

solvability of the Vector Variational Inequality (2.14).

Theorem 3.1. Let GVI be given according to (2.15). IT the VVI (see (2.14»

xES, Yz E o:F(x), :F(x):= Az(y)G(x), yET, G acc.to (2.4),

(-Yz,r-z)1:F(x)-:F(r) VrES (3.1)

is solvable, where each element Ak(y), k E J~+p , of Az(y) is positive and A(y)G : S -+
R~~+P is properly R++l-convex , then in GVI (2.15) equality holds.
62

Proof. Since Yz , occurring in the vector Variational Inequality (3.1), is a matrix-

subgradient of Az(y)G at xES, we have:

Az(y)G(r) ;::: Az(y)G(x) + Yz(r - x) Vr E S

and thus, regarding (2.11):

Ez(Yr - A(y)G(r» ~ Yzx - Az(y)G(x) Vr E S.

Using Definition 2.7, we obtain by Lemma 2.1

Ez sup{Yr - A(y)G(r) IrE S} = EzA(y)G*(A -l(y)Y) = Yzx - Az(y)G(x), (3.2)

yET, Y E A(y)D(G*), D(G*) according to (2.22).

Setting in (3.2)
-A(y)G*(A -l(y)Y) =: V(y) (3.3)
it follows:
Yzx = Az(y)G(x) - EzV(y), xES, yET. o

Theorem 3.2. Let VVI and GVI be given according to (2.14) and (2.15), respectively.
If in GVI the matrix Yz is chosen according to

V'Gk(X), k E J~+p, according to (2.16),

then the Vector Variational Inequality (2.14) is solvable.

Proof. Because of (3.4) we deduce the equality in (2.15). Then, regarding (3.3), it
follows

Yzx = Az(y)G(x) + EzA(y)G*(A- 1 (y)Y). (3.5)

Setting :F(x) := Az(y)G(x) , we obtain by Lemma 2.1

:F*(Yz):= EzF*(Y) = EzA(y)G*(A -l(y)Y).

Then (3.5) gives
Yzx = :F(x) + :F*(Yz).
By Lemma 2.2(ii) we get Yz E 8:F(x), that means Yz is a matrix- subgradient of:F at
xES:
:F(r) ;::: :F(x) + Yz(r - x) Vr E S. (3.6)

From (3.6) we imply immediately the existence of a vector xES and a matrix Yz such
that
(-Yz, r - x) ~ :F(x) - :F(r)Vr E S.

Thus, (2.14) is solvable. o

 63

4. GEOMETRIC VECTOR OPTIMIZATION ON THE VIEWPOINT

OFVVI

To discuss relationships between solvability of VVI (2.14) and solutions of dual

geometric vector optimization problems let us derive such problems by the following
way: Multiplying (2.15) by a vector z E K O according to

( 4.1)

and assuming

rEP, Pc R n a linear subspace, 0 < dimP =m <n (4.2a)

Al(y) = EI Vy E T (see (2.8), (2.10a», (4.2b)

Yi z =: Y(z) E pl., pl. the orthogonal complement ofP, (4.2c)

it follows from (4.1)

0:::; zTG(r) + >:(yfG(r) - zTV(y), (4.3)

x E pn S, y satisfying (4.2b), (4.2c),

because of
V(y) := Ez V(y), ( 4.4)
G(r) := (Go(r), ... G.(r)f,
._ (G(r»)
G(r).- G(r) , (4.5a)
G(r) := (G.+I(r), ... , G.+p(r))T
and

(4.5b)

where

Inequality (4.3), derived from (2.2), gives now motivation for establishing dual
geometric vector optimization problems [8]:

(P): G(x):= (Go(x), ... ,G.(x)f -t v -min (4.7)

x E B:= {x I x E Snp,G(x):= (G.+I(x), ... ,G.+p(x)f:::; O}

with S according to (2.2), and Gk(x), k E J~+p, according to (2.4).

VO(Y)]
(P*) : V(y) := [ : + II zZll2 L Vk(y) -t v - max (4.8)
Vo(y) kEJO

y := y(z) E B*:= U B;, K O := int lR~+I,

zEKO
64

B;:= {y ERR lyE T,A1(y) = El,Ylz =: Y(z) E p.L}

with T according to (2.3) ,

Vk(Y), k E J~+p, according to (2.5). (4.9)

For the vector optimization problems we define properly efficient solutions accord-
ing to [18).

Definition 4.1. A point xO E B (yO E B*) is called a properly efficient solu.tion of (P)
((P*», if there is azo E KO such that

Now we give a necessary condition for properly efficient solutions of (P) .

Theorem 4.1. Let xO E RR be a properly efficient solution of (P) . Then the point
xO is a solution of the following VVI:

Proof. Because the function

G(x) := (Go(x), ... , G.(x»T,

._ (G(X»)
G(x).- G(x) ,
G(r) := (G.+t(x), ... , G&+p(x»T

is satisfying (2.15), we infer from Theorem 2.2 that

G: B -+ R&+t+p, B:= {x ERR I x E snp, G(x) $ O}

is R·+1-convex and thus this holds also for

Then, there exists an element ZO E KO, such that each coordinate function

(4.10)

of

is a convex function [171,[211.

 65

Furthermore, because of the assumed differentiability of Gk for all k E J2+ p , each Fk

is differentiable, too. Hence each Fk, k E J2 , is differentiable.
Using convexity and differentiability of Fk, k E J2 , we conclude

Recall that in (4.11) each function Fk, k E J2 , depends only on certain coordinates of
x E B , since
Fk(X) = >'k(yo)Gk(X),Gk(X) according to (2.4).
Thus, it follows from (4.11) for all k E J2 :
8:F.(x)
~
Vi E [kJ,
{
aFk(X):= 0
Vi f/. [kJ, (4.12)
ax;
-!L 8:F,(x) Vi E [IJ, I E J O•
zUII.oi: 8Xi

By (4.11) and (4.12) we obtain

F(x) ;::: F(xO) + EzolF(xO)(x - XO) "Ix E B, (4.13)

where

Since

the inequality (4.13) is equivalent to

that means Azo(yO)Ja(xO) =: Y~o is a matrix-subgradient of

Azo(yO)G at xO E B :
( 4.15)
Y~o E 8A z o(yO)G(xO).
To prove ( - Y~o, X - xO) If.J we use the relation

Azo(yO)G(x) - Azo(yO)G(xo)

= G(x) - G(xO) - II :0°112 X(yO)[G(xO) - G(x)J "Ix E B

recall that Azo(yO) = (EI I ZOA2(yO». Assuming Azo(yO)(G(x) - G(xO» E KO and

taking into account the properly efficiency of xO E B , it follows

and hence a contradiction for x = xO . Therefore, we have Azo(yO)(G(x) - G(x O) ¢

KO and by the Corollary in [4J (Oi-a ;::: b implies bj{J) it follows from (4.14) that
(-Y~o,x-XO}If.J. 0
 66

Theorem 4.2. Let the Vector Variational Inequality

xEB, Y z oE8A z o(y)G(x), yET,(-Yzo,r-x)::;O VxEB, (4.16)

be solvable, where B and T are given according to (4.7) and (2.3), respectively,

(4.17)

._ (G(x»)
G(x).- G(x) according to (2.4) . (4.18)

Suppose that for an appropriate ZO EKO

(4.19)

(4.20)

Then x is a properly efficient solution of (P).

Proof. Since Yzo, occurring in (4.16), is a matrix-subgradient of Azo(y)G at x E B ,

we have:
Azo(y)G(r) ~ Azo(y)G(x) + Yzo(r - x) Vx E B.
Multiplying this inequality by an appropriate ZO E KO , it follows by (4.19) and by
r - x E P (recall that r, x E B):

ZOT Azo(y)G(r) ~ zOT Azo(y)G(x) VrE B.

Using (4.17),(4.18) and (4.20), we conclude

Therefore, x is a properly efficient solution of (P). o

REFERENCES

[1] G.-Y. Chen, "Vector Variational Inequality and its application for multiobjective optimization".
Chinese Science Bulletin, 34, 17, 969-972, 1989.
[2] G.-Y. Chen, "Existence of solutions for a Vector Variational Inequality". An extension of
Hartmann-Stampacchia theorem". lou. Optim. Th. Appl., V. 74, No.3, pp. 445-456, 1992.
[3] G.-Y. Chen and G.M. Cheng, "Vector Variational Inequality and vector optimization". In
Sawaragi, Y., Inoue, K. and Nakayama, H. (eds.), Toward Interactive and Intelligent Deci-
sion Support Systems, Vol. 1 , Proceedings 1986. Lecture Notes in Econ.and Math. Syst., 285,
Springer-Verlag, Berlin·Heidelberg·New York, 1987.
[4] G.-Y. Chen, and X.-Q. Yang, "The vector complementarity problem and its equivalence with
the weak minimal element in ordered spaces". l.Math.Anal.Appl. 153, 136-158, 1990.
 67

[5] J. Crank, "Free and Moving Boundary Problems". Clarendon Press, Oxford, 1984.
[6] S. Dafermos, "Exchange price equilibrium and Variational Inequalities". Math. Programming
46, 391-402, 1990.
[7] K.-H. Elster and R. Elster, "The geometric vector inequality and its properties". Report
3.200(735} Univ. of Pisa, Dept.of Mathematics (Optimization and OR). Forthcoming in Op-
timization, 1993.
[8] K.-H. Eelster, R. Elster and A. Gopfert, "On approaches to duality theory in geometric vector
optimi- zation", In "Methods of Operations Research", Anton Hain,Meisenheim/ Frankfurt
a.M., 60, 23-37, 1990.
[9] R. Elster, "Geometrische Vektoroptimierung". Dissertation B, Technische H och.chule M er.e-
burg, 1990.
[10] F. Giannessi, "Theorems of alternative, quadratic programs and complementarity problems". In
Cottle R.W., Giannessi F. and Lions J.-L.(eds.), "Variational Inequalities and Complementarity
Problems. Theory and Applications". J .Wiley & Sons, New York, pp.151-186, 1980.
[11] R. Glowinski, J .-1. Lions and R. Tremolieres, "Numerical Analysis of Variational Inequalities",
North-Holland, Amsterdam, 198!.
(12) P.T. Harker and J.-S. Pang, "Finite-dimensional Variational Inequality and nonlinear comple-
mentarity problems: A survey of theory, algorithms and applications" Mathem.Programming
48, 161-220, 1990.
(13) A. Kaplan and R. Tichatschke, "Stable Methods for Ill-Posed Variational Problems" . Akademie
Verlag, Berlin, 1990.
[14] O.G. Mancino and G. Stampacchia, "Convex programming and Variational Inequalities".
J.Optim. Theory and Appl. 9,1,3-23,1972.
[15] A. Maugeri, "Convex programming, Variational Inequalities,and applications to the traffic equi-
libriumproblem". Appl.Math.Optim., 16,169-185,1987.
[16) M.Aslam Noor, "Variational Inequalities in physical oceanography". In Rahmen, M. (ed.),
"Ocean Waves Engineering". Comput. Mechanics Publ.,U.K., pp.201-226, 1994.
[17] Y. Sawaragi, H. Nakayama and T. Tanino, "Theory of Multiobjective Optimization". Academic
Press Inc.,Orlando, 1985.
[18] K.P. Schonfeld, "Some duality theorems for the nonlinear vector maximum problem". Un-
ternehmen.forschung 14, 1, 51-63, 1970.
[19] C. Tammer, "A generalization of Ekeland's variational principle". Optimization 25,2-3, 129-141,
1992.
[20) YANG; X.-Q., "Vector Variational Inequality and its duality. Nonlinear Analysis". Theory,
Methods and Applications 21, 11, 869-877, 1993.
[21) P.-L. Yu, "Multiple-Criteria Decision Making". Plenum Press, New York and London, 1985.
TESTING A NEW CLASS OF ALGORITHMS
FOR NONLINEAR COMPLEMENTARITY PROBLEMS

Facchinei Francisco I and Joao Soares 2

I Dipartimento di Informatica e Sistemistica

Universita di Roma "La Sapienza"
Via Buonarroti 12, 00185 Roma, Italy
e-mail: soler@peano.dis.uniromal.it

2 Departamento de Matematica
Universidade de Coimbra
3000 Coimbra, Portugal
e-mail: jics@dragao.mat.uc.pt

Abstract: We investigate the numerical behavior of a new, simple algorithm for the
solution of nonlinear complementarity problems. The algorithm is based on a recently
proposed merit function which possesses some interesting theoretical properties. One of
the aims of the paper is to show that algorithms based on this merit function can be
viable also from the numerical point of view.

Key Words: Nonlinear complementarity problem, merit function, SCI function, global
convergence, superlinear convergence.

1. INTRODUCTION We consider the nonlinear complementarity problem:

F(x) ~ 0, x ~ 0, F(x?x = 0, (NC)

where F : lRn -+ lRn is continuously differentiable. Recent research on the numerical

solution of Problem (NC) has focused on the development of globally convergent algo-
rithms. To this end, two approaches have been investigated: the transformation of the
nonlinear complementarity problem into a system of (nonsmooth) equations and the
use of continuation methods. Strictly related to the first approach is the minimization
approach, which tries to solve Problem (NC) by performing the unconstrained or con-
strained minimization of some suitable merit function, while interior-point methods are
close to the continuation approach.
The main purpose of this paper is to illustrate the numerical behavior of an algorithm
for the solution of nonlinear complementarity problems, based on the unconstrained

69
70

minimization of a new merit function, and to shortly review some recent results on
merit functions.
In the next section we give some basic definitions which are needed in the sequel. In
Sect. 3 we review the properties of some existing merit functions and we present the
new merit function, which is shown to possess some definite advantages over previous
merit functions. In Sect. 4 we describe a new algorithm for the solution of Problem
(NC) which is based on the new merit function. Finally, in Sect. 5 numerical results are
presented.
A few words about the notation. If M is an n x n matrix with elements M i ;, i =
1, ... n, and I and J are index sets such that I, J ~ {i = 1, ... n}, we denote by MIJ the
III x IJI submatrix of M consisting of elements Mij, i E I, j E J. If v is an n vector,
we denote by VI the subvector with components Vi, i E I. A superscript k indicates the
iteration number; finally, II . II indicates the euclidean norm.

2. BASIC DEFINITIONS

In this section we review some definitions related to nonlinear complementarity prob-

lems and to differentiable properties of functions which will be used in the sequel.
A solution to the nonlinear complementarity problem (NC) is a vector x E lRn such
that
F(x) ~ 0, x ~ 0, F(xyx = o.
Associated to the solution x we define three index sets:
ct:= {ilxi > OJ, = 0 = F;(x)},
(3:= {ilxi 1:= {ilF;(x) > O}
The solution x is said to be nondegenerate iff (3 = 0.
We say that the solution x is b-regular iff, for every index set 6 : ct ~ 6 ~ ct U (3 the
principal submatrix V'F66(X) is nonsingular.
We say that the solution x is R-regular iff V' F",,( x) is nonsingular and the Schur
complement of V'F",,(x) in
( V'F",,(x) V'F"p(x»)
V'Fp,,(x) V'Fpp(x)
is a P-matrix (see below). We recall that the above mentioned Schur complement is
defined by V'Fpp(x) - V'Fp"(x)V'F",,(xt1V'F,,p(x). Note that this condition has been
called s-regularity in [22] but actually coincides with the notion of regularity introduced
by Robinson in [30] (see also [291, where the same condition is called strong regularity).
It is known that R-regularity implies b-regularity and that both these conditions
guarantee the local uniqueness of the solution x.
We recall the following definitions

- A matrix A E lRnxn is a Po-matrix iff every of its principal submatrices is non-negative.

- A matrix A E lRnxn is a P-matrix iff every of its principal submatrices is positive.

- A function F : lRn -+ lRn is a Po-function iff, for every x and y in lRn with x :F y,
there is an index i such that

Xi :FYi, (Xi - Yi)[Fi(X) - Fi(Y)1 ~ O.

 71

- A function F : IRn -+ IRn is a P-function iff, for every x and y in IRn with x i: y,
there is an index i such that

(Xi - Yi)[Fi(X) - Fi(Y)] > O.

- A function F : IRn -+ IRn is a uniform P-function iff there exists a positive constant
p such that, for every x and Y in IRn, there is an index i such that

(Xi - Yi)[Fi(x) - Fi(Y)] ;:: pllY - x1l 2 •

- A function F : IRn -+ IRn is monotone iff, for every x and Y in IRR,

(x - yY[F(x) - F(y)] ;:: O.

- A function F : IRn -+ IRn is strictly monotone iff, for every x and y in IRn with xi: y,
(x - ynF(x) - F(y)] > O.

- A function F : IRn -+ IRn is strongly monotone iff there is a positive constant p such
that, for every x and y in IRn,

(x - yY[F(x) - F(y)] ;:: plly - x1l 2 •

It is obvious that every monotone function is a Po-function, every strictly mono-

tone function is a P-function, and that every strongly monotone function is a uniform
P-function. Furthermore it is known that the Jacobian of every continuously differen-
tiable Po-function is a Po-matrix and that if the Jacobian of a continuously differentiable
function is a P-matrix for every x, then the function is a P-function.
In the remaining part of this section we recall some basic definitions about semis-
moothness and SCI functions.
Semi smooth functions were introduced in [20] and immediately showed to be relevant
to optimization algorithms. Recently the concept of semismoothness has been extended
to vector valued functions [28].
Definition 2.1 Let F : IRn -+ IRm be locally Lipschitz at x E IRn. We say that F is
semismooth at x iff
lim Hv' (2.1)
He8F(x+tv'}
v' ...... v,t!O

exists for any v E IRn, where 8F(y) is the generalized Jacobian of Clarke at y.
Semismooth functions lie between Lipschitz functions and CI functions. Note that
this class is strictly contained in the class of B-differentiable functions.
It is known that [20,28]:
(a) Continuously differentiable functions and convex functions are semismoothj also the
compositions of semismooth functions are semismooth.
(b) If a function F is semismooth at x, then F is directionally differentiable at x, and
the directional derivative F'(xj d) is equal to the limit (2.1).
We can now give the definition of SCI function.

Definition 2.2 A function f : IRn -+ IR is said to be an S(J1 function iff f is continu-

ously differentiable and its gradient is semismooth.
72

Semi smooth functions can be viewed as functions which lie between CI and C2 func-
tions. Semismooth systems of equations form an important class, since they often occur
in practice and many of the classical methods for their solution (e.g. Newton's method)
can be extended to solve such problems [23,28,26]. Analogously, many classical results
concerning the minimization of C 2 functions can be extended to the minimization of
SCI functions (see e.g. [24,2] and references therein), which, in turn, play an important
role in many optimization problems. Under very mild differentiability assumptions on
F, the new merit function we will introduce in the next section is an SCI function.

3. MERIT FUNCTIONS FOR NCP

Let C ~ IRn be given. A merit function for Problem (NC) is a nonnegative function
M : IRn -+ IR such that x is a solution of Problem (NC) iff x E C and M(x) = 0, i.e. iff
the global solutions of the problem

min M(x), x E C, (PM)

coincide with the solutions of Problems (NC). Usually either C = IRi. or C = IRn.
It is not difficult to find a merit function for Problem (NC), the challenging task
is to find a merit function which enjoys some properties which are useful from the
computational point of view. For example one could consider the merit function M(x) =
F(xyx whose global minimizers on the set C:= {xix ~ O,F(x) ~ O} are the solutions
of the complementarity problem (NC). But seeking these global minimizers is not easy
because, even in very simple cases, the structure of C may be very complicated and the
minimization problem can have stationary points which are not global solutions.
Properties which seem particularly attractive, and that we would like the merit func-
tion to enjoy, are the following:

• M is "smooth";
• every stationary point of Problem (PM) is a global solution of Problem (PM);

• the sets L(o):= {x E IRn: x E C,f(x) S o} are bounded.

These properties are obviously important from the computational point of view and
if they are met one can easily solve the nonlinear complementarity problem by applying
standard minimization algorithms to Problem (PM).
We now pass to describe two recently proposed merit functions and the new merit
function.

Fukushima merit function

The regularized gap function proposed by Fukushima [8] (see also [1]) is defined for
variational inequalities. When specialized to nonlinear complementarity problems it
becomes (to simplify we have fixed a free parameter)

Mfa is a merit function with C = IRi., so that solving (NC) is equivalent to finding
the global solutions of the simply constrained minimization problem {min Mfa(x) : x E
IRi.}. Furthermore the merit function Mfa enjoys the following properties:
- Mfa is continuously differentiable;
 73

- if the Jacobian of the map F is a positive definite matrix for every x, then every
stationary point of Problem (PM) is a global minimum point of Problem (PM) [8];

- if the map F is strongly monotone then the sets L(o:) are bounded [8].

Mangasarian-Solodov merit function

The implicit Lagrangian proposed by Mangasarian and Solodov [18] is defined as (to
simplify we have fixed a free parameter)

Mm. is a merit function with C = IRn, so that solving (NC) is equivalent to finding the
unconstrained global solutions of the problem {min M m .( x)}. Furthermore the merit
function Mm. enjoys the following properties:

- Mm. is continuously differentiable;

- if the Jacobian of the map F is a positive definite matrix for every x then every
stationary point of Problem (PM) is a global minimum point of Problem (PM)
[35];

- if the map F is strongly monotone and globally Lipschitzian, then the sets L( 0:) are
bounded [35].

We note that on one hand the implicit Lagrangian merit function is simpler than
the regularized gap function, since it only requires an unconstrained minimization, but,
on the other hand, the condition to have bounded level sets is much stronger for the
implicit Lagrangian than for the regularized gap function.
We point out that most of the existing global algorithms for the solution of nonlinear
complementarity problems are based, in a more or less explicit way, on a merit function.
Unfortunately often the analyses of these algorithms are based on conditions on the limit
points of the sequence generated by the algorithm itself; this is most unfortunate because
it makes the properties of the algorithm depend on the behavior of the algorithm itself.
Furthermore this way of analyzing algorithms makes it extremely difficult to identify in
a clear way classes of problems (e.g. monotone, strictly monotone, etc) for which the
algorithms work. Thus in this short review we have only considered those merit functions
which are analyzed in the framework illustrated at the beginning of the section. The
interested reader is referred to [17,4,14,25] for further examples of merit functions.

The new merit function

The new merit function we are about to describe is based on the following, simple,
two variables, convex function:

¢>(a, b) := Va 2 + b2 - (a + b).

The most interesting property of this function is that

¢>(a, b) = 0 a 2: 0, b 2: 0, ab = 0; (3.1)

note also that ¢> is continuously differentiable everywhere but in the origin. The function
¢> has been introduced by Fischer in [5,6], since then it has attracted the attention of
 74

many researchers and has shown to enjoy many favorable properties; for example it has
been used in [7,14,15,16,9,33,27).
In view of (3.1) it is obvious that the function
n
'l1(x):= LcP(x;,F;(x)?
;=1

is a merit function with C = m.n , and that solving (NC) is equivalent to finding the
unconstrained global solutions of the problem {min 'l1( x)}. The merit function 'l1 enjoys
the following properties [3]:

- 'l1 is continuously differentiable; furthermore, if every F; is an SCI function, then also

'l1 is an SCI function;

- if F is a Po-function then every stationary point of Problem (PM) is a global minimum

point of Problem (PM);

- if F is a uniform P-function then the sets L(a) are bounded.

We remark that the theoretical properties of 'l1 are superior to those of the regularized
gap function and of the implicit Lagrangian. In fact, on one hand 'l1 allows to solve the
nonlinear complementarity problem by an unconstrained minimization, on the other
hand the conditions under which we can be sure that every stationary point of the merit
function is a global minimizer and that the level sets are bounded are substantially
weaker. Furthermore also the differentiability properties of 'l1 seem more interesting,
and actually the SCI property is very important from an algorithmic point of view.
The merit function 'l1 has also been independently introduced by Geiger and Kanzow
[9). Their results are however weaker than those reported above or simply of a different
nature. In particular they showed that every stationary point of the merit function is a
global minimum point if F is monotone, while the level sets of 'l1 are bounded if F is
strictly monotone. The analysis of the second order properties of 'l1 is cruder than ours
and, to define superlinear convergent algorithms for the solution of the complementarity
problem, they require the solutions to be nondegenerate, which is not the case for the
algorithm described in the next section. On the other hand Geiger and Kanzow describe
an interesting algorithm for the solution of strictly monotone complementarity problems
which does not require the evaluation of the Jacobian of F.

4. AN ALGORITHM FOR NCP

The merit function 'l1 can be used in several ways to define globally convergent
algorithms for the solution of nonlinear complementarity problems: for example one
could simply use an off-the-shelf unconstrained algorithm to minimize 'l1. In this section
we use the merit function in a different, but classical way. We first define a fast, local
algorithm for the solution of Problem (NC). Then we globalize this local algorithm by
performing an Armijo-type linesearch using the "local" direction, but reverting to the
antigradient when the "local" direction is not a good descent direction for the merit
function. Note that this scheme follows exactly the same lines used in the classical
stabilization scheme for Newton's method for the unconstrained minimization of a twice
continuously differentiable function. The crucial point will be to show that eventually
the gradient direction is never used and the stepsize of one is accepted, so that locally the
global algorithm coincides with the local one thus ensuring a fast asymptotic convergence
 75

rate. To this end the SCI smoothness turns out to be a crucial property. We remark
that this is neither the only way to exploit the function IV, nor, possibly, the best one.
However, we note that the local algorithm enjoys several interesting properties and that
the overall global algorithm, in spite of its simplicity, performs surprisingly well.

The local algorithm

In this section we describe a local algorithm for the solution of nonlinear comple-
mentarity problems. The algorithm generates a sequence of points {xk} as

xk+ I = xk + dk.
To motivate the local algorithm we first consider a simplified situation. Suppose that
x is a solution of Problem (NC), that x is nondegenerate and that we know the sets A
and N of variables which are 0 or positive at x

A:= {ilx; = O}, N := {ilx; > O}.

Then, to determine XN we would only need to solve the system of equations F;(XN' OA) =
0, i E N. Provided that V FNN(XN, OA) is nonsingular we could also apply Newton's
method to this system, thus setting x';tt
= x~ + d~, where d~ is the solution of the
following linear system

(4.1)

Obviously, in general we do not know the sets A and N, and furthermore we would like
to avoid the nondegeneracy assumption, which is often not met in practice. We then
define dk in two steps. At each iteration we first estimate the sets A and N, thus fixing
some of the components of dk , then we calculate the remaining part of dk by solving a
reduced linear system. We approximate the sets A and N by the sets Ak and Nk defined
by

where e is a fixed positive constant. It is very easy to check that the following result
holds.

Proposition 4.1 Suppose that x is a solution of Problem (NC). Then, for every fixed
e there exists a neighborhood n of x such that, for every xk belonging to n

l' ~ Ak ~ l' U ,8
a ~ N k ~ aU,8.

Furthermore, if x is nondegenerate, then l' = Ak and a = N k.

Based on this results it then seems reasonable to define d k in the following way.

(4.2)

while d~k is the solution of the following linear system

(V FNkNk(xk>r d~k = -FNk(xk) + (V FAkNk(xk>r X~k. (4.3)

The definition of d~k is very natural, since if we estimate that A k is the set of variables
which are zero at x, by (4.2) we obtain
kH
XAk = 0. (4.4)
76

Regarding (4.3) we note that, if x is nondegenerate, Ak = 'Y by Proposition 4.1, so

that, since x~tl = 0 by (4.4), (4.3) reduces to (4.1). Roughly speaking the extra term
(VFA'N.(xk)yx~. in (4.3) is needed to deal with degeneracy. The following result can
be proved [3).
Theorem 4.2 Suppose that x is a b-regular solution of Problem (NC). Then there exists
a neighborhood n of x such that, if XO belongs to n, the algorithm defined above is such
that
a. All the linear systems which have to be solved aTe uniquely solvable.
b. {xk} -+ x.
c. The converge rate of the sequence {xk} to x is at least superlinear; if the Jacobian of
F is locally Lipscitzian at x, then the convergence rate is at least quadratic.
The properties reported in the theorem are the natural extensions of the classical
results for Newton's method for systems of smooth equations. It is worth pointing out
the following points.
- No nondegeneracy assumption is needed.

- Only reduced linear systems are solved at each iteration.

- The points generated can violate the constraint x ;::: O.

Although very simple we think the local algorithm outlined above enjoys some inter-
esting properties. If we compare it to the classical local linearization method of Josephy
and Robinson [30,13) we see that we have two advantages: the regularity assumption
required (b-regularity) is weaker than the R-regularity assumption used in [30); fur-
thermore the methods described in [30] require, at each iteration, the solution of a full
dimensional linear complementarity problem, which is obviously a computationally more
intensive task than solving a linear system. Recently Pang [21) has shown that it is pos-
sible to relax the R-regularity assumption, however, using this weaker assumption, the
linear complementarity problem that has to be solved at each iteration can have multiple
solutions, and a suitable one has to be chosen, and this is by no means an easy task.
There exist other local methods which solve, at each iteration, only a linear system, see,
e.g., [14,15,18,31,9,4). However, as far as we are aware of, all these methods require non-
degeneracy of the solution to get superlinear convergence; furthermore at each iteration
a full dimen3ionallinear system has to be solved.

Globalizing the local algorithm using the function 1lo

In this section we exploit the merit function to globalize, in a simple way, the local
algorithm. The global scheme is now described.
Global Algorithm
Data: Xo E IRn, c > 0, p> 0, p > 2, f3 E (0,1/2), u E (0,1).

Step 0: Set k =0
Step 1: (stopping criterion) If the stopping criterion is satisfied stop.
Step 2: Calculate the "local direction" dk according to (4.2)-(4.3).
If system (4.3) is not solvable set dk = -V1lo(x k ).
Step 3: If
(4.5)
 77

set xk+l = x" + d"

set k +- k + 1 and go to Step 1.
Step 4: (linesearch) If d" does not satisfy the following test

(4.6)

set d" = - Vw( x"). Find the smallest i = 0,1,2, ... such that

(4.7)

set x"+1 = x" + 2- i d"

set k +- k + 1 and go to Step 1.

A few comments are in order. At Step 1 any reasonable stopping criterion can be
used. Note that in our case we can use classical measures of optimality, like the norm of
the vector of residuals, but also measures connected to the merit function as, for example,
the norm of the gradient of w. At Step 2 we try to calculate the "local" search direction
defined by (4.2)-(4.3). If this direction is not well defined we switch to the antigradient
of the merit function. Then we exploit the fact that, if the nonlinear complementarity
problem is solvable, the optimal value of W is o. So, if, for some constant u E (0,1),
test (4.5) is satisfied, we accept the stepsize of one. If this test is passed an infinite
number of times this will obviously lead to the function value tending to zero as desired.
Should test (4.5) not be satisfied, we perform in Step 4 a classicallinesearch procedure
to determine the step size. In this latter case we possibly switch to the antigradient, see
test (4.6), in order to ensure that the search direction is "sufficiently" downhill.
The aim of the acceptability test of Step 3 is twofold. On one hand it gives us one
more chance to accept the stepsize of one, on the other hand it makes it easier to prove
the superlinear converge rate of the algorithm (see [3]). A test close to (4.5) has been
proposed, with similar purposes, in [26].
The following result can be proved [3].

Theorem 4.3 It holds that:

a. each accumulation point of the sequence {x"} generated by the algorithm is a station-
ary point of w;
h. if one of the limit points of the sequence {x"} is a b-regular solution of Problem (NC),
then {x"} -+ X;
c. if {x"} -+ if and if is an R-regular solution of Problem (NC) and each component F;
of F is S(Ji in a neighborhood of x, then:
1. eventually d" is always the "local" direction defined in the previous subsection
(i. e. the antigradient is never used eventually);
~. eventually the stepsize of one is always accepted so that x"+1 = x" + d" ;
9. the convergence rate is quadratic.

We note that in general what we can guarantee is that every limit point x (if any)
is a stationary point of w. If W(if) = 0 then x is also a solution of the nonlinear
complementarity problem. According to what seen in Sect. 3, we can be sure that every
limit point of the sequence generated by the algorithm is a solution of Problem (NC) is
F is a Po-function. If F is a uniform P-function, we can also guarantee the existence
78

of a limit point. Actually in this latter case, it is elementary to show that the whole
sequence converges to the unique solution of the complementarity problem.
In this section we have considered a basic algorithm; some simple modifications
which preserve all the properties described here, but which can be useful from the
computational point of view, will be illustrated in the next section.

5. NUMERICAL RESULTS
In this section we report on some numerical results obtained with the basic algo-
rithm described in the previous sections and with some variants of it. To motivate these
variants we first recall that it has been often observed in the field of nonlinear com-
plementarity algorithms that the linesearch test used to enforce global convergence can
lead to very small stepsizes; in turn this can bring to very slow convergence and even to
a numerical failure of the algorithm. To circumvent this problem many heuristics have
been used (see, e.g., [11,22,14]). Here we propose to substitute the linesearch test (4.7)
by the following non monotone linesearch:

(5.1)

where the Mk are nonnegative integers bounded above for any k. This kind of linesearch
has been first proposed in [10] and since then it has proved very useful in the uncon-
strained minimization of smooth functions. Adopting the same proof techniques used in
[10] it is easy to see that all the results described in the previous section still hold if we
substitute the linesearch (5.1) to (4.7) in the Global Algorithm.
A variant of test (4.5) in line with the nonmonotone approach and which still pre-
serves all the properties of the Global Algorithm can also be envisaged.
llt(x k + dk ) ~ u. max k llt(xk-i), (5.2)
J=O,l, ... ,M

where, again, Mk are nonnegative integers bounded above for any k. Exploiting this
ideas we have then tried three algorithms.
Algorithm A: A pure monotone algorithm. This is the basic Global Algorithm de-
scribed in the previous section.

Algorithm B: A pure nonmonotone algorithm. This is the basic Global Algorithm

described in the previous section where we have used the linesearch (5.1) and the
test (5.2). We set
Mk = minCk, 10).
This simply amounts to setting Mk = 10 except in the first 9 iterations when 10
previous function values are not available and then only those available are used.

Algorithm C: A mixed algorithm. Since it is known that the nonmonotone strategy is

very effective when a Newton-like direction is used, but its use can be troublesome
with the antigradient, we tried a mixed strategy; this is also suggested by some
recent research on nonmonotone strategies [32]. The basic idea is: use the non-
monotone approach if it seems that we are already in a "safe" zone, otherwise use
the more conservative monotone strategy until a "safer" (closer to the solution)
zone is reached. We realized this basic idea in a very rough and yet seemingly
effective way. If in the first iteration the stepsize of one is accepted we use the
nomonotone strategy of Algorithm B, otherwise we use algorithm A for the first
five iterations and then switch to Algorithm B.
 79

Table 1. Results for Nonlinear Complementarity Problems

Problem Dimen. Start. point Alg. A Alg. B Alg. C
it. F it. F it. F
Kojima-Shindo 4 a 8 17 9 14 9 14
4 b 7 15 12* 37 7 15
Spatial eq. 42 a 27 50 31 60 27 44
42 b 26 41 34 51 25 38
Traffic eq. 50 a 32 208 19 36 19 36
Nash-Cournot 5 a 1 2 1 2 1 2
5 b 6 7 6 7 6 7
5 c 9 10 9 10 9 10
Nash-Cournot 10 a 6 7 6 7 6 7
10 b 9 10 9 10 9 10
Walrasian eq. 1 4 a 8 11 8 11 8 11
Walrasian eq. 2 4 a 22 45 22 45 22 45
HS34 8 a 20 75 14 41 14 41
HS66 4 a 5 6 5 6 5 6
Watson 1 5 a 20 21 20 21 20 21
5 b 16 20 16 20 16 20
5 c 22 40 21 35 22 40
5 d 1 2 1 2 1 2
Josephy 4 a 7 13 13 31 13 31
4 b 7 16 12 37 7 16
4 c 8 14 14 32 14 32
4 d 4 5 4 5 4 5
4 e 3 4 3 4 3 4
4 f 6 10 6 10 6 10
80

Table 2. Results for Linear Complementarity Problems

Problem Dimen. Start. point Aig. A Aig. B Aig. C
it. F it. F it. F
Murty 1 10 a 6 9 10 11 10 11
20 a 14 39 20 21 20 21
30 a 32 106 17 19 17 19
40 a 19 64 25 27 25 27
50 a 52 223 33 35 33 35
Murty 2 10 a 5 16 12 20 5 16
20 a 22 25 23 34 23 35
30 a 5 16 12 20 5 16
40 a 6 22 43 57 6 22
50 a 8 31 53 67 55 79
Kanzow 1 20 a 1 2 1 2 1 2
Kanzow 2 20 a 1 2 1 2 1 2
Watson 2 5 a 2 8 2 8 2 8
5 b 1 2 1 2 1 2
5 c 1 2 1 2 1 2
5 d 2 3 2 3 2 3
HS35 4 a 2 3 2 3 2 3
HS76 7 a 3 8 3 6 3 8

In all the algorithms the following constants were used:

c = 1, j3 = 10-4 , p= 2.1, u=0.9.

We tried the algorithm on several test problems taken from the literature, we con-
sidered nonlinear (Table 1) but also linear problems (Table 2); the test set includes
problems that are not Po and problems that are not R-regular or b-regular at the solu-
tion. Some details on these problems, the starting points, and adequate references are
reported in the appendix. For each algorithm we report the number of iterations and
the number of function evaluations needed to reach a point where the euclidean norm of
the residual, min( xl:, F( xl:», is less than 10- 6 • The asterisk indicates that convergence
occurred to a different solution (see Appendix). By these results it is apparent that
the three algorithms seem to be robust. Algorithm A, as expected, requires in some
cases many functions evaluations because in some iterations many halvings are needed
to satisfy the acceptability test. Algorithm B behaves better in this respect, however
the nonmonotonicity is not always beneficial. On the whole the hybrid Algorithm C
seems the best one, since it combines the good features of the other two algorithms.
Algorithm C compares well also with the results reported in the literature (see, e.g.,
[4,9,11,14,15,22,31]) especially if one considers that the computational cost per iteration
is low.

Acknowledgement. We are grateful to Christian Kanzow, who read a previous version

of this paper and gave some useful suggestions.

APPENDIX

In this appendix we report some relevant data, the source and the starting points for
 81

the test problems used in the previous section. The source reported is not necessarily
the original one.
Kojima-Shindo problem: See [22]. F(x) is not a Po-function. This problem has
two solutions: xl = (0.5V6, 0, 0, 0.5) and x 2 = (1,0,3,0); xl fails to be R-regular and
is degenerate. In the results of Table 1 convergence always occurred to Xl except in
the case indicated by an asterisk. The linearized complementarity problem at 0 has no
solution.
Starting points: (a) (0,0,0,0), (b) (I, I, I, 1).
Spatial price equilibrium problem: See [22]. F is a P-function and the unique
solution is therefore R-regular.
Starting points: (a) (0,0, ... ,0), (b) (I, 1, ... ,1).
Traffic equilibrium problem: See [22].
Starting points: (a) All the components are 0 except Xl, X2, X3, XlO, Xu, X20, X21l X22, X29, X30,
x40, X4S which are 1, X39,X42,X43,X46 which are 7, X4t,X47,X48,XSO which are 6, and X44

and X49 which are 10.

Nash-Cournot production problem: See [22]. F is not twice continuously dif-
ferentiable. F is a P-function on the strictly positive orthant and since the solutions
obtained have all the components strictly positive, they are all R-regular.
Starting points for n= 5: (a) (36.931,41.8175,43.7060,42.6588,39.1786), (b) (10, 10, ... ,10),
(c) (1,1, ... ,1).
Starting points for n= 10: (a) (10,10, ... ,10), (b) (1,1, ... ,1).
Walrasian equilibrium problem: The Mathiesen example of a Walrasian equilib-
rium model. see [22]. F is not defined everywhere and does not belong to any known
class offunctions. Two set of parameters have been tried (see [22]): (1) (0.75,1,0.5) and
(2) (0.75,1,2). In both cases the solutions are not b-regular.
Starting points (a) (1,1,1,1).
HS34 problem: This problem represents the KKT conditions for the 34th problem
in [12]. The resulting F is monotone on the positive orthant but not even Po on JRn.
Starting points: (a) (0,1.05,2.9,0,0,0,0,0).
HS66 problem: This problem represents the KKT conditions for the 66th problem
in [12]. The resulting F is monotone on the positive orthant but not even Po on JRn.
Starting points: (a) (0,1.05,2.9,0,0,0,0,0).
Watson fourth problem: See [34]. This problem represents the KKT conditions for
a convex programming problem involving exponentials. The resulting F is monotone on
the positive orthant but not even Po on JRn.
Starting points: (a) (0,0, ... ,0), (b) (1,1, ... ,1), (c) (1,2,1,2,1), (d) (3,2,1,2,3).
Josephy problem: See [13]. F(x) is not a Po-function. The problem has a unique
solution which is not R-regular.
Starting points: (0,0,0,0), (b) (1,1,1,1), (c) (10 3 ,103 ,103 ,103 ,), (d) (1,0,1,0), (e)
(1,0,0,0), (f) (0,1,1,0).
Murty first problem: See [19]. Murty's first example of an LCP for which Lemke's
and Cottle and Dantzing's algorithms require an exponential number of steps. The
matrix defining F is a P-matrix.
Starting points: (a) (0,0, ... ,0).
Murty second problem: See [19]. Murty's second example of an LCP for which
Lemke's and Cottle and Dantzing's algorithms require an exponential number of steps.
82

The matrix defining F is a positive definite.

Starting point3: (a) (0,0, ... ,0).
Kanzow first problem: See [9). An LCP defined by a tridiagonal positive definite
matrix.
Starting points: (a) (0,0, ... ,0).
Kanzow second problem: See [9). An LCP defined by a diagonal positive definite
matrix whose condition number increases with the dimension.
Starting points: (a) (0,0, ... ,0).
Watson second problem: See [34). This is a linear complementarity problem defined
by a matrix which is strictly semimonotone, but neither P nor strictly copositive.
Starting points: (a) (0, ... ,0), (b) (1, ... ,1), (c) (106 , ••• , 106 ), (d) -(1,10,100,1000,10000).

HS35 problem: This problem represents the KKT conditions for the 35th problem
in [12). The resulting F is monotone but not strictly monotone.
Starting points: (a) (0.5,0.5,0.5,0).
HS76 problem: This problem represents the KKT conditions for the 76th problem
in [12). The resulting F is monotone but not strictly monotone.
Starting points: (a) (0.5,0.5,0.5,0.5,0,0,0).

REFERENCES
[1] G. Auchmuty. "Variational principles for variational inequalities". Numerical Functional Analysis
and Optimization, 10, pp. 863-874, 1989.
[2] F. Facchinei. "Minimization of SCI functions and the Maratos effect". DIS Technical Report 10.94.
To appear in Operations Research Letters.
[3] F. Facchinei and J. Soares. "A new merit function for nonlinear complementarity problems and a
related algorithm". DIS Technical Report 15.94, 1994.
[4] M. Ferris and S. Lucidi. "Globally convergent methods for nonlinear equations". Journal of Opti-
mization Theory and Applications, 81(1), 1994.
[5] A. Fischer. "A special Newton-type optimization method". Optimization, 24, pp. 269-284, 1992.
[6] A. Fischer. "A special Newton-type method for positive semidefinite linear complementarity prob-
lems". Technical Report, Institute of Numerical Mathematics, Dresden University of Technology,
Dresden, Germany, 1992. To appear in Journal of Optimization Theory and Applications.
[7] A. Fischer. "On the local superlinear convergence of a Newton-type method for LCP under weak
conditions". Technical Report MATH-NM-07-1993, Institute of Numerical Mathematics, Dresden
University of Technology, Dresden, Germany, 1993.
[8] M. Fukushima. "Equivalent differentiable optimization problems and descent methods for asymmet-
ric variational inequality problems". Mathematical Programming, Series A, 53, pp. 99-110, 1992.
[9] C. Geiger and C. Kanzow. "On the resolution of monotone complementarity problems". Preprint
82, Institute of Applied Mathematics, University of Hamburg, Hamburg, Germany, 1994.
[10] L. Grippo, F. Lampariello and S. Lucidi. "A nonmonotone line-search technique for Newton's
method". SIAM Journal on Numerical Analysis, 23, pp. 707-716, 1986.
[11] P.T. Harker and B. Xiao. "Newton's method for the nonlinear complementarity problem: A B-
differentiable equation approach". Mathematical Programming, Series A, 48, pp. 339-357, 1990.
[12] W. Hock and K. Schittkowski. "Test examples for nonlinear programming codes". Lectures Notes
in Economics and Mathematical Systems 187, Springer -Verlag, Berlin, 1981.
[13] N.H. Josephy. "Newton's methods for generalized equations". MRC Technical Summary Report
1965, Mathematics Research Center, University of Wisconsin-Madison, Wisconsin, 1979.
[l4] C. Kanzow. "Some equation-based methods for the nonlinear complementarity problem". Preprint
63, Institute of Applied Mathematics, University of Hamburg, Hamburg, Germany, 1993.
 83

[15] C. Kanzow. "Nonlinear complementarity as unconstrained optimization". Preprint 67, Institute of

Applied Mathematics, University of Hamburg, Hamburg, Germany, 1993.
[16] C. Kanzow. "Global convergence properties of some iterative methods for linear complementar-
ity problems". Preprint 72, Institute of Applied Mathematics, University of Hamburg, Hamburg,
Germany, 1993, revised 1994.
[17] O.L. Mangasarian. "Equivalence ofthe complementarity problem to a system of nonlinear equations".
SIAM Journal on Applied Mathematics, 31, pp. 89-92, 1976.
[18] O.L. Mangasarian and M.V. Solodov. "Nonlinear complementarity as unconstrained and constrained
minimization". Mathematical Programming, Series B, 62, pp. 277-297, 1993.
[19] K.G. Murty. "Linear complementarity, linear and nonlinear programming". Heldermann Verlag,
Berlin, 1988.
[20] R. Mifflin. "Semismooth and semiconvex functions in constrained optimization". SIAM Journal on
Control and Optimization, 15, pp. 957-972, 1977.
[21] J .-S. Pang. "Convergence of splitting and Newton methods for complementarity problems: an appli-
cation of some sensitivity results". Mathematical Programming, Series A, 58, pp. 149-160, 1993.
[22] J.-S. Pang and S.A. Gabriel. "NE/SQP: A robust algorithm for the nonlinear complementarity
problem". Mathematical Programming, Series A, 60, pp. 295-337, 1993.
[23] J .-S. Pang and L. Qi. "Nonsmooth equations: motivation and algorithms". SIAM Journal on
Optimization, 3, pp. 443-465, 1993.
[24] J.-S. Pang and L. Qi. "A globally convergent Newton methods for convex SCI minimization prob-
lems". To appear in Journal of Optimization Theory and Applications.
[25] M. Patriksson. "A unified framework of descent algorithms for nonlinear programs and variational
inequalities". PhD dissertation, Department of Mathematics, Linkoping University, Sweden, 1993.
[26] L. Qi. "A convergence analysis of some algorithms for solving nonsmooth equations". Mathematics
of Operations Research., 18, pp. 227-244, 1993.
[27] L. Qi and H. Jiang. "Karush-Kuhn-Tucker equations and convergence analysis of Newton methods
and Quasi-Newton methods for solving these equations". Technical report AMR 94/5, School of
Mathematics, University of New South Wales, Australia, 1994.
[28] L. Qi and J. Sun. "A nonsmooth version of Newton's methods". Mathematical Programming, Series
A, 58, pp. 353-368, 1993.
[29] S.M. Robinson. "Strongly regular generalized equations". Mathematics of Operations Research, 5,
pp. 43-62, 1980.
[30] S.M. Robinson. "Generalized equations". In "Mathematical programming: the state of the art",
A. Bachem, M. Groetschel and B. Korte editors, pp. 346-367, Springer-Verlag, Berlin, 1983.
[31] P.K. Subramanian. "Gauss-Newton methods for the complementarity problem". Journal of Opti-
mization Theory and Applications, 77, pp. 467-482, 1993.
[32] P.L. Toint. "An assessment of non-monotone linesearch techniques for unconstrained optimization".
Technical report 94/14, Department of Mathematics, Facultes Universitaires ND de la Paix, Namur,
Belgium, 1994.
[33] P. Tseng. "Growth behavior of a class of merit functions for the nonlinear complementarity problem" .
Manuscript, 1994.
[34] L.T. Watson. "Solving the nonlinear complementarity problem by a Homotopy method". SIAM
Journal on Control and Optimization, 17, pp. 36-46, 1979.
[35] N. Yamashita and M. Fukushima. "On stationary points of the implicit Lagrangian for nonlinear
complementarity problems". Information Science Technical Report, Nara Institute of Science and
Technology, Nara, Japan, 1993. To appear in Journal of Optimization Theory and Applications.
EQUILIBRIUM IN TRANSPORT NETWORKS WITH
CAPACITY CONSTRAINTS

Ferrari Paolo

Dipartimento di Costruzioni Stradali e Trasporti

Universita di Pisa
Via Diotisalvi 2 - 56126 Pisa, Italy
Fax: ++3950555057

Abstract. This paper considers the problem of multimodal transport networks with
elastic demand and capacity constraints. It is shown that, when capacity constraints
are active, there is not equilibrium in these networks. However an equilibrium solution
can be assured in any case if additional costs are imposed on network links. The paper
presents a model of multimodal networks where capacity constraints, and both demand
and cost functions, are symmetric: in this case the additional costs are the Lagrange
multipliers of capacity constraints. A method to calculate these additional costs is
proposed and is used in a simple numerical example.

Key Words: Transport network equilibrium, Elastic demand, Capacity constraints,

Lagrange multipliers.

1. INTRODUCTION

The hypothesis that capacity constraints are not binding is very often used in the
study of transport networks. In this hypothesis an equilibrium solution always exists
if demand is fixed, whereas some restrictions have to be imposed on both demand and
cost functions in order to assure the existence of equilibrium when demand is elastic.
But road capacity constraints are often active in real life: when this happens
there is not equilibrium in transport networks. However the existence of a network
equilibrium can be assured also when capacity constraints are active, if additional
costs are imposed on network links.
This paper, after a discussion about the problem of equilibrium in transport net-
works with capacity constraints, considers a model of multimodal transport network
where capacity constraints, and both demand and cost functions, are symmetric. It is
shown that in this case the additional costs that assure the existence of equilibrium

85
86
are the Lagrange multipliers of capacity constraints. A method to calculate this "road
pricing" is proposed and is used in a simple numerical example.

2. THE CONCEPT OF EQUILIBRIUM

Consider a multimodal urban transport network, represented by a graph G(N, L),

where N is the set of nodes and L the set of directed links, each of which is used by
only one transport mode. Each link i has a cost function Ci(J) associated to it, i.e. a
relationship between the average Ci of costs perceived by users travelling along the link
and the flow vector f on all network links. Link flows have to satisfy some capacity
constraints 9 r (J) ::; o.
Some nodes of the network are origin/ destination (0 /D) centroids. At a certain
time let Nk be the generation capacity of the centroid k, i.e. the number of the
individuals that are in k and that can decide to move from it. The pattern of the
network utilization, i.e. the demand between the different pairs of centroids and its
distribution among the various transport modes and the various paths, is a consequence
of choices made by the individuals who belong to the generation capacity of the various
centroids.
Consider an individual who belongs to Nk and is characterized by a vector Sk of
socioeconomic attributes. He chooses the destination j of his trip (j could coincide
with k, in this case he would decide not to move at the time we are considering),
the transport mode and the path he will use to go to j. So each choice alternative is
defined by a destination, a transport mode and a path, and is characterized by a vector
of attributes which represent the attractiveness of the destination, the comfort and the
charges of the mode, and the path characteristics, i.e. journey time, fuel consumption,
etc. A vector of attributes completely defines an alternative, so that when an individual
makes a choice, he actually draws a vector of attributes from the choice set Bk, which
is in general different from an individual to another (McFadden, 1975).
The individual associates an utility value to each alternative. This utility is a
function U(X,Sk) of the vector x of the alternative attributes and of the vector Sk of
socioeconomic attributes of the individual. He draws from the set Bk the vector Xj
which maximizes his utility:

(2.1)

The utility associated to a vector x E Bk varies for random reasons from an

individual to another even if they have the same vector Sk, and it is different also for
the same individual in different time periods: so it is a random variable. Consider a
sequence Tl, T2 . .. Tn . .. of epochs having the same joint probability distribution function
of U, x, Sk. These epochs can be found out easily: e.g. they are the time periods when
people go to work early in the morning, or when they go home in the evening, and so
on. It can be noted that this joint distribution is a conditional distribution, given the
values that the attributes of alternatives assume in epochs which are different from
 87

those of the sequence Tl, T2 ... T n ... : e.g. the probability that an individual receives a
given utility from making or not a trip in a given day period depends on the costs of
this trip in other day periods.
An individual characterized by the vector Sk associates to the vector x E Bk in
any epoch of the sequence Tl . .. Tn an utility given by:

(2.2)

where Vex, Sk) is the regression of U on x and Sk, and p is a random variable with zero
mean. If p is a Wei bull random variable and if the utilities associated to the various
alternatives are independent, the probability Pkj that the individual chooses the vector
Xj is given by the logit model (McFadden, 1975):

(2.3)

Let y, v, w be the vectors of attributes of the destination, of the mode and of the
path into which the vector x can be partitioned:

(2.4)

We assume that the utility regression V has the following additively separable
form (Domencich and McFadden, 1975):

(2.5)

It follows from Eq. (2.5) that an individual belonging to N k , who has chosen to go
to the destination j using the t mode, travels along the path from which he expects the
maximum utility, i.e. the minimum cost: thus this minimum cost becomes an attribute
of the transport mode chosen to go from k to j.
Assume that all individuals belonging to Nk have the same socioeconomic char-
acteristics. In this case we have:

(2.6)

where aj is the component of f) due to the attractiveness attributes of the centroid

j, AL is the average of the costs expected by users who want to travel on mode t
between k and j along the path considered the most convenient by each of them, and
f3 is a scale factor (the sign minus on the right side of Eq. (2.6) is due to the fact that
the utility deriving from the choice of the transport mode is negative, whereas both f3
and AL are considered positive).
Let Tkj be the set of transport modes connecting k and j. By substituting the
expression (2.6) into (2.3), and taking into account that Atk = 0, we have the following
88

expression of the probability P~j that an individual chooses to go from k to j using

mode t:
t exp[aj - .B.\~il
(2.7)
Pl<j = exp[akl + E E exp[a/ - .B.\~/I·
/#l<tETkl

If the choices of the individuals belonging to NI< are independent, the trips made
in the epoch under examination from k to the other centroids are the components of
a multinomial random vector. Thus the average Dt j of demand from k to j on mode
t in this epoch is given by:

(2.8)

and the average of the number of individuals that remain in k is:

D -N -
I< - I<
E E Dt
/ .... 1< tET
-N
1</ - I< exp[al<l +E E
exp[al<l
exp[a/ - .B.\~,I·
(2.9)
... kl 1#1< tETkl

If aj and .\~j Vj, t are constant during the sequence Tl •• •T n ••• , the same happens
for the probability given by Eq. (2.7); in this case the succession of the demand values
during the sequence is the realization of a stationary process whose average is given by
Eqs. (2.8) and (2.9). Moreover a multinomial random variable converges in probability
to its average; this means that if N" is large, the probability of a considerable shift
of demand from its average is negligible. For this reason we can consider the demand
value constant during all the sequence: Eqs. (2.8) and (2.9) give the demand function,
and its inverse, i.e. the vector of the averages .\~j Vk,j, t of the expected costs that
yield the demand, is obtained from Eqs. (2.8) and (2.9):

.\~j = ~[aj - a" -lnDtj + In(N,, - E E D~,)I Vk,j, t (2.10)

1#" tETkl

where we take lnD~j = 0 if DL :5 1 in order to assure the continuity of AL(D).

Until the socioeconomic characteristics and the attractiveness attributes of the
various zones of a town do not vary, the aj values can be considered constant in the
successive epochs of the sequence. The same happens for .\~j if the link flows are low,
because in this case the costs on network links are independent of the flows. On the
other hand, if flows are not low, costs depend on them, and it is possible that .\L
are not constant. As a matter of fact the individuals that belong to the generation
capacity of the various centroids make their choices at a certain epoch on the basis of
the expected trip costs whose averages are .\tj . These choices give rise to the demand
between the various 0 jD pairs and on the various transport modes, and to flows on
network links on which the costs the individuals bear during their trips depend. If
these costs are different from those the individuals expected, in the successive epoch
they will change their choices, giving rise to a new demand, new link flows and new
costs on links, from which new trip costs derive, and so on. It is possible that a certain
 89

epoch such a situation is reached that the trip costs the individuals bear coincide or
are very close to those they expected: in this case the individuals will repeat their
choices in the successive epochs, giving rise to constant values of >"L
and of demand.
However this happens only if an equilibrium solution for the transport network exists.

3. THE EQUILIBRIUM SOLUTION

Let (' indicates the transpose of a vector or of a matrix):

w = (k, j), k =f. j = an ordered pair of centroids

W = the set of the ordered w pairs

n= the number of pairs w E W

S = the set of centroids

Tw = the set of transport modes joining the pair w EW

T = U Tw = the set of all transport modes of the network

wEW

u = the number of elements in T

rw = the number of all transport modes joining w E W

d!v = the number of individuals who travel between i,j = w on mode t during an
unit time period within the epoch T under examination: if we assume that the
demand D!., given by Eq. (2.8) is uniformly distributed during T, d!v is obtained
dividing D!v by the duration of T.

d! = (.. .d~ ... )' =

the vector of demand during the unit time period on mode t
between all w E W which are connected by mode t

d = (.. .d t ' ••• )'

the vector of demand during the unit time period on all modes
-

t E T between all w E W. Hence forward we will omit, for the sake of brevity, the
words "during the unit of time" when we will refer to vector d and its components.

P! = the set of paths p joining w on mode t

Pw = U P! = the set of all paths p joining w

tETw

pt = U P! = the set of all paths p joining all w E W on mode t E T

wEW

p = U p w = the set of all paths p joining all w E W on all modes t E T

wEW
 90

Mt = the number of paths p E pt

M = the number of paths pEP

hp = the flow on paths pEP

h t E R~' =the vector of flows on all paths p E pi

h = ( ... h t' ... )' E R~ = the vector of flows on all paths pEP

At = the incidence matrix link-paths travelled on mode t

p = At h t = the vector of link flows generated by mode t

f = ( .. .p' ... )' = the vector of link flows generated by all modes.
Given:

1
we have f = Ah.
Let Bt be the incidence matrix between pairs wand paths travelled on mode t,
and

1
we have:
Bh - d = O. (3.1)

Suppose that the network has v capacity constraints:

gr(f)::;O jE(1.2 ... v)

where gr(f) are convex functions, in general linear (Ruberti, 1994).

Any vector [~] ~ 0 which satisfies Eq. (3.1) is called solution vector. The set:

is the supply set. The set:

is the demand set. Both 0 1 and O 2 are compact and convex.

The set 0 = 0 1 n02 is the set of feasible solutions: generally in transport networks,
whose study is of actual interest, 0 1 C02 and thus 0 = 0 1 . The set of feasible link
flow vectors is:
 91

Let:

Ci(f) = the average of trip costs perceived by users who travel on link i E L; Ci(f) is
continuous function of f in 8.

C p( h) = L Ci(f)l!ip = the average of trip costs perceived by users who travel on path
iEL
pEP, where l!ip = 1 if p traverses i, 0 otherwise; Cp( h) is continuous function
of h in f!.

Alj = the inverse of the demand function, obtained by substituting in Eq. (2.10)
D!j = dljT

At = ( .. . A:" ... ),

A = ( .. . At ' .. . )'.
A vector [~] E f! is an equilibrium solution if and only if, for every pair w, every

mode t E T w , and every path pEP! (Dafermos, 1982):

hp > 0 =} Cp(h) - A:"(£1) =0

(3.2)
hp = 0 =} Cp(h) - A:"(£1) 2: o.
It can be shown that, if capacity constraints are written as strict inequalities,
conditions (3.2) are equivalent to the following Variational Inequality:

(3.3)

where hp and d:" are the components of any vector [~] E f!.
The following demonstration of equivalence is a modification of that reported in
(Dafermos, 1982) in order to take into account the capacity constraints.
We note that (3.2) =} (3.3). Indeed, given a vector [~] E f!, (3.2) implies:

(3.4)

for every pair w, every mode t E Tw and every path pEP!. Inequality (3.4) follows
from the fact that, when hp > 0 for pEP!, we have Cp(h) - A:"(£1) = 0, so that (3.4)
is satisfied. When hp = 0 it follows that Cp(h) - A:" (d) 2: 0, and (3.4) is satisfied again
because hp 2: O. By summing (3.4) over all pEP!, the following relation holds:

L [Cp(h). (hp - hp)] - A:"(£1). (d~ - d~) 2: 0 (3.5)

pEP:'

and summing (3.5) for all t and all w we obtain (3.3).

Conversely assume that, for a certain [~] E f!, (3.3) is satisfied for every [~] E

f!. We show that [~] satisfies the equilibrium conditions (3.2). Indeed, for a fixed
92

pair Wn E Wand a mode tk E Twn , and given a path Pk E P!':., we consider a vector
[J] E 11 such that hp = hp Vp ::f. Pk, but hpk ::f. hpk ' As B h - d = Bh - d = 0, it
follows that d~ = d~ if (t, w) ::f. (tk' w n) and d!:;n = d!:;n - hpk + hpk ' In this case (3.3)
reduces to:
CpkCh). (h pk - hpk ) - A!:;Jd). (hpk - hpk ) ~ o. (3.6)
If hpk = 0, as hpk ~ 0 (3.6) implies:

Cpk(h) - A!:;Jd) ~ o. (3.7)

We suppose, as it was said before, that 111 C112 and capacity constraints are written
as strict inequalities. So, if hpk > 0 we can choose hpk ~hpk' and (3.6) implies:
<
(3.8)

As (3.7) and (3.8) hold for every w, t and p, we obtain (3.2). It is worth noting that
the condition that capacity constraints are strict inequalities, so that we can always
choose hpk > hpk , is essential in order to demonstrate the equivalence between (3.2)
and (3.3). On the other hand it is easy to verify that if capacity constraints are written
as :::;, the user equilibrium problem (3.2) and the Variational Inequality (3.3) are not
equivalent, because the existence of a solution of (3.3) would not imply the same for the
problem (3.2). Consider for example the simple netwotk represented in fig. 1, where a
pair of nodes (WI, W2) is connected by a link characterized by the cost function c(f).
The link flow is equal to the demand between the two nodes; let A( d) be the inverse
of demand function, and let H be the link capacity. No equilibrium solution exists,
because for any demand value in the closed interval [0, HI the link cost is less than A( d),
so that the equilibrium conditions (3.2) are not verified. If the capacity constraint

c, }..

W1 W2

0>--------..-<0

o f~ f H f,d

Figure 1 - An example illustrating that the solution H of the Variational Inequality

is not equilibrium solution.

were written as f :::; H, the interval of feasible solution [0, H] would be closed, and the
point f =H would be a solution of the Variational Inequality (3.3); in fact we would
 93

have, for any I E [0, H] and =I- H, I < H so that:

[e(H) - A(H)] . U - H) > ° (3.9)

whereas H is not an equilibrium solution. On the other hand, if the capacity constraints
is I < H, the point I = H is not a feasible solution, and no point r E [0, H) is a
r
solution of the Variational Inequality: in fact a point I > II E [0, H) always exists,
so that:
[e(r) - AU·)]' U - r) < ° (3.10)

IT capacity constraints are strict inequalities, 0 is not closed, so that the Varia-
tional Inequality (3.3) does not always admit a solution (Kinderleher and Stampacchia,
1980, pp. 13-14). IT (3.3) has no solution, the transport network has no equilibrium
point belonging to 0: this means, from a practical point of view, that it has no equi-
librium solution at all.

4. A MODIFICATION OF COST FUNCTIONS THAT ASSURES THE

EXISTENCE OF EQUILIBRIUM

It has been shown (Ferrari, 1994) that it is always possible to obtain an equilibrium
solution for a network with capacity constraints if one modifies in a suitable way the
link cost functions. In this section we illustrate this procedure when cost functions,
the inverse of demand functions and capacity constraints are symmetric. This means
that a link cost depends only on the flow travelling on this link:

ejU) = ejU;) Vi E L (4.1)

the cost A!., depends only on the demand d!.,:

A!.,(d) = A!.,(d!.,) Vt E Tw , Vw E W (4.2)

and only one link flows appears in each capacity constraint:

(4.3)

where Hi is the capacity of link j, and I is the set of links whose capacity constraints
are considered in the network. The case of symmetry is interesting because, among
other things, the modification of link cost functions assumes a particular meaning.
Let X be the open set of vectors hERM which satisfy the constraints defined by
the generation capacity of centroids:

(4.4)
 94

Consider the following problem of nonlinear programming:

min[R(h):L,hpDjp~Hj
pEP
VjEI, h2:0, hEX] (4.5)

where
(E h,6;, f E h.6!...
R(h) = L, }O·EP ci(x)dx - L, }o·EP >'~(y)dy (4.6)
iEL 0 fET", 0
"'EW

and D~p = 1 if P connects the w pair on mode t, 0 otherwise.

The set defined by constraints which appear in the problem (4.5) is contained in
X, as it has been said in the previous section; thus it is compact. So the problem (4.5)
has at least a solution h, which is a Kuhn-Tucker (KT) point, because the constraints
are linear.
The KT condition in h is written:

VR(h) + L,ujVgj(h) + L,zpVsP(h) =0 (4.7)

jEi pEP
wheregj(h) = E hpDjp-Hj ~ OVj E I are the capacity constraints, sP(h) = -hp ~ 0
pEP
Vp E P are the non-negativity constraints, U j 2: 0 and zp 2: 0 are the corresponding
Lagrange multipliers.
The Pk-components of Eq. (4.7), Pk E P!, is:

where w is the only pair connected on mode t by path Pk.

As we have:

Eq. (4.8) becomes:

(4.9)

If Uj = 0 Vj E I, Eq. (4.9) written for all Pk E P coincides with the definition

of equilibrium (3.2). As a matter of fact, if hp. > 0 we have ZPk =
0 and so Cpk(h) -
>'~(d~) = 0; whereas if hpk = 0 we have ZPk 2: 0 and so Cpk(h) - >'~(d~) 2: O. Thus,
if capacity constraints are not binding in a KT point, this is an equilibrium solution;
and if no KT point has the multiplier vector U = 0, no equilibrium solution exists.
Suppose that one increases the link cost function Ci(fi) by adding a road-pricing
equal to the corresponding multiplier Ui that appears in Eq. (4.9). Let

crUd = CiUi) Vi ~ I
(4.10)
c';(f;) = Ci(f;) + Ui Vi E I
 95

be the new cost functions. The modified path cost C;.(h), Pk E P!, is:
C;.(h) = 'L>:Cf;)D;Pk = L Ci(/i)Dip> + L UjDiP. = Cpo(h) + L UjDip.' (4.11)
iEL iEL iEI iEI
By substituting the expression (4.11) into Eq. (4.9), the latter becomes:

(4.12)

and coincides with the equilibrium definition (3.2). So, by using the cost functions
(4.10), the point h becomes an equilibrium solution.
The traditional method used to calculate an equilibrium solution solves the prob-
lem (4.5) neglecting the capacity constraints. Thus the problem solved is:

min[R(h) : h ~ 0, hE Xl. (4.13)

If one obtains a solution that satisfies the capacity constraints, it is an equilibrium

solution. If no solution so calculated satisfies the capacity constraints, no equilibrium
solution exists.
We propose a different approach, which produces at the same time both the equi-
librium solution and the values of road pricing. It solves the problem (4.13) by using
new cost functions Cj(fi) on all links i E I:

Ci(fi) = Ci(fi) if fi = L hpDip ::; Hi

pEP (4.14)
Ci(fi) = Ci(f;) + ~(fi
f
- Hi), f > 0 if fi > Hi.
If one uses the cost functions (4.14), the expression of the minimandofthe problem
(4.5) becomes:
R(h) = R(h) + etCh) (4.15)
where
2

r " h 6· I: hpDip - Hi
pEP
[ ]

etCh) = L",i }/EP

L.. p.p 1
(x - Hi)-dx = L ",i-=------"-- ( 4.16)
iEI Hi € iEI 2€

with "'i = 0 if pEP

L: hpDip - Hi ::; 0 and "'i = 1 if L:
pEP
hpDip - Hi > 0, is a differentiable
penalty function. It can be shown (Bazaraa, Sheraly and Shetty, 1993, pp. 366-368)
that when € -+ 0 the solution of the problem:

min[R(h) : h ~ 0, hE Xl (4.17)

tends to the solution h ofthe problem (4.5). If h, is the solution of the problem (4.17)
for a particular value of €, the KT condition for f sufficiently small can be written:

"V R(h,) + L u,S [L(h'PDi P) - Hi] + L z'P "V sP(h,) =0 (4.18)

iEJ pEP pEP
96
E
where Uti = pEP <
h!p 6 ip-Hi
, J is the set of links whose capacity constraints are active

in hand Z<p are the Lagrange multipliers of the non-negativity constraints sP( h) :::; o.
The Pk-component of (4.18) is:

Gpk(h.) - A~(d!w) +L U.ibipk - Z'Pk =0 (4.19)

iEJ

where d!w = L: h.pb~p. When f ..... 0 h • ..... h. By comparing Eq. (4.9) with
pEP
Eq. (4.19) when f ..... 0, and taking into account that there exists unique Lagrange
multipliers Ui ~ 0, ZPk ~ 0 that solve Eq. (4.9), we have that the lim U.i = Uj Vi E I .
• -0
Thus the solution of the problem (4.17) for f ..... 0 gives the solution h of the
problem (4.5) and at the same time the values of Lagrange multipliers Ui. If Ui = 0
Vi E I, h is an equilibrium solution for the network with the original cost functions. If
the multipliers Ui are> 0 for some links i E I, they represent the "road pricing" that
has to be imposed on these links in order to transform h into an equilibrium solution.
The solution of the problem (4.17) can be made arbitrarily close to the solution
of the problem (4.5) by choosing f sufficiently small. However high computational
difficulties can arise if we use a very small f value. For this reasons the popular
approach to the solution of problems that use penalty functions employs a sequence of
decreasing parameters f. With each new value of € a problem (4.17) is solved, starting
with the solution corresponding to the previously chosen parameter value.

5. A COMPUTATIONAL EXAMPLE

The method illustrated in the previous section has been applied to the small
network reported in fig. 2, which is travelled by two transport modes: car and transit.
Nodes 1. .. 5 are centroids: DID flows depart from them and arrive at them, but cannot
go through them. Every pair of centroids is joined on transit by only one link, without
need of passing through other nodes; every centroid is connected by a dummy link to
car links. Car and transit do not interfere with each other. Transport cost coincides
with time.

The journey time on transit is independent of flow, and is given by the sum of
line-haul time t~ reported in Tab. 1 for every pair w, and of access and waiting time ta
which is equal to 10 minutes for all pairs w. The time on dummy links is 0.1 minutes;
the time of every car link depends only on the flow Ii on it and is given by:

t; = ai + bi [ 10~O
r ]4 (5.1)

where parameters ai and bi are reported in Tab. 2 for every link i. Flows on both
car and transit links are measured in passengers per hour (Pph). Capacity constraints
express the condition that flow Ii on every car link has to be less than the EGi values
reported in Tab. 2.
 97

(0:. 6- - - - - - - 7-------8~:Pl
1 2

I I I
I I I
I I I
I I I

o Centroids
I
I
I
I
I
L ______ -Q-
3
________
I
I
...J
o Nodes I 9 J J •••• 10 11

- Car links I
I
r------ JL - - - - - - - - - . . ,I
- - Transit links I I I
I I
........ Dummy links I I
I I
I12 ) 13 14 I
.I .~l
41~ - - - - - - - - - - - -
- - - ~'" 5
L _________________ J
~-----------------
Figure 2 Bimodal network considered in the computational example.

Tab. 1 - Transit line-haul times (min) between the OlD of the network in fig. 2

OlD 1 2 3 4 5

1 - 29 12 28 27

2 29 - 17 32 32

3 12 17 - 16 15

4 28 32 16 - 31

5 27 32 15 31 -
98

Tab. 2 - Parameters of cost functions and capacities ECi (Pph) on car links of the
network in fig. 2

Link ai bi ECi Link ai bi ECi

6-7 6.00 7.26 1200.0 10-11 10.00 3.00 1000.0

6-9 4.00 2.94 4000.0 10-13 7.00 1.33 1400.0
7-6 6.00 7.26 1200.0 11-8 7.00 2.10 1300.0
7-8 8.00 7.04 1300.0 11-10 10.00 3.00 1000.0
7-10 6.00 2.94 1200.0 11-14 8.00 1.92 1400.0
8-7 8.00 7.04 1300.0 12-9 6.00 1.44 1300.0
8-11 7.00 2.10 1300.0 12-13 9.00 5.94 1200.0
9-6 4.00 2.94 4000.0 13-10 7.00 1.33 1400.0
9-10 9.00 3.42 1300.0 13-12 9.00 5.94 1200.0
9-12 6.00 1.44 1300.0 13-14 8.00 2.40 1200.0
10-7 6.00 2.94 1200.0 14-11 8.00 1.92 1400.0
10-9 9.00 3.42 1300.0 14-13 8.00 2.40 1200.0

The total transport demand d w on both modes for every pair w is fixed and is
given in Tab. 3. The demand for car travel d~ is obtained from Eq. (2.8), which in
the case under examination becomes:
de - d exp[-,Bt~]
(5.2)
w - Wexp[-,Bt:'u] + exp[-'Yta - ,Btt,]
where t:;, is the journey time by car for pair w, and coefficients ,B and 'Y are equal to
0.1 and 0.15 respectively.

Tab. 3 - Transport demand (Pph) between the OlD of the network in fig. 2

OlD 1 2 3 4 5

1 - 2500 2300 2500 2700

2 3000 - 3450 3750 4050

3 2300 1840 - 2000 2160

4 2400 2760 3360 - 3240

5 2400 2070 2520 2160 -

 99

Tab. 4 - Flows Ii (Pph) and estimates uE;(min) of Lagrange multipliers for car links
of the network in fig. 2 corresponding to various value of parameter f

Link f=100.0 f=6.25 f=0.195 f=0.0061 f=0.00019

i Ii Ufi Ii Ufi Ii UEi Ii Ufj Ii UEi

6-7 1286.75 0.867 1219.34 3.094 1201.19 6.071 1200.03 4.480 1200.00 4.748

6-9 1614.38 0.000 1569.93 0.000 1559.46 0.000 1559.02 0.000 1559.00 0.000

7-6 1278.77 0.788 1210.66 1.705 1200.37 1.916 1200.01 1.489 1200.00 1.904

7-8 1281.33 0.000 1276.26 0.000 1271.16 0.000 1270.46 0.000 1270.45 0.000

7-10 1128.86 0.000 1076.05 0.000 1041.43 0.000 1040.59 0.000 1040.57 0.000

8-7 1367.32 0.673 1331.28 5.004 1301.40 7.149 1300.05 7.487 1300.00 9.004

8-11 1743.15 4.431 1405.87 16.939 1304.12 21.118 1300.12 20.295 1300.00 21.221

9-6 1585.67 0.000 1527.32 0.000 1502.28 0.000 1501.63 0.000 1501.59 0.000

9-10 1090.09 0.000 1038.43 0.000 1020.03 0.000 1019.49 0.000 1019.49 0.000

9-12 1706.88 4.069 1392.62 14.819 1303.16 16.169 1300.10 16.542 1300.00 17.832

10-7 1034.89 0.000 1012.35 0.000 1010.38 0.000 1010.99 0.000 1011.02 0.000

10-9 1001.04 0.000 980.39 0.000 961.64 0.000 962.07 0.000 962.07 0.000

10-11 856.80 0.000 842.98 0.000 840.69 0.000 840.85 0.000 840.87 0.000

10-13 1136.87 0.000 1101.58 0.000 1082.63 0.000 1082.25 0.000 1082.23 0.000

11-8 1606.53 3.065 1360.34 9.654 1302.15 10.994 1300.07 11.613 1300.00 11.925

11-10 940.32 0.000 887.05 0.000 842.54 0.000 840.91 0.000 840.87 0.000

11-14 1639.52 2.395 1451.57 8.252 1402.03 10.372 1400.06 9.177 1400.00 9.934

12-9 1767.21 4.672 1408.05 17.288 1304.38 22.410 1300.13 21.654 1300.00 21.534

12-13 1356.61 1.566 1253.00 8.481 1202.14 10.962 1200.07 11.644 1200.00 12.883

13-10 1181.94 0.000 1114.61 0.000 1083.45 0.000 1082.29 0.000 1082.23 0.000

13-12 1276.94 0.769 1229.90 4.784 1201.28 6.539 1200.04 5.766 1200.00 6.263

13-14 1571.71 3.717 1280.95 12.952 1202.86 14.658 1200.08 13.803 1200.00 13.504

14-11 1586.41 1.864 1450.11 8.018 1401.91 9.756 1400.06 9.697 1400.00 9.415

14-13 1537.11 3.371 1270.88 11.340 1202.82 14.432 1200.09 14.907 1200.00 14.759
100

The problem under examination is equivalent to that obtained if one considers

only the car network with elastic demand given by (5.2): flow on the transit link
between the pair w is d~ = dw - d~. In this case the demand d can be assigned to
the bimodal network by the usual Frank-Wolfe algorithm, if one attributes to every
transit link j the following dummy cost function (Sheffi, 1985, pp. 155-157):

ti = 731[1n dwd~
_ dt, + fib
tw + ,ta ] . (5.3)

A cost function ii(fi) modified following (4.14) was attributed to every car link:

i- =
I
a'
I
+ b· [~]4
I 1000 if f,.I > EG·I
(5.4)
.
t·I = a'I + b·
I
h]4 + -(f,.
[-1000 1
f
- EG·)
I I ,
f > 0 if h ~ EG;.

The Frank-Wolfe assignment procedure was repeated 20 times, starting with f =

100 and taking in each successive assignment half the f value of the preceding one. The
equilibrium flow vector obtained in each assignment was assumed as stRl-ting point in
the successive one.
The flows Ifi on car links and the estimates U.i of Lagrange multipliers obtained
for different values of f are reported in tab. 4. It can be noted the progressive approach
of link flows If; to capacities EGi as f approaches O.

REFERENCES

[1) S. Bazaraa, D. Sheraly and C.M. Shetty, "Nonlinear Programming: Theory and Algorithms",
John Wiley & Sons, New York, N.Y. ,1993.
[2) S. Dafermos, "The general multimodal network equilibrium problem with elastic demand", Net-
worL 12, 57-72, 1982.
[3) T.A. Domencich and D. McFadden, "Urban Travel Demand", North Holland, Amsterdam, 1975.
[4) P. Ferrari, "Road pricing and network equilibrium" , forthcoming in Transpn. Res. B 1994.
[5) D. Kinderleher and G. Stampacchia, "An Introduction to Variational Inequalities and Their
Applications" , Academic Press, New York, 1980.
[6) D. McFadden, "Conditionallogit analysis of qualitative choice behaviour", Frontiers in Econo-
metrics, Academic Press, New York, NY, 105-142, 1975.
[7) G. Ruberti, "I vincoli di capacita delle reti stradali urbane" "Internal Report of Istituto di Strade
e Trasporti dell'Universita di Pisa", 1994.
[8) Y. Sheffy, "Urban Transportation Networks: Equilibrium Analysis with Mathematical Program
ming Methods" , Prentice-Hall, Englewood Cliffs, N.J., 1985.
SEPARATION OF SETS AND GAP FUNCTIONS FOR
QUASI-VARIATIONAL INEQUALITIES

Giannessi Franco

Department of Mathematics
University of Pisa
Via F. Buonarroti 2, 56127 Pisa, Italy
e-mail: gianness@gauss.dm.unipi.it

Abstract. A Quasi-Variational Inequality, whose domain is intersection between a

set and the level set of a function having finite dimensional image, is embedded into a
separation scheme. Then, it is shown that a gap function can be viewed as a by-product
of the separation scheme. Connections with duality and some further investigations
are discussed.

Key Words. Quasi-Variational Inequality, separation of sets, gap function, duality,

network equilibrium

1. INTRODUCTION

We deal with Quasi-Variational Inequalities, whose domain can be expressed as

intersection between a set and the level set of a function in a real Hilbert space and
with finite dimensional image. The existence of solutions is expressed as disjunction
of two sets, and then the separation theory is exploited. Such an investigation allows
us, among other things, to associate, to a Quasi-Variational Inequality, two families
of optimization problems (weak and strong), even in cases where classic assumptions
of monotonicity type for the operator and convexity of the domain are not fulfilled.
Thus, it is natural to think of exploiting such families in order to "force" the Quasi-
Variational Inequality to enjoy some desired properties. This is shown in detail by
the example of Sect.2 for a Variational Inequality. Sect.4 considers the particular case
of a Variational Inequality having a polyhedral domain; it is shown how the analysis
developed in the previous sections can lead to new algorithms, that take advantage
of the tools of Nonconvex Optimization. Sect.5 shows also that the above mentioned
analysis can help us to deepen other topics, like duality, which nowadays requires
further investigation. Sect.6 shows that the study of Variational and Quasi-Variational

101
102

Inequalities may give a contribution to a better definition of the equilibrium of flows

on a network.

2. GAP FUNCTIONS FOR QUASI-VARIATIONAL INEQUALITIES.

THE WEAK CASE

Let 3 be a real Hilbert space, F: 3 -+ 3, X: 3=*3, 9 : X(y) x X(x) -+ Rm;

consider a Quasi-Variational Inequality (in short, QVI) in the form: to find y E K(y),
such that

(F(y), x - y) ~ 0, Vx E K(y) := {x E X(y) : g(y; x) E C} =1= 0, (2.1)

where C is a closed and convex cone with apex at the origin. This form can be in-
terpreted as the search, among the fixed-points of the point-to-set map K, for one
which satisfies (2.1); it obviously includes a classic form of QVI, when Vy E X(y) the
condition g(y;x) E C is identically true (with respect to x) on X(y); the fact that the
format (2.1) embraces such kinds of 9 will be exploited later to recover known gap
functions.
When X(y) and g(y; x) are independent of y (in this case they will be denoted by
X and g(x), respectively), then (2.1) collapses to a Variational Inequality, which will
be denoted by VI. If, moreover, K(y) == X(y) and, Vy,K(y) is a closed and convex
cone with apex at the origin, then (2.1) collapses to a Quasi-Complementarity System
(in a Hilbert space):

YE K(y) , F(y) E K*(y) , (F(y), y) = 0, (2.1),

where K*(y) denotes the (positive) polar of K(y). When K(y) is independent of y,
then (2.1)' becomes the Classical Complementarity System.
In [8J, Sect. 3, it has been proposed to associate an image space to a VI; this is
now done for a QVI. Such an approach will lead to define a wide class of gap functions,
containing the known ones. The gap functions we will obtain have a Lagrangian taste,
in the sense that they allow us to free ourselves from the constraints g(y; x) E C in
defining the gap function. The approach proposed in [8J starts with the obvious remark
that y E K(y) is a solution of (2.1) iff the system (in the unknown x):

u:= (F(y),y - x) > 0, v:= g(y;x) E C , x E X(y), (2.2)

is infeasible. The space where (u, v) runs is the image space associated to (2.1), and
the set:

K(y) := {(u, v) E R x R m : u = (F(y), y - x), v = g(y; x), x E X(y)}

is the image of (2.1). To system (2.2) we associate the set 1{ := {(u, v) E R x R m : u >
0, v E C}, which depends on the types of inequalities only; another obvious remark is
 103

that the impossibility of (2.2) is equivalent to 'Ii n K(y) = 0. To show this disjunction
in [8] a separation scheme is proposed; this approach has been developed in [9]. It will
now be slightly modified and applied to (2.1) in order to define a general class of gap
functions; concepts and notations of [9] will be here understood.
In the remaining part of this section we will develop the particular case where

1:= {l, ... ,m}, g(y;X) = (gi(Y;X), iEI),

Op:= (0, ... ,0) E RP C = Op x R~-P ,

where p,m E ~+ with 0 ~ p ~ m, and C = R~ or C = Om according to p = 0 or

p = m, respectively. Consider the function:

w(y;u,v;,\,W):=u+{,\,G(y;v;W») ,uER, vERm, '\EC·, wEQ, (2.3)

where C· := {,\ E R m : '\i ~ 0, i = p + 1, ... , m} is the (positive) polar of C, and

G(y;v;W):= (Gi(y;Vi;Wi), i E I) ,
m
Gi : :::: x R xQi -+ R, W = (Wi, i E I), Wi E Qi , Q = X Qi;
i=1

G and Q must be such that, Vy E X(y),

lev>ow :::> 'Ii ; n

"'Ell
lev>ow = cl'li. (2.4)

where the level sets are considered with respect to (u, v) only. Under these conditions
(2.3) is a weak separation function in the sense of [9].
Each G i may be considered as a transformation of gi; for this reason,
Vy E X(y), VWi E Qi, Gi must be such that

(Gi(y;gi(y;X);Wi), i E I) E C <==> (gi(Y;X), i E I) E C. (2.5)

In the case of a VI, if p = 0, so that C = R~, examples of Gi (considered also in [7],

p. 352) are:
G(y;Vi;Wi) = Viexp(-WiVi), Wi E Qi:= R+, (2.3)'

G(y; Vi; Wi) = 1 - exp( -WiVi), Wi E Q i := R+ . (2.3)"

Remark 2.1. According to [9), (2.3) is a particular case of weak separation function
and then, in further investigation, might be useful to replace it with a more general
function:
w(y; u, v;w) := (Ju + ,(V; v;w), wE Q ,

where, : B x R m xQ -+ Rand Q must be such that lev>ow :2 'Ii. In this section

we are dealing with the particular case, = (,\, G). The explicit dependence of w on
y is motivated by the fact that, in spite of what happens for constrained extremum
problems, the change of y does not merely imply a shift of K(y).
104
The above comments lead us to consider the transformed image of (2.1):

K(y;w) := {(u, v) E R X Rm : u = (F(y), y - x}, v = G(y; g(y; x);w), x E X(y)}

and its conic extension:

£(y;w):= K(y;w) - cl1{ ,
where cl denotes closure.

Definition 2.1. A function 'IjJ : KO -+ R with KO := {y E =: : y E K(y)} is said to be

a gap function iff 'IjJ(y) ~ 0 Vy E KO and 'IjJ(y) = 0 iff y is a solution of (2.1); KO is the
set of fixed~points of the point~to-set map K.
Since we will set up a gap function as a by~product of the separation scheme in the
image space, it is natural to expect to find two classes of gap functions, corresponding
to weak and strong separation functions [9]. The preceding definitions and notations
correspond to the weak separation; hence the function (where the dependence on w is
understood):

'ljJw(y) := min max

).EC* xEX(y)
[(F(y), y - x} + (A, G(y; g(y; x); w»)], (2.6)

will be shown to be a gap funtion for (2.1), and will be called weak gap function, since
it comes from (2.3).
In the definition of 'ljJw there is a maximization; this is performed on X(y), i.e. on
the relaxation of the domain of the QVI, and not on K(y), unlike what happens (in
the case of a VI) to the existing gap functions.
Note that, in the case (2.1)', we easily find:

min 'ljJw(y) = mm (F(y), y), (2.6)'

YEK(y) .EK(.)
F(.)EK* (.)

which shows a gap function for a Quasi~Complementarity System.

Theorem 2.1. Let y E KO. Assume that the extrema in (2.6) exist and for each
y E K(y) there exists w(y) E Q, such that
(i) £(y;w(y» is convex;
(ii) U := {(u,v) E 1{ : v = O} <1 T(&(y;w(y))), where T denotes Bouligand
tangent cone at (u = 0, v = g(y; y».
Then y is a solution of (2.1) iff 'ljJw(y) = o.

Proof. Let y E KO be a solution of (2.1), so that g(y;y) E C. Then, taking into

account that «(F(y), 0) , g(y; y» E cC 1{, the inclusion lev>ow:::> 1{ implies:

w(y; (F(y),y - y) , g(y;y); A,W(y» ~ 0, VA E C'.

Hence
max w(y;(F(y),y-x), g(Y;X);A,W(Y»~O, VAEC',
xEX(y)
 105

Hence
max W(Yi (F(y), Y - x), 9(Yi x)i A,w(y» ~ 0, 'VA E C*,
"EX(y)

so that tPw(Y) ~ O. Ab absurdo, assume that tPw(Y) > O. Then 3 a > 0, such that

max W(Yi(F(y),y-x), 9(YiX)iA,w(y»~a>0, 'VAEC*,

"EX(y)

and hence
(2.7)

Since Y solves (2.1), i.e. (2.2) is impossible or 'H n K(YiW(Y» = 0 or, equivalently,
'H n t'(YiW(Y)) = 0, then (i) and the obvious convexity of'H imply the existence of at
least one hyperplane which separates 'H and t'(Yiw(y)). Because of Theorem 2.1 of
[9], (ii) is necessary and sufficient in order that at least one of the above hyperplanes,
say H, does not contain Ui and this implies that there exists disjunctive separation,
namely t'(YiW(Y)) can be included in one closed halfspace defined by Hand 'H in its
complement, or 3 XE C· such that:

t'(YiW(Y» C;; {(u, v) E R X Rm : u + (X, v) :5 O},

and hence
W(Yi (F(y), Y - x), 9(Yi X)i X,w(y» :::; 0, 'Vx E XCV),

which implies
(2.8)

and contradicts (2.7). Now assume that tPw(Y) = o. Then 3 AE c· such that

max W(Yi (F(y),y - x), 9(YiX)iA,W(Y» = 0,

"'EX(y)

so that
(F(y),y-X) + (X,G(Yi9(YiX)iW(Y») :::;0, 'VxEX(y). (2.9)

Ab absurdo, assume that 3x E K(y) such that

(F(y), Y - x) > O.

Then, since 'VA E C· we have (A, 9(Yi x») ;::: 0, and, because of assumption (2.5), also
(A, G(Yi 9(Yi X)i w(y))) ~ 0 , we find

(F(y),y - x) + (A,G(Yi9(YiX)iW(Y») > 0 , 'VA E C*.

This inequality, at A = X, contradicts (2.9). o

106

Remark 2.3. IfVy E X(y) the condition g(y;x) E C is identically true on X(y),
so that K(y) = X(y), then Theorem 2.1 holds without (i)-(ii), so that tPw(Y) is a
gap function whatever F(y) and X(y) may be. To see this the proof of Theorem 2.1
requires a few changes only. In fact, according to Remark 2.2, the sufficiency does not
require any change. The necessity holds without (i)-(ii) up to (2.7). At this point note
that now

JC(y) ~ V:= {(u,v) E Rx R m : Vi = 0, i = 1, ... ,p; Vi ~ 0, i = p+ 1, ... ,m}.

Since JC(y) n 11 = 0, we have JC(y) ~ V\ 11, so that any hyperplane of R1+m, whose
equation is
p
u +L .xi Vi = 0 , if p > 0, or u = 0, if p = 0 ,
i=1

implies a disjunctive separation between JC(y) and 11. Then we again achieve (2.8) and
the absurd.

If the domain of the QVI is given as a set K(y) and not furtherly specified like in
(2.1), then - as said at the beginning of this section - suitable 9 and C can be found so
that, Vy E X(y), 9(Yi x) E C be identically true on X(y). Hence, as shown in Remark
2.3, (2.6) is gap function whatever G may be. For instance, we can set m = 1, C = R+,
and g(y;x) = 0, Vx E X(y), and choose G(y;v;w) = V Vv E R, Vw En:= R (the
definition n is of no importance); then (2.6) collapses to:

tPw(y)= max (F(y),y-x) . (2.10)

zEX(!I)

When X is independent of y, or QVI becomes a VI, then (2.10) becomes the gap
function introduced by Auslander [2].
In Theorem 2.1 we do not assume that K(y) is convex even when K(y) is indepen-
dent of y, unlike what happens in [2,7,16]. However it may happen that, K(y) being
not convex, (i) of Theorem 2.1 turns out to be satisfied. This is shown by the following
example.

Example. Consider (2.1) as VI (K(y) is independent of y; X(y) and K(y) are denoted
by X and K, respectively) in the particular case, where Xc R 2 , C = R+, m = 1, and

D;
set:
F(y) = G~: =: X = {(XI, X2) E R2 : ~ ~ Xi ~ 1, i = 1,2};

g(x)=x~+x~-1i

so that F is strictly isotone, but K is not convex. JC(y) is the set of (u, v) E R X R,
such that:
 107

or, if Y2 =I- !,
V= (1-~Y2)2 {u 2 + 2[YI + Y2 - 2(yi + yn + (2YI - I)XI]U+
{
+2[2(Yi + y~ - YI - Y2) + l]xi+
+2[(YI + Y2 - 2y; - 2yD(2YI -l)]XI + (YI + Y2 - 2yi - 2y~)} - 1,

1-~Y2 [u + YI + Y2 - 2(Yr + y~) + (2YI - I)XI] E [t, 1] ;

if Y2 = !, so that YI =I- !, we exchange Y2 with YI; (YI = !, Y2 = !) rI. K. Hence K(y)
is a family of arcs of parabolas. For instance, at y = (4,4) and Xl = 4 such an
arc is given by:

3V2 5 -3V2
v = (3 + 2V2)u 2 - (2 + V2)u - V2 - 4 , 2 - -2- ~ u ~ 4

and is art arc of a parabola; it does not intersect 'H., since such an Y solves (2.1), as it is
easy to see. If we select 'Y = v, namely if we choose a linear separation function (2.3),
then [(V; w) is not convex. While, if we consider (2.3)" and set 'Y = 1 - exp( -wv),
then simple calculations let us see that at w > 8 the above arcs of parabolas pass
from convexity to concavity. For instance, with w = 10 consider the arc of K(y; w) at
Y -_ (./2 ./2)
2'2'
X =
I 2'
./2.
3V2 5 - 3V2
v=1-exp{-10[(3+2v'2)u 2 -(V2+2)u]}, 2--2-~uE 4 ;

we find

v" = 1O{2(3 + 2V2) - 20[2(17 + 12V2)u 2 - 2(10 + 7V2)u + 3 + 2V2]}.

.exp{-10[(3 + 2V2)u 2 -
3V2 , 5 - 43V2] '
(V2 + 2)u]) < 0, Vu E [ 2 - -2-

and hence the concavity of the arc follows. Now [(V; 10) is convex. (i) of Theorem 2.1
is now fulfilled, while before was not. The above arcs are the images of the segments
t
{(XI,X2) E R2: Xl =constartt; ~ X2 ~ I}, ~ Xl ~ 1. t
Since (F(y), Y - x) is linear in x, such arcs become concave because of the trartsforma-
tion of g; more precisely 'Y can be interpreted as the convexification of the restrictions
of 9 on the rays {(XI,X2) E R2: Xl =constant; X2 ~ 2~}' t ~ Xl ~ 1, and hence on
the above segments. In conclusion, since (ii) is obviously verified, Theorem 2.1 Cart be
applied, notwithstanding the fact that K is not convex.
Same results are obtained if the above operator F is replaced with an anti tone
operator, as
F(y)=(2-2 YI ).
2 - 2Y2
108

3. THE STRONG CASE

Now it will be shown that, correspondingly to what happens for the general situ-
ation of the Theory of Optimization, we have the strong case here too. This leads us
to define another class of gap functions. Consider again (2.1), and the function:

S(YjU,VjW):= u - h(Yjvjw) , u E R, v E R m , wE 11, Vy E K O , (3.1)

where h: B x R m x11 -+ R is, Vy E KO, such that lev>o S C 'H, and h and 11 must be
such that, Vy E K O , the following conditions are satisfied:

3w E 11 s.t. K;(y) n lev=o s(y;u,Vjw) =F 0j (3.2a)

'HnK;(y) =F 0 => 3W E 11 and 3(u, v) E 'HnK;(y) s.t. s(YjU,Vjw) > 0, (3.2b)

where the level sets are considered with respect (u, v) only.
Given a kind of h we may have several possible sets 11. In this case we can choose
11 in order to simplify the subsequent development. Instead of (2.6) we consider now
the function:

t/Ja(Y):= max max [(F(y),y - x) - h(yjg(yjX)jW)], (3.3)

",EO ",EX(,1)

which will be called ~trong gap functionj the term strong is motivated by the fact that
(3.1) is a strong separation function in the sense of [9].

Theorem 3.1. Let y E K O• Assume that the maxima in (3.3) exist, and that
conditions (3.2) hold. Then y is a solution of (2.1) iff t/J.(y) = O.

Proof. Let y E K O be a solution to (2.1). Because of (3.2a) 3w E 11 and 3x E K(y)

such that:
S(yj (F(y),y - x), g(yjx)jw) = OJ
the inclusion lev>o S C 'H implies:

S(yj (F(y),y - x), g(yjx)jw) ~ 0, Vx E X(y) , Vw E 11.

Then the necessity follows. Now assume that t/J.(y) = o. Then:

max s(yj(F(y),y-x), g(yjx)jw)~O, VwE11. (3.4)

",EX(,1)

Ab absurdo, assume that 3x E K(y) such that (F(y), y - x) > o. Then:

«(F(y),y - x), g(Yjx)) E 'Hn K;(y),

so that 'H n K;(y) =F 0. Hence, because of (3.2b), 3w E 11 and 3x E K(y) such that:
S(Yj (F(y), Y - x), g(Yj x); w) > O.

This inequality contradicts (3.4) at W = w. o

 109

Remark 3.1. Note that we have tf;.(y) 2: 0, Vy E KO, as shown by the proof of
Theorem 3.1. Moreover, the proof of Theorem 3.1 does not require the existence of
the two maxima in (3.3).

Now consider a particular case of (2.1), namely:

(F(y),x - y) 2: 0, "Ix E K(y):= {x EX: g(y;x) E C} =f 0. (3.5)
Of course, when g(y; x) E C is identically true on X2, then (3.5) becomes a VI. The
QVI (3.5) is useful to show the nature, in terms of the separation scheme, of some
known gap functions for VI. Indeed, they turn out to be strong gap functions for the
QVI (3.5) in which the given VI has been embedded by means of an identically true
constraint g(y;x) E C. To see this we consider the particular case where:=: = Rn,
assume that X be convex, and set m = 1, C = R+,g(y; x) = (x - y, A(x - y)), with A
positive definite square matrix of order n. Since the constraint g(y; x) 2: 0 is identically
true, so that K(y) = X and K(y) ~ {(u, v) E R x R : v 2: OJ, with the position

o"( ·v·w ) - { wv , if v 2: 0
y" - +00, if v < 0
we have that the function s obviously satisfies the inclusion lev>o s C H. At x = y
we find that (u = 0, v = 0) E K:(y) and s(y; 0, 0; w) = 0, Vy EX, Vw 2: 0; hence (3.2a)
is verified at any w 2: o. In order to discuss (3.2b), note that, Vy E X, we have:
K(y) = {(u,v) E R x R: u = (F(y),y - x) ,
(3.6)
v=(x-y,A(x-y)), xEX}C{(u,v)ERxR:v2:0};
x EX} { 3£ E X s.t.
(3.7)
(F(y), y - x) > 0 => (F(y), y - £) > (£ - y, A(£ - y)).
(3.6) is obvious. Now let us prove [15] (3.7). Suppose that there exists x E X such
that (F(y), y - x) > o. Let us set ii := (F(y), y - x) , /3:= (x - y , A(x - y)). (Note
that (ii,/3) E K(y)). If ii > /3, then £ := x is the required point. Assume that /3 2: ii.
We set x(t) = (1 - t)y + tx, t E [0,1]. Since y, x E X and X is convex, then x(t) EX.
Thus we have (a(t), f3(t)) E K(y), where
a(t) := (F(y), y - x(t)) , f3(t) := (x(t) - y, A(x(t) - y)) .
We have
a(t) = (F(y), y - x(t)) = (F(y), y - (1 - t)y - tx)
= t(F(y), y - x) = tii,
f3(t) = (x(t) - y, A(x(t) - y)) = (t(x - y), A(t(x - y))}
= t 2 (x - y, A(x - y)) = t 2 /3.
Since
f3(t) t 2 /3 /3
--
a( t)
= -tii
= t- --+ 0
ii
as t --+ 0,

then there exists i E] 0, 1 [ such that

f3( ~) < 1 <=> f3( i) < a( i).
a(t)
Thus £ := x(i) = (1 - i)y + ix is the required point. o
110

Thus (3.3) becomes

¢.(y) = maxmax[(F(y),y
"'Ell ",EX
- x) - (x - y, wA(x - y»)],

which is the gap function considered by M. Fukushima [7]. Note that we could have
chosen other sets n. We have made the simplest choice: a singleton. This is not unique.
We might have chosen n = {w > O} finding the same result, or n = {O} finding again
(2.10) with X in place of X(y). The Auslander gap function is so particular that, with
it the two approaches become the same thing.
In a quite analogous way, by setting g(y;x) = A(x,y), A being non-negative,
continuously differentiable on X x X, strongly convex on X with respect to x, Vy E X,
and such that A(x,x) = 0, and

V.,A(y,y) =0 , Vy E X,

then we recover the gap function considered by D.L. Zhu and P. Marcotte [16].
In fact, we can prove the following [15].

Let g(y; x) = A(x, y), where A satisfies the following properties:

(i) A(x,y) ~ 0 , Vx,y.
(ii) Vy the function A(., y) is differentiable at y;
(iii) A(y,y) = 0, V",A(y,y) = 0, Vy.
Arguing similarly as above we can show that (3.7) holds true. Indeed, assume
that there exists x E X such that (F(y),y - x) > o.
Let a := (F(y), y - i) , ~:= A(i, y).
If a > ~, then x := x is the required point.
Assume that ~ ~ a and let x(t) := (1 - t)y + tx, t E [0, I]. Since y, x E X and X
is convex, then x(t) EX. Hence

(a(t),f1(t» E K(y):= {(u,v): u = (F(y),y - x), v = A(x,y), x EX},

where a(t):= (F(y),y - x(t»), f1(t):= A(x(t),y). We have

aCt) := (F(y), y - x(t») = t(F(y), y - i) = tao

Noting that x(O) = y and using (ii) and (iii), we have

f1(t) = A(x(t), y) = A(x(O), y) + V",A(x(O), y)(x(t) - x(O» + o(lIx(t) - x(O)I!) =
= A(y, y) + V",A(y, y)(x(t) - y) + o(lIx(t) - ylD = o(lIti - tylD = oCt).
(In fact we have x =I y, because a = (F(y), y - i) > 0). Consequently,

f1(t) _ oCt) 0
at( )
-
ta
--+ as t --+ o.

Hence, 3i E ]0, 1[ such that ~~2 < 1, and x := xCi) = (1- i)y+tx is the required point.
This completes the proof. 0
 111

4. FINITE DIMENSIONAL VI WITH POLYHEDRAL DOMAIN

Now consider again (2.1) as VI (X(y), K(y) and g(y; x) are replaced with X, K and
g(x), respectively) in the particular case where X = :=: = IRn , C = 1R+'; furthermore
let us set:
g(x)=Ax-b, S1=lR+" w='\, yT=(Yl, ... ,Yn), ( 4.1)

where b is a column-vector of R m, ,\ is a row vector of R, m, A is a m x n matrix, and

T as apex marks transposition. The weak separation function is selected among the
linear ones (F(y) is here a row-vector; G(y; v; w) = v; lC(y; w) = lC(y );£(y; w) = f(y)):

w = (F(y), y} - ('\, b) + ('\A - F(y), x}. (4.2)

It is easily seen that now K(y) is affine so that fey), being the sum of an affine
variety and a closed convex cone (namely -cRH), is convex - so that (i) of Theorem
2.1 is fulfilled - and coincides with its Bouligand tangent cone at (u = 0, v = g(y)).
Now note that 1-{ n K(y) = 0 implies the existence of a hyperplane, say H, which
contains K(y) and such that 1-{ C H+, H+ being one open halfspace defined by H;
therefore, T( f(y)) equals f(y) and, being included in the complement of H+, does not
intersect 1-{ and hence U, so that (ii) of Theorem 2.1 is satisfied at those y which solve
the present VI. Since this is enough (as noted in Remark 2.2), then Theorem 2.1 can
be applied.
Now we have to analyse the minimization of (2.6). To this end note, first of all,
that, without any loss of generality, we can assume that the system ,\ 2': 0, ,\A = F(y)
be feasible. In fact, in the contrary case, Vy E K, a well known theorem of the
alternative for linear systems (see [12], page 33, Th. 10) would give the existence of
t E lR n (depending on y) such that:

(F(y), t} =1 At ~ O.

By setting x := y - t, we would get Ax 2': Ay 2': b and thus deduce the existence,
Vy E K, of x E K such that:

(F(y),x - y) < 0

or the impossibility of the VI.

Since in (2.6) the existence of the maximum is not required (so that we can read
supremum in place of maximum), and since we can assume that ,\A = F(y) be feasible,
then, in the present case (4.2), (2.6) becomes:

1/Jw(y) = (F(y), y) + 'A=F(.)

min (.x, -b) ( 4.3)
.\~o

Remark 4.1. The symbols adopted in the present Remark are generic and indepen-
dent of those used elsewhere. Assuming that the minima exist, note that we have:

p/:= min [fey)

YER(y)
+ min
(y,A)ES
g(y, ,\)] = min [fey)
.ER(.)
+ g(y, ,\)] := p", ( 4.4a)
( •• ')ES
112

where the I-st minimization is performed with respect to y, the 2-nd with respect to
A, and the the 3-rd with respect to (y,A). Denote by P',P",Q(y) the problems in
(4.4a) and that inside P', respectively; by M" and A(y) the sets of minimum points
of P" and Q(y), respectively. Let M' be the set of pairs (y, A(Y)), where A(y) E A(y)
and y is a minimum point of P'. Then we have

M'=M" . (4.4b)

The proof is quite obvious: ab absurdo let (y., A·) E M"\M' and (y,). := A(Y)) EM',
so that, A· being a feasible solution for Q(y.) and y. for P', we have:

,.,,' = fey) + g(y,).) < f(y·) + g(y., A·) = P," ;

since (y,).) is a feasible solution for P" we get the absurdo and hence M"\M' = 0.
Now let (y,). := A(Y» E M'\M", and (y*, A·) E M", so that, (y,).) being feasible
solution for P", we have:

P," = f(y·) + g(y., A·) < fey) + g(y,).) = ,.,,' ; (4.5)

since A· is feasible for Q(y.), we have:

P," = f(y·) + g(y*, A·) ::::: f(y·) + (y.min

,A)ES
g(y. , A) ::::: p,'. (4.6)

Collecting (4.5) and (4.6) leads us to the absurdo. Hence (4.4) follow. o

According to Remark 4.1 the minimization of (2.6) on K(y) becomes equivalently:

min [(F(y), y) - H(y, A)l, (4.7a)

.EK(.)
.).EC·

where the minimization is performed with respect to (y, A), and where (w is considered
as fixed):
H(y,A):= - max [(A,G(y;g(y;x);w») - (F(y),x)].
xEX(,1)

In the case (4.1)-(4.2) we find H(y, A) = (A,b), so that, taking into account (4.3) and
the preceding analysis, (4.7a) becomes:

~;l! [(F(y),y) - (A,b)] (4.7b)

F(.)-XA=O
,,~o

where the minimization is meant with respect to (y, A).

Remark 4.2 In (4.1) K is defined by inequality only. If there are both equalities,
say p, and inequalities, say m - p, we set C = RP X R+-P j then the preceding analysis
receives only formal changes.
 113

Remark 4.3 The preceding development receives only formal changes, when g(x) =
Ax - b is replaced with g(y; x) = Ax + By - b, with B of order m x n, so that we have
now a QVI. Instead of (4. 7b), we find:

min
(A+B).~'
[(F(y), y) + (A, By - b)]. (4.7)'
F(.)-~A=O
,,~o

The case where F is linear has particular interest in the applications:

F(yf = C+ Cy (4.8)

where cT := (Cl,""C n ) and C:= (cij,i,j = 1, ... ,n). In this case (4.7b) becomes a
so~called quadratic programming problem, since C can be obviously replaced with
C := t(C + CT), if not symmetric. If moreover C is positive semidefinite, then (4.7b)
is a convex quadratic program. Note that in th<; case (4.8) the VI is a l~st order
optimality condition of problem min (y, C+ ~Cy) subject to y E K.
In some applications, like in the study of equilibrium of flows in a network, it is
useful to consider explicitly bounds on the variables. Hence X is replaced with

X := {x E R n : 0 ::; x ::; d} (4.9)

where dT := (d l , ... , d n ). In place of (4.2) we find now:

w = (c,y) + (y,Cy) - (A,b) + (AA - C - yTC,x). (4.10)

Now we will discuss two ways of solving the present VI by minimizing the gap function.
Because of Remark 4.1 the minimization of 'ljJw(y) now becomes:

mi~'ljJw(Y)
yEh
= min[(c,y)
.EK
+ (y,Cy) - (A,b) + cI>(y,A)], (4.11)
~~o

where (the present cI> is independent of that of Sect. 2):

cI>(y, A) := max (AA - C - yTC , x) .

xEX

Note that cI>(y, A), being the maximum of a parametric linear problem, where the
gradient of the objective function depends linearly on the parameter (y, A), is convex
and piecewise linear; call Cl, ... ,Cr the cones where cI>(y,A) is linear.
Another way consists in setting J = J(y, A) := {j = 1, ... , n : {>.A - yTC}j 2: Cj};
thus we easily find:

min'ljJw(y) = mil![(c, y) + (y, Cy) + IPw(Y)], (4.12)

yEK yEh

where

IPw(Y) : = mln[2: (f.

jEJ
dj
i=l
Aiaij - Cj - t
r=l
YrCjr) - f.
i=l
Aibi] =

2: + t f (l:
(4.13a)

=- (Cj YrCjr) d j + min djaij - bi) Ai,

jEJ r=l i=1 jEJ
114

!
subject to

n
~
L.J Aa"
I IJ
>
- Cj + L: YrCjr, j E J,
i=I r=I
m n
(4.13b)
L: Aiaij S Cj + L: YrCjr, j ~ J,
i=I r=I

A ~ O.
In fact, since now X is an interval of Rn, the maximization which appears in (2.6)
is easily solved in terms of the sign of the elements of vector AA - yTC - cT.
Note that i.f'w(Y), being the minimum of a parametric linear problem whose right-
hand side depends linearly on the parameter y, is convex and piecewise linear; call
DI, ... ,D. the cones where i.f'w(y) is linear.
If C is positive semidefinite, then (4.11) and (4.12) are convex and thus we can
use any method of convex quadratic optimization to minimize it on K.
If C is negative semidefinite, then the objective functions of (4.11) and (4.12) can
be viewed either as continuous and piecewise concave on the polyhedra
DI n K, ... , D. n K, or as a difference of convex functions, namely <I>(y, A) - (A, b)
or i.f'w(y) and -(c, y) - (y, Cy). For the former case some interesting analysis already
exists in [5]. In the latter case the minimization falls into a class of problems which
have received a lot of attention [11]; below some details will be discussed.
If C is indefinite, then we can use classic decomposition techniques to reduce
ourselves to subproblems of the above kinds.
Now assume that C be negative semidefinite and see (4.12) as a difference of two
convex functions (in short, DC-functions), so that the minimization of (4.12) can be
written as
min[1/!w(Y) = i.f'w(y) -u(y)], s.t. Y E K, (4.14)
y

where u(y) : -(c, y) - (y, Cy); i.f'w and u are obviously convex. We assume that

.((; := sup 1/!w(y) < +00 ; rp:= inf i.f'w(y) > -00 ;
yEK yEK
(4.15)
if := sup u(y) < +00.
yEK

In the applications these assumptions do not seem to be restrictive. It is easy to show

that (4.14) is equivalent to the problem (see [11], Th. 19, p. 39):

minr, s.t. i.f'w(y)-sSO; u(y)+r-s~O; yEK; rER; sES, (4.16)

y,r,s

where now R := [0, .((;]C It , and S := [rp, if + .((;]C R. In fact, by a well known device
(4.14) is obviously equivalent to

minr, s.t. i.f'w(y) -u(y) S r ; y E K ; (r E R). ( 4.17)

Y.r

Since i.f'w(y) S u(y) + riff 3s E R such that i.f'w(y) S s S u(y) + r, problem (4.17)
turns out to be equivalent to

minr, s.t. i.f'w(y) S s; s S u(y) +r; y E K. (4.18)

Y,r,s
 115

Now, to see the equivalence between (4.14) and (4.16) it is enough to note that, by
Theorem 2.1 and the subsequent remark, tPw(Y) ;:::: 0 Vy E K and thus on the feasible
region of (4.17) we have r ;:::: 0, and to take into account (4.15). Problem (4.18) is of
the kinds studied in [11], where some solving algorithms are proposed. Quite similar
remark can be done for (4.11).

5. CONNECTIONS WITH DUALITY

This section contains some remarks, which aim to stimulate the investigation on
the duality theory for QVI. Among the several aspects, which concern the present
topic, only one is considered here: the connections between gap functions and duality
for extremum problems. Such connections can be analyzed in at least two ways. With
the former we perform suitable transformations directly on the gap function. With the
latter we replace the minimization of the gap function with a simplified minimization,
like (4.7), and perform the transformations on the simplified problem.
Now let us describe the former way. Consider again (2.6) and the transformed
operator Fo(y) := F(y) - F(O). Within square brackets of (2.6) substract and add
~(Fo(Y),Y); then the minimization of (2.6) becomes:

m~n
YEK(y)
{(F(Y) - ~Fo(Y),y)
2
+ ~(Fo(Y),y)+
2
(5.1)
min max [(A,G(y;g(y;x);w)) - (F(Y),x)]} ,
AEC· xEX(y)

where w is fixed. Then from (5.1) we deduce:

min tPw(y)::::: m~n (F(y) - -21 Fo(y),y) - max [H(y,A) - ~(Fo(Y),y)]

YEK(y) YEK(y) .~:c(:) 2
(5.2)
;:::: min (F(y) - ~Fo(Y),y) - m<L]C[H(y,A) - ~(Fo(Y),y)],
YEK(y) 2 feet. 2

where H has been defined in Sect. 4.

When (2.1) is a VI and X,g,G,n are as in (4.1), (4.8), (4.9), then (5.2) takes the
form:
min tPw(Y) :::: min(F(y) -
yEK yEK
~Fo(Y),
2
y)-
1 (5.3)
- max [(A, b) - -2(Fo(Y), y) - max(A T A - F(y), x)]
.ERR xEX
AER+

= miI!(F(y) - -21 Fo(Y),Y) - max [(A,b) - -21 (Fo(y),y) - 8*(AT A - F(y);X)],

yEK .ERn
.\ERfn

where 8* denotes support function. Instead of (5.3) we can equivalently (in the sense
116

that the maxima appearing in them are equal) consider the following:

min!f1w(Y)
yEK
~ min(F(y)
yEK
- -21 Fo(y),y)-
1 (5.4)
- 'E~;~R+ [(A, b) - 2(Fo(y), y) - (fI, d)]
pOtAT'-F(.)
pERi-

When d = +00, in (5.4) we must put fI = O.

In the particular case where F(y) = e ERR, so that Fo(y) = 0, (5.4) becomes

min !f1w(y) = min(e, x) - max [(A, b) - (d, fI)] = 0, (5.5)

yEK "'EK P.,p)EK*

where K* := {(A'fI) E Rf.' x Rf, : ATA - fI ~ e}, and where the equalities are moti-
vated by the fact that Th.2.1 can be applied and implies that the LHS of (5.5) is zero,
and by the so-called weak duality theorem for linear programs. As a consequence of
(5.5) we obtain the so-called strong duality theorem for linear programs (on polytopes).
In the particular case where F(y) = e + Cy (as in (4.1), (4.8), (4.9)), so that
Fo(Y) = Cy, (5.4) becomes:

miq !f1w(y) ~ min(c + -2

1 Cy, y) - max [(A, b) - (fI, d) - ~(Cy, y)], (5.6)
yEK yEK (y,'\,p)EK* 2

where K* := {(y, A, fI) E RR x Rf.' x Rf, : AT A - fI - Cy ~ e} collapses to the above

K* at C = O. If C is symmetric, the so-called weak duality for quadratic programs
implies that the RHS of (5.6) is ~ O. Hence, when the LHS is = 0, (as in Th. 2.1),
then in (5.6), the ~ is verified as = and the RHS is = O. Thus we derive the so-called
strong duality theorem for quadratic programs.
Consider again the case F(y) = c, where (2.1) becomes:

(e, x - y) ~ 0 , \:Ix E K, (5.7)

and is obviously a (l-st order) necessary and sufficient optimality condition for y to
be minimum point of the I-st problem in the 2-nd side of (5.5). Hence, it is natural
to consider the analogous (l-st order) necessary and sufficient optimality condition for
the 2-nd problem in the 2-nd side of (5.5), namely

(-b,d) , (A - X,fI- it») ~ 0 , \:I(A,fI) E K*, (5.8)

where K* is that of (5.5), and (X, jt) is the unknown; when d = +00, then in (5.8) we
must put fI = it = O. It is natural to call (5.8) dual variational inequality of (5.7).
Note that the operator of (5.7) is not injective. It is useful, even if quite obvious, to
note that, if we consider (5.8) in place of (2.1), and we repeat the development which
has led to (5.5) (note that now the domain of the VI is K*, so that now d = +00 and
A, b are respectively:
 117

and F(y) is replaced with (-b, d), so that Fo == 0), then, in place of the problems which
appear in the 2-nd side of (5.5), we find respectively:

min [(A,-b)(d,p)] =- max [(A, b) - (d,p)] ,

(>',Il)EK" (>',Il)EK"

and
max
>'ERm ;JlERnjIER+
[-(c, y)] = min
Ay>b
(c, y).
O:s;i~d
I'ER+irER+;
_,T AT+IITS-bTill+r:S.d

Hence we find again the problems in the 2-nd side of (5.5), but in reverse order.
Consider again the case F(y) = c + Cy, where (2.1) becomes:

(c + Cy, x - y) ~ ° , "Ix E K. (5.9)

The 1-st order condition for the 2-nd problem in the 2-nd side of (5.6) is

(-b,d,Cy),(A->',p-P"x-y))~O , V(A,p,x)EK*, (5.10)

where K* is now that of (5.6), and the unknown is (A,p,y). We call (5.10) dual
variational inequality of (5.9). We consider (5.10) in place of (2.1) and we repeat the
development which has led to (5.6). Now the domain of the VI is K*, so that d = +00
and A, b are respectively

and F(y) is replaced with (-b,d,Cy), so that Fo(A,p,y) = (O,O,Cy). Then, in place
of the problems which appear in the 2-nd side of (5.6), we find respectively:

min [(A,-b)+(d,p) + ~(Cy,y)] =

(>',Il,y)EK" 2
1
max [(A,b) + (d,p) - -(Cy,y)] ,
(>',Il,y)EK" 2

and

max
.\ERm i}lER" iYER+.
[-(c, y) - -21 (Cy, y)] = min [(c, y)
Ay>b
+ -21 (Cy, y)] .
vER+ iTER+.
1;5
0;5 d

_liT AT+"T5-bTiy+r5d
.TC$(C.)T

Hence we find again the problems in the 2-nd side of (5.5), but in reverse order.
Now let us consider a second way for establishing relationships between QVI and
duality for extremum problems. (5.1) has been obtained through a special transforma-
tion of (2.6). Now we start with (4. 7a) instead of (2.6). Since in (2.6) the existence of
the maximum is not required (so that we can read supremum in place of maximum),
118

then the minimization of (2.6) is equivalent to (4. 7a). Let f : :::: -t R be a generic
function, and set
L(y; >.) := fey) - (F(y), t) + H(y, >.).
By exploiting (4.7a) we find:

min ~w(y) = min {fey) + [-L(y; >.)]} ~ min fey) - maxL(y; >.), (5.11)
YEK(y) .~:c(!) YEK(y) AEC·

where the 1-st and 3-rd minimizations are done with respect to y, and the 2-nd and
4-th with respect to (y, >.).
In the case (4.1)-(4.2), taking into account that H(y,>.) = (>.,b) and that we can
assume F(y) = >.A, so that L(y; >.) = fey) - (>.,Ay - b), (5.11) becomes:

mil! ~w(y) ~ min f(y) - max L(y; >.). (5.12)

yEK Ay>b F(,)-AA=O
- .\2:0

Denote by P and P' respectively the problems in the RHS of (5.12). Let (y', >") be
an optimal solution of po. If y' is an optimal solution of P (so that Ay' ~ b), then
we easily deduce that (>", b) :::; (>", Ay*), and thus:

L(y*; >'*) :::; f(y*) - (>.*, Ay') + (>", Ay') = fey'),

where the inequality holds as equality if (>",Ay*) = 0 (complementarity); it follows
that the RHS of (5.12) is ~ O. If, in addition, the assumptions of Theorem 2.1 are
fulfilled, so that the LHS of (5.12) is =0, then we have

min fey)
Ay>O
= max
F(,)-AA=O
L(y; >.). (5.13)
- .\~o

When F(y) is the gradient of a convex function, which is chosen as fey), then P*
turns out to be the Lagrangian (Wolfe) dual [12J of P.

6. SOME EXTENSIONS
We shall now briefly discuss some questions that arise from the development car-
ried out in the previous sections.
The analysis of Sect.5 has been done by exploiting a weak gap function. It can be
done analogously in the case of a strong gap function of Sect.3.
In Theorem 2.1 we assume the existence of w. This is guaranteed if 9 is C-concave
and K convex. More general conditions, which ensure the linear separability of t: and
'H, are useful, as well as the comparison with the existing ones. Analogous question
arises in Theorem 3.1. An important class of problems, recently studied in terms
of VI and QVI, is the one that considers with the equilibrium of flows on a road
or on a computer network. Same questions for (ii) of Theorem 2.1; under convexity
assumptions the well known Slater condition (or its generalization) guarantees it.
 119

In (2.1) (F(y), x - y) ~ 0 can be replaced with the more general Variational

Inequality (F(y),x - y) ~ fey) - f(x), where (-,.) is the pairing between a Hilbert
space and its dual and f : K --4]- 00, +00]. In such a case, f being or not identically
zero, it is useful to distinguish the cases where the image of 9 is finite dimensional
(and in this case the approach of the preceding sections works) from the case where
the image is infinite dimensional. In the second case a multifunction approach may be
adopted.
Conditions under which 1/Jw and 1/J. are (semi)continuous, or (semi)
differentiable, or convex, or quasiconvex, or pseudoconvex, or geodesically convex,
and so on, are extremely useful. Some results can be found in [7,13,15].
The example of SectA shows that G can be used to treat VI or QVI where the
domain is not convex or the operator is not isotone. When G is separable, its compo-
nents, namely Gi, can be interpreted as a convexification of parts of the domain, just
as in Example of Sect.2.
The properties of 1/Jw and 1/J. might give more insight on the topics related to the
integrability of (2.1).
Sect.5 is merely a proposal for deepening the relationships between the separation
scheme and duality theory, in particular the relations among the 3 problems which
appear in (5.2) or (5.3) or (504) and those in (5.11) or (5.12). Of course, these should
be connected with the existing theory for VI. In the particular case, where (2.1) is
the formulation of an equilibrium problem - for instance in a network -, it would be
useful to know when A represents a potential. The particular case (4.1),(4.8), (4.9)
where C is not symmetric but has non-negative eigenvalues, is a test problem for such
investigations.
Again with reference to equilibrium problems in a network, the equilibrium of flows
is defined by the so-called Wardrop Principle [5]. In this case (2.1) in the particular
form (4.1), (4.8),(4.9) (hence (2.1) is a VI) and with dj = +00, j = 1, ... , n, expresses
the equilibrium; y is the vector of flows on several paths, and F(y) is the vector of
weights (time, cost, ... ) of several paths. Unfortunately the assumption dj = +00
is unrealistic. More realistic models should be considered. For instance, we might
consider the case dj < +00; if in correspondence of a solution of the VI the "reac-
tions" of the constraints Yj ::::; d j are zero, then the solution represents an equilibrium.
Otherwise, the reactions offer an index of the lack of equilibrium and might suggest
modifications of the networks to achieve equilibrium. The above question is extended
to the quasi-variational case; this means to study the "equilibrium" flows in a net-
work with capacities in the arcs and with elastic demand. Existence and uniqueness of
equilibrium flows are now open questions. The following rem81·k may be of some help.
In [14] there is a sufficient condition for the existence of the minimum of a con-
strained extremum, which generalizes the known ones of Weierstrass type by using the
image space approach. In this approach the starting point is the separation between
two sets, like here. Hence it is conceivable to exploit such a condition (or, more gen-
erally, the same approach) for studying the existence of solutions to (2.1) in order to
generalize the existing conditions. However, we must notice a substantial difference
120

between the separation scheme adopted in (14) for extremum problems and the one
exploited here. In the former case the image of an extremum problem receives a trans-
lation in the direction of u axis, if the candidate y for a minimum point is changed;
here, as (2.2) shows, the image receives any transformation, if y changes.
The uniqueness of solutions to a VI or to a QVI may be reduced to that for
the corresponding gap function, by using the following obvious proposition. Given a
function (not necessarily the gap function) t/J : K ~ JRn -+ JR, y is unique minimum
point iff, Va E R,n, the system (in the unknown z)

(a,y-z) >0 ; t/J(y)-t/J(z)~O; zEK

is impossible. In the particular case of VI with C = JR+ and t/J as the Auslander gap
function, it becomes: a solution y of the VI is unique iff, Va E R,n, the system (in the
unknown z):

(a,y - z) > 0 ; (F(z),z - x) ~ 0 , "Ix E K ; z E K,

is impossible. All this can be extended to QVI and to the infinite dimensional case.
The format (2.1) is not the most general one for VI. In [3) a more general scheme
is considered: to find y E K such that

F(y,x)~O, VxEK,

where F : K X K -+ R is such that F( x, x) = 0 , "Ix E K; and several interesting

properties are established. From one side it is interesting to extend the above study to
the QVI; this means that K becomes K(y) of Sect.2. From the other side the approach
of Sections 2-5 can be applied to such general formats; in this case, for instance, (2.2)
is replaced with

u:=F(y,x»O, v:=g(y,X)EC , xEX(y).

This differs from a general system, like that considered in [9), for the fact that
F( x, x) = 0 and for the dependence of the domain X on the unknown. The format
(2.1) contains also infinite dimensional QVI, like the case where g(y; x) E C represents
an integral constraint, but does not contain the case where g(y; x) E C is a geodesic
constraint or a differential equation or inclusion. Extension of the results of Sect.s 2-5
to these cases may be conceived by means of the multifunction approach outlined in
Sect.4 of [10).

ACKNOWLEDGMENT

The research has been partially supported by the Project "Trasporti" of Italian
National Research Council (CNR), which is here gratefully acknowledged.
 121

REFERENCES

[I] C. Antoni, "On a separation approach to Variational Inequalities". This volume.

[2] A. Auslender, "Optimization. Methodes numeriques". Masson, Paris, 1976.
[3] E. Blum and W. Oettli, "From Optimization and Variational inequalities to equilibrium prob-
lems". Proceedings of the 57-th Conference of the Indian Mathern. So., Published in "The
Mathematics Student", Vol. 63, N. 1-4, pp. 1-23, 1993.
[4] M. Castellani and G. Mastroeni, "On the duality theory for finite dimensional Variational In-
equalities". This volume.
[5] M. De Luca and A. Maugeri, "Quasi-Variational inequalities and applications to equilibrium
problems with elastic demand". In "Nonsmooth optimization and related topics", F.H. Clarke
et al. eds., Plenum, New York, pp. 61-77, 1989.
[6] M. De Luca and A. Maugeri, "Discontinuous quasi-variational inequalities and applications to
equilibrium problems". In "Nonsmooth optimization. Methods and applications", F. Giannessi
ed., Gordon and Breach U.K., pp. 70-75, 1992.
[7] M. Fukushima, "Equivalent differentiable optimization problems and descent methods for asym-
metric variational inequality problems". Mathematical Progr., Vol. 53, pp. 99-110, 1992.
[8] F. Giannessi, "Theorems of the alternative, quadratic programs and complementarity problems" .
In "Variational inequalities and complementarity problems", R. W. Cottle et al. eds. J. Wiley,
New York, pp. 151-186, 1980.
[9] F. Giannessi, "Theorems of the alternative and optimality conditions". Jou. Optim. Theory
Appl., Vol. 42, N. 3, pp. 331-365, 1984.
[10] F. Giannessi, "Theorems of the alternative for multifunctions with applications to optimization:
General results". Jou. Optim. Theory Appl., Vol. 55, N. 2, pp. 233-256, 1987.
[11] R. Horst and H. Tuy, "Global optimization". Springer-Verlag, Berlin, 1990.
[12] O.L. Mangasarian, "Nonlinear programming". McGraww-Hill, New York, 1969.
[13] T. Rapcsak, "On the connectedness of the solution set to nonlinear complemantarity systems".
Jou. Optim. Th. Appl., Vol. 80, No.3, pp. 501-512, March 1994.
[14] F. Tardella, "On the image of a constrained extremum problem and some applications to the
existence of a minimum". Jou. Optim. Th. Appl., Vol. 60, N. 1, pp. 93-104, 1989.
[15] N. D. Yen, private communication.
[16] D. L. Zhu and P. Marcotte, "A general descent framework for monotone variational inequalities".
To appear in Jou. Optimiz. Th. Applic.
STABILITY OF MONOTONE VARIATIONAL INEQUALITIES
WITH VARIOUS APPLICATIONS

Gwinner Joachim

Institute of Mathematics
Department of Aerospace Engineering
University of the Federal Army Munich
D - 85577 Neubiberg, Germany
e-mail: 11lbgwin@rz.unibw-muenchen.de

Dedicated to Professor W. Krabs on the occasion of his sixtieth birthday

Abstract: The convergence and stability theory of Mosco is extended by the more
general approach of monotone - convex functionals and by weakening uniform equi-
monotonicity and equicoervity assumptions to semicoervity conditions. The abstract
theory is firstly applied to a finite dimensional Variational Inequality that models dis-
tributed market equilibria with box constraints providing a stability result with respect
to the upper bounds. Then it is applied to obtain stability results with respect to co-
efficients and unilateral constraints in p-harmonic elliptic unilateral boundary value
problems that can be considered as scalar models of the unilateral contact problem
without, respectively with given friction.

Key Words: Variational Inequality, Convergence, Stability, Monotonicity, Market

equilibria, Contact Problems.

1. INTRODUCTION

There is already a growing literature on stability for Variational Inequalities and

related equilibrium problems. In the pioneering work [36) Mosco deals with the conver-
gence of the solutions of perturbed Variational Inequalities where both the monotone
operators and the convex sets are perturbed using different topologies for upper and
lower limits. Doktor and Kucera [12) use the Hausdorff distance and a distance given
by projections for convex closed sets in Hilbert space and give norm estimates for the
distance between the solution sets. In [32) Maugeri studies a class of parametric Varia-
tional Inequalities and presents stability results in the sense that the solution set does
not change with a special change of the parameter. A related global stability concept
has been investigated by Smith [43) for traffic equilibria. Also for network equilibria

123
124

Daiermos and Nagurney [l1J give a stability result for perturbations in the cost func-
tions. More recently McLinden [33J comprehensively investigates stability understood
as nonemptiness of the solution set of locally perturbed Variational Inequalities.
In this paper we follow the stability concept of Mosco and use his notion of set
convergence, what is today called "Mosco convergence". Instead of considering single
valued monotone operators we more generally work with monotone-convex functionals
as initiated by Minty and Oettli [35,5J. This approach allows to consider also set-
valued monotone operators. However, the main difference between our contribution
and previous work is that we relax uniform monotonicity and hence coercivity to
semicoercivity conditions. Under these conditions solvability and even more stability
of the Variational Inequality depend in a subtile manner on the data of the problem,
namely on the monotonicity and growth behaviour of the nonlinear monotone-convex
functional, the constraints and the inhomogenous term at the right hand side. Thus
our analysis do not only cover the well-posed case [13, Chapter II, Section 5J, but also
treats the ill-posed case [14, section 7.3.3J, [26, Chapter VIIJ, where uniqueness may
be lost. Therefore we arrive at new results also in finite dimensions.
As the monograph of Glowinski [18J shows, Mosco convergence can be refined to
give an excellent tool for the analysis of the convergence of discretization methods, in
particular finite element methods applied to elliptic Variational Inequalities. Also the
abstract stability result presented in this paper can be understood as an approximation
result that applies to the convergence of finite element approximations for semicoercive
nonlinear elliptic unilateral boundary value problems; for details we refer to [22J. A
related discretization result in the case of bilinear forms, where uniform monotonicity
or equivalently positive definiteness is relaxed to a Garding inequality, is given in [20J
and applied in [21J to the convergence of boundary element approximations to contact
problems in linear elastostatics with Signorini boundary conditions. II-posed elliptic
Variational Inequalities in terms of weakly coercive bilinear forms and their application
to contact problems are treated by iterative multi-step prox-regularization methods in
the recent monograph [26, Chapter VIIJ. For earlier convergence analysis of Galerkin
discretization of monotone operator equations under relaxed coercivity conditions we
refer to the poetic paper [49J.
In the present paper convergence results for discretization methods are outside of
the scope. So to illustrate the power of our abstract stability result we derive sta-
bility results for perturbed data, like perturbed constraints and perturbed operators,
in concrete Variational Inequalities. These Variational Inequalities model some equi-
librium problems drawn from operations research, respectively model some unilateral
boundary value problems stemming from mathematical physics.

2. THE VARIATIONAL INEQUALITY PROBLEM

Let us formulate the Variational Inequality problem in the following way: Find ii.
in a real reflexive Banach space (V, II . II) such that
(P) ii.EK,cp(ii.,v)~'\(v-ii.), VvEK.
Here K ~ V is nonempty, closed, and convex, ,\ belongs to the dual space V·, and
cp : K x K -+ IR satisfies appropriate convexity and monotonicity assumptions. In
particular, following Minty-Oettli [35,5J, cp is termed monotone, if

(2.1)
 125

Moreover, for any u E K, <pC u, .) is assumed to be convex and lower semi continuous on
K, and to satisfy <pC u, u) = O. Note that a set-valued monotone operator T : V ---> V·
gives rise to a monotone-convex functional by

<p(u, v) = sup{(v - u) : (E T(u)} .

Also <P(Vl,V2) = t/J(V2) - t/J(Vl) is monotone-convex for a given convex lower semi-
continuous functional t/J. Hence the above formulation includes so-called Variational
Inequalities of the second kind [18] without pain.
To introduce further assumptions, let I . I be a continuous seminorm on V. Thus

Y = {y E V : Iyl = O}
is a closed subspace. With ~ denoting weak convergence in contrast to strong conver-
gence denoted by -+ , we require:
(AI) For any sequence {Vn}nEN with Ivnl-+ 0, Vn ~ v and IIvnll ;::: ." for some.,,> 0
there exists a subsequence {v n • hEN such that Vn.-+ v.
Obviously in finite dimensions, (AI) trivially holds, since then weak and strong con-
vergence coincide. To state more concrete sufficient conditions for (AI) to hold, we
insert from [25, Sect. 5.3 (25)], [19, Sect. 3.4] the following
Lemma I The IL8sumption (A 1) is satisfied in any of the foliowing two cases:

(i) 11·11 is equivalent to 111·111 + 1·1, where 111·111 is another norm on V; the dimension
of Y is finite; there exists a > 0 such that

inf IIlv - ylll :$ aivi Vv E V.

IIEY

(ii) V is compactly embedded in another Banach space (X, II . IIx); II 'IIv == II· II is
equivalent to 1·1 + II . IIx .
Moreover the seminorm 1·1 is continuous.
Proof. The inequality 1·1 :$ 111·111 + 1·1, respectively 1·1 :$ 1I·lIx + 1·1 and the equivalence
of the sum to the original norm 1I·lIv implies the continuity of the seminorm 1·1.
To prove (i) let {Vn}nEN a sequence in V such that Vn ~ V, Ivnl-+ o. Then for any
n choose Yn E Y such that

IIlvn - Ynlll :$ inf IlIvn - ylli + .!. :$ alvnl + !.n .

yEY n

Taking Wn = Vn - Yn gives IIlwnlll-+ 0 and by

and by the equivalence of the sum to the original norm 1I·lIv one obtains IIwnliv -+ O.
Thus Yn = Vn - Wn ~ v in the finite dimensional subspace (Y, II . II) whose unit ball
is compact. Hence for the bounded sequence {yn} there exists a subsequence {Yn.}
strongly converging to v. Altogether one concludes IIvn • - vllv = llYn. +W n • - vllv -+ O.
To prove (ii) let {Vn}nEN a sequence in V such that Vn ~ V, Ivnl -+ O. Since the
seminorm I· I is weakly lower semicontinuous, Ivi = 0 holds. By the boundedness of
the sequence {v n } and by the compact embedding of V in (X, II . IIx) there exists a
subsequence {v n .} strongly converging in II . IIx to some x EX. Since X* ~ V* and
126

the weak topology separates points, x = v holds. Finally the equivalence of the sum
to the original norm 1I·lIv again implies IIvnk - vllv --+ O. 0

As we shall see later, the assumptions of Lemma 1 can be satisfied in Sobolev spaces
in virtue of the Kondrachov-Rellich embedding theorem.
Now we introduce several assumptions concerning the functional cpo Having our
applications in mind we can take for granted that there exists some y E Y n K and
some constants ao ~ 0, al ~ 0 such that

cp(y, w) ~ -ao - ally - wi 'Vw E K. (2.2)

Note that this condition is satisfied, if the affine minoranat, which is obtained by
subdifferentiability of cp(y, .) at y, is continuous with respect to I . I. Since we need
detailed information about the monotonicity, growth, and Lipschitz behaviour of cp,
we require:

(A2) lu-vlg(lu-vl) < -{cp(u,v)+cp(v,u)} 'Vu,vEK,

(A3) Icp(u,v)-cp(u,w)1 10+'Ilul+G(luDlv-wl 'Vu,v,w E K,
~

where 10 ~ 0 ,,1 ~ 0 , and the functions 9 and G : [0, 00) --+ R are monotonously
increasing and positive for positive arguments. Moreover for t --+ 00 we require that
get) --+ 00 and G(t) --+ 00 and that

(A4) lim G(t) =0.

t ... oo tg(t)
Note that in (A2) we can fix u = Y E YnK and derive from (2.2) that cp is semicoercive
in the sense that
-cp( v, y)/Ivl --+ +00 as v E K, Ivl --+ 00 .
Moreover (A2) implies that cp is monotone as defined above. More precisely (A2)
relaxes the assumption of uniform monotonicity by replacing the norm II . II by the
semi-norm I . I . (A3) describes a growth condition in the first variable and also a
relaxed form of Lipschitz continuity in the second variable. As a weak continuity
condition with respect to the first variable, we demand that cp is hemicontinuous:

For any v E K, w E V, the function t E IR f-+ cp(v + tw,v) is upper

semicontinuous (the argument v + tw varies in K only).

Before we address the existence of a solution u to (P), we insert the following useful
result for monotone-convex functionals which goes back to Minty [34] and was firstly
formulated in [35].

Lemma 2 If cp is monotone and hemicontinuous, (P) is equivalent to

11 E K , cp( v, 11) ~ >.(11 - v) "Iv E K . (2.3)

Proof. (P) => (2.3) is immediate from the monotonicity. To prove (2.3) => (P) fix
v E K. Then, for any s E [0,1), w. = v + s(u - v) E K, hence
cp(W., 11) ~ >.(u - w.) = (1 - s)>.(11 - v).

By convexity of cp( w., .),

0= cp(W.,w.) ~ (1- s) cp(W.,v) +s(l- s)>.(11 - v),

 127

hence
SA(V - u) ~ Cf'(w.,v).
Thus by hemicontinuity, we arrive in the limit S --+ 1 at (P). o
Since we do not presuppose that K is bounded we use the concept of the asymptotic
cone of K, due to Fenchel [15], which is defined for some fixed Vo E K as follows:

acK = {v E V : Vo + tv E K, '<It> O}.

Note that by the closedness of K, this definition is independent of the chosen Vo E K.
Now the existence of a solution u to the Variational Inequality (P) can be guar-
anteed (see [25, §5.3, p. 185] in the case of (pseudo-)monotone operators A), if the
following additional assumption is met:
(AS)
Y n ac K is bounded
or A = Al + >'2 satisfies the two conditions
>'I(Y) < 0, '<Iy E Y n acK\{O}, (2.4)
(2.5)

3. AN ABSTRACT STABILITY RESULT

In this section we present our abstract stability analysis. Thus in addition to the
real valued functional Cf' defined on K x K we are given a sequence {K" }"EN of nonvoid
convex closed subsets of V and a sequence {Cf'v}VEN of real valued functionals defined
on K" x K". We do not suppose that all the K" are contained in K. Instead we follow
the concept of Mosco [36] (see also [44,13]) of set convergence ("Mosco convergence")
introducing the subsequent two hypotheses:

(HI) If {V"}"EN weakly converges to v and for all". E IN, v" lies in K"" for a subsequence
{Kv"},,EN of the sequence {K,,}vEN, then the weak limit v belongs to K.
(H2) For any v E K there exists a sequence {VV}"EN strongly converging to v such
that v" lies in Kv for all large v.
These two hypotheses can be shortly phrased as

w-limsupK" S;; K, s-liminf K" :2 K ,

"-+00 "-+00

where in the sense of Kuratowski (see e.g. [2,29]) w-limsup, respectively s-liminf de-
notes the limes superior with respect to weak convergence, respectively the limes infe-
rior with respect to strong convergence. Note that only for simplicity we work with
sequences, all the analysis to follow carries through also with nets or filters.
Let us remark that Stummel [47] obviously independently developed this concept of
set convergence in his study of perturbations of linear elliptic boundary value problems
in Sobolev spaces. He notes in [47, p. 11] that these two hypotheses are equivalent to

w-limsupK" S;; K S;; w-liminfK", s-lim sup K" S;; K S;; s-lim inf Kv ,
"-+00 11-+00 11-+00

since strong convergence implies weak convergence. For further characterizations of

Mosco convergence, in particular in terms of projections and distance functions we
refer to [13,44].
 128

In his convergenc~ analysis of discretization methods for elliptic variational inequal-

ities Glowinski [18, p. 9) refined this concept of set convergence by replacing (H2) by
the following condition - here stated for sequences instead of nets in the discretization
parameter h > 0:

(H2') There exist a dense subset D of K and for all II E IN mappings r" : D -+ V with
the property that, for each v E D, r,,(v) -+ v (II -+ 00) and rAv) E K" for all
II ~ 1I0( v) for some 1I0( v) > o.

The emphasis of (H2') lies upon the fact that when applied to finite element approx-
imation, the mappings r ll can be explicitly stated by interpolation only on a dense
subset, given by smooth enough functions. Since we are here not interested in dis-
cretization analysis, we stick to the original Mosco convergence. On the other hand,
Stummel [47, p. 11,12] (see also [46, section 4.1(9) p.78]) proved:

The relation K ~ s-liminf K" holds, if and only if there exists a dense
subset D ~ K with the property D ~ s-liminf K".

In addition to (HI) and (H2) it is important for our stability analysis that the inter-
section K n n"EN K" should be nonvoid. In view of our applications it is not hard to
require that there is even some ii E K n Y such that ii E K., and (2.2) is uniformly
satisfied in II E IN, that is there exist some constants bo ~ 0, b1 ~ 0 such that

(3.1)

Also concerning the sequence {'>I',,}"EN, we now require that the assumptions (A2) and
(A3) hold uniformly in /I E IN. In order to approximate '>I' by {'>I',,} , we need the
following hypothesis:

(H3) Let v E K and according to (H2) let {v,,} be an approximating sequence to v,

i.e. 11lim
.... 00
IIvlI - vII = 0 and v" E K" for /I > 110 = 1I0(V). Let in addition {W"}"EN
be a sequence such that W" E KII for all /I E IN and W" -'" W as II -+ 00. Then we
have

This hypothesis corresponds to the assumption (Ass I) with Mosco; more precisely in
the special setting '>1'.,( v"' w,,) = (T,,( VII)' wlI - VII)' '>1'( v, w) = (T( v), W- v) for single val-
ued monotone operators Til, T (Ass I) requires the norm convergence of Til ( VII) to T( v)
and hence implies (H3). Moreover in the case Cf'1I(V." w,,) = ,p1l(W II ) - ,p1l(VII ), Cf'(v, w) =
,p(w) - ,p(v) for convex functionals ,pll',p Mosco convergence ,pll -+ ,p [13, Chapter IV,
Section 8], [48] implies (H3). We point out that our stability theory does not need
a uniform equimonotonicity assumption like (Ass III) with [36] or an equicoercivity
assumption like (Ass IV) with [36].
Altogether instead of the problem (P) we consider the perturbed problem of finding

(P,,)
By the existence theory discussed in the previous section, also solvability of the per-
turbed problems (P,,) can be guaranteed.
Now we can present our basic stability result extending Theorem A in the classical
paper of Mosco [36].
 129

Theorem 3 Let the seminorm 1·1 and "P,>',K satisfy (AI) and (AS). Let the func-
tionals "Pv satisfy (A2) and (A9) uniformly with respect to v E IN together with (A4)·
Suppose that the sets K, Kv and the functionals "P, "Pv satisfy (HI)-(H9). Then there
exists a subsequence {u v.} weakly converging to u, where uv• is a solution of (Pv.)
and u is a solution of (P). If (A9) holds with 10 = II = 0, then in addition we have
lim lu v - ul = O. If moreover the solution u of (P) is unique, then v_oo
1(_00 K
lim lIu v - ull = 0
holds.

Proof. We divide the proof in five parts. We first show a priori estimates for {u v },
before we can establish the convergence results. In the following co, CI, • .• are generic
positive constants.
(1) An I· I-estimate for {u v }. Fix Vo E K, according to (H2) let {v v } be an
lim IIv v - voll = 0 and Vv E Kv for v > vo = vo(vo).
approximating sequence to Vo ,i.e. v_oo
Then with U v, a solution of (P v), we have

By (A2), we get for v > Vo using (3.1) that

lu v - iii g(lu v - iiI) ~ bo + blluv - iii + l"Pv( U v, vv) - "Pv( uv, ii)1 + >.( U v - vv) .

Since Vv converges to vo, >.(vv) and Ivv - iii are bounded. By (A3) we can estimate
Xv := U v - ii by

Ixvlg(IXvl) < co(bo"o,ii,>') + cI(b!'II)lxvl +c2G(lxvl) + >'(xv) (3.2)

~ Co + C2 G(lxvl) + c3l1xvll. (3.3)

(2) Norm-boundedness of {u v }. Here we modify a contradiction argument which in

the case of a bilinear form goes back to Fichera [16] and Stampacchia [45] and is used in
existence proofs for semi coercive variational inequalities (see e.g. [23, p. 646-648]'[19]);
another variant can be found with [4,20]. We assume there exists a subsequence {un} :=
{U Vn } such that lIunll--+ +00 (n --+ (0). With Zn := IIxnll-1xn = lIu n - iill-1(u n - ii) in
the reflexive Banach space V, we can extract a subsequence, again denoted by {zn},
that converges weakly to some £ E V. In virtue of (3.3), we get

(3.4)

We claim that IZn I --+ O. Assume not. Then for a subsequence IZn.1 ;::: c > 0, hence
:= Ixn.1 = IIxn.lllzn.1 --+ 00. By the continuity of I . I
tk

Therefore we obtain from (3.4) that

This leads by (A4) to a contradiction, proving our claim. Thus we have IZnl--+ 0, Ilznll =
1 and Zn weakly convergent to £. By (AI), we can extract a subsequence, again denoted
by {zn}, that converges strongly to £. It follows that 11£11 = 1, in particular £ =I O.
Now we claim that £ E ac (Y n K) = Y n acK. Since by continuity of 1·1, £ E Y
and since ii E K, we only have to show that, for any fixed s > 0, ii + s£ belongs to K.
130

Because of JlunJi -+ +00, there exists an index nI such that Jlu n - YII > s for n ~ nI.
On the other hand, y E K"n and Un E K"n. By convexity for n ~ nI,

and by (HI) Y + s2 belongs to the set K, proving our claim.

If Y n ac K is bounded, then the existence of 2 E Y n ac K with 2 =I- 0 leads already
to a contradiction. Otherwise we conclude from (3.2)

Concerning the behaviour of IXnl = IUn- yl there are two cases.

Firstly, for some subsequence {un,}, IX", I is bounded. Then -G(lxnD and { ... } in
the left hand side of (3.5) are bounded below by some constant, which is not necessarily
positive. By Ilun,lI -+ 00, the left hand side of (3.5) can be estimated below by some
zero sequence for the subsequence considered.
Secondly we have Iunl -+ 00, hence IXnl -+ 00. Then in view of (A4), {... } is
positive for large enough n.
Thus in both cases, the lim inf of the left hand side is nonnegative, and in the limit
we get
A(2) = AI(2) + A2(2) ~ o.
By (2.5) in the assumption (A6) and by 121 = 0, we arrive at

for 2 E Y n ac K , 2 =I- o. A contradiction to (2.4) is finally reached.

(3) Any weak limit point u* 0/ {u,,} solves (P). By the preceding step, there exists
a subsequence, again denoted by {un}, such that Un ~ U*. By (HI), U* belongs to K.
We show that u* is a solution to (P). Fix Vo E K, according to (H2) let {v,,} be an
approximating sequence to Vo ,i.e. lim Jlv il - voJl = 0 and v" E K" for v > Vo = vo( vo).
Then with Un, a solution of (P IIn ), we have for n > nllQ
" .... 00

Since CPn is monotone, if follows

In virtue of the limit condition (H3),

cp(vo,U*) :::; A(U· - vo)

holds for any Vo E K. By Lemma 2 we conclude that for any v E K

cp(U·,v) ~ A(V - u*).

(4) Convergence with respect to 1·1. Here we modify the arguments of Glowinski
[18, p. 11]. Suppose that (A3) holds with /0 = /1 = O. Then as above, let Un solve
(P"n) and let Un converge weakly to u*, a solution of (P). By (A2),
 131

for any vn E K n , then by (A3)

Since 'P(u·,·) is convex and lower semicontinuous, hence [25, §4.3, p. 153] weakly lower
semicontinuous and 'P( u·, u*) = 0,

limsup [-'P(u*,u n )] ~
n_oo
o.
Now fix Vo E K, according to (H2) let {v~} be an approximating sequence to Vo. Then
Vn= V~n converges strongly to Vo and Un - Vn converges weakly to U* - Vo. Hence
lim sup
n--+oo
IVn - u*1 ~ lim
n--+oo
IVn - vol + Ivo - u*1 = Ivo - u*l·

Moreover, G(luni) is bounded above, say by c* > O. therefore (3.6) entails in the limit,
for any Vo E K,

o ~ lim sup
n_oo
IU n - u*1 g(lu n - u*i) ~ c*lvo - u*1 + A(U* - vo).

The choice Vo = u* leads to the desired I . I-convergence.

(5) Convergence with respect to 11·11. Let il be the unique solution of (P). Assume
there exists a sequence {un} such that Un solves (P ~n) and lIu n - illl :::: fJ > o. By part
(2), Ilu n - illl is bounded and therefore we can subtract a subsequence, again denoted
by {un} such that Un - il converges weakly to some w E V. By part (3), il + w solves
(P)j hence by uniqueness w = Ov. Moreover by part (4), IU n - ill - t 0 as n - t 00. In
virtue of (AI), there exists a subsequence {un.hEN such that lIu n• - illl - t 0, and a
contradiction is reached. 0

4. AN APPLICATION TO DISTRIBUTED MARKET EQUILIBRIA

WITH BOUNDS

In this section we formulate a finite dimensional variational inequality to describe

a constrained equilibrium of spatially distributed economic markets with given bounds
on prices and transportation fluxes thereby extending a recent "disequilibrium" model
due to Nagurney and Zhao [37]. Then we address the resulting question of stability
with respect to the given bounds and derive a stability result using the abstract theory
of the preceding section.

The Market Model

Let us consider a single commodity that is produced at n supply markets and

consumed at m demand markets. There is a total supply gi in each supply market i,
where i = 1, ... ,n. Likewise there is a total demand Ii in each demand market j, where
j = 1, ... , m. Since the markets are spatially separated, Xij units of the commodity are
transported from i to j. Introducing the excess supply Si and the excess demand tj we
must have
m
gi LXij +Si, i = 1, ... ,nj (4.1)
j=l
n
fj LXij +tj, j = 1, ... ,mj ( 4.2)
i=l
132

Moreover the transportation from i to j gives rise to unit costs 7rij. Further we associate
with each supply market i a supply price Pi and with each demand market j a demand
price qj. As in the recent model of Nagurney and Zhao, we assume there is given a
fixed minimum supply price Pi ~ 0 ('price floor') for each supply market i and also
a fixed maximum demand price qj > 0 ('price ceiling') for each demand market j.
These bounds can be absent and the standard traffic network equilibrium model due
to Dafermos ([9], [10]) results, where the markets are required to be cleared, i.e.
Si =0 for i = 1, ... , nj tj =0 for j = 1, ... , m
are required to hold. Since Si ~ 0 and tj ~ 0 are admitted, the model is also termed a
disequilibrium model. In addition to the model of [37] we also include upper bounds
Xij > 0 for the transportation fluxes Xij.
Assuming perfect equilibrium the economic market conditions take the following
form

Si > 0 * Pi = Pi ' Pi> Pi * Si = 0 i = 1, ... ,nj (4.3)

tj > 0 * qj = qj , qj < qj * tj = 0 j = 1, ... ,mj (4.4)
if Xij = 0
> q'
Pi + 7rij { ::q;
if 0 < Xij < Xij i=l, ... ,njj=l, ... ,m. (4.5)
:5 qj if Xij = Xij
The last condition (4.5) extends the well-known Samuelson [41] equilibrium conditions
in that Pi + 7rij < qj can occur because of the flux constraint Xi; :5 Xij. For notational
simplicity we group the introduced quantities in vectors omitting the indices i and j.
Thus we have the total supply vector 9 E Rn, the supply price vector P E Rn, the total
demand vector f E Rm, the demand price vector q E Rm, the flux vector X E RRm,
and the unit cost vector 7r E Rnm. As in unconstrained market equilibria ([9],[10]) we
assume that we are given the functions
9 = g(p),f = j(q),7r = *(x).
This is in contrast to [37] where the prices are assumed to depend on the total supply
g, resp. the total demand f, what leads to a somewhat simpler structured comple-
mentarity problem. In our constrained economic market model the feasible set for the
vectors u = [p, q, x] is given by the product set
n m n m
M:= II [Pi' 00) X II[O,qj] x II II[O,Xijj.
i=1 j=1 i=1 j=1

The Variational Inequality

Now we can characterize a market equilibrium u = (p, q, x) introduced above as a
solution to a Variational Inequality in the general nonlinear case.
Proposition 4 Suppose that for each i = 1, ... , njj = 1, ... , m there holds
qj=O*jj(q)~Oj Xij>O**ij(X»O. (4.6)
Then u = (p,q,x) E M satisfies the market equilibrium conditions (4.1) - (4.5), if and
only if u is a solution to
n m m n
EGj;(p) - E Xij)(Pi - Pi) - E(ij(q) - E Xij)(qj - qj)
i=1 j=1 j=1 i=1
n m
+L L(Pi + 1rij(X) - qj)(Xij - Xij) ~ 0, Vii = (p,q,x) EM. (4.7)
i=1 j=1
 133

Proof. Let u = (p,q,x) E M satisfy the market equilibrium conditions (4.1) - (4.5)
with
Si = ij;(p) - I>ij ~ 0, tj = /j(q) - I>ij ~ O.
i

If Pi= p., then Pi - Pi ~ 0 and the product in the first sum in (4.7) is nonnegative,
otherwi;~ by (4.3) Si = 0 holds and the product vanishs. By similar case distinctions
one obtains that each product in the second sum is nonpositive and each product in
the third sum is nonnegative. This proves (4.7).
Conversely let (4.7) hold. By the choices Pk = Pk, ifl = ql, Xkl = X kl for all indices k
except some fixed index i and for all indices 1 except some fixed index j one obtains
that in (4.7) all products in the first and in the third sum are nonnegative, whereas all
products in the second sum are nonpositive.
By the choice Pi> Pi (4.1) follows with Si ~ O. To verify (4.3) suppose for some i
that Si > 0 and Pi > p. both hold. Then the choice Pi = p. leads to a contradiction.
Now turn to the third sum. If Xij = 0, choose Xij = -iij > OJ if Xij = Xij, choose
Xij = OJ if Xij E (0, Xij) both choices are possible. This shows (4.5).
Let us consider the second sum. If qj E (0, qj]' choose ifj = O. Then obviously tj
given by (4.2) is nonnegative. Thus the case qj = 0 remains. Suppose tj < O. Then
n
LXij>f;=h(q)~O
i=l

by assumption. Therefore there exists some index i E {l, ... ,n}, such that Xij > O.
The already proved condition (4.5) entails

Pi + 7f'ij ~ qj =0.
Hence in virtue of Pi ~ p. ~ 0 we obtain 7f'ij = *ij(X) ~ 0, what contradicts that by
assumption *ij(X) > 0 h~lds because of Xij > O.
To verify finally (4.4) apply an analogous argument as in the proof of (4.3): tj > 0
and qj < qj cannot hold simultaneously. In view of the just proved bound tj ~ 0 this
proves (4.4). 0
The Variational Inequality (4.7) leads to the following functional

cp(u,u) := tUj;(p) - 'Exij)(Pi - Pi) - 'E(h(q) - tXij)(ifj - qj)

i=1 j=1 j=1 i=1
n m
+ L L(Pi + *ij(X) - qj)(Xij - Xij) ,
i=lj=1

where u = (p, q, x) ,u = (p, if, x) E M. vObviously this functional is monotone, if and

only if the mappings 9 : R~ --+ Rn , - f : R~ --+ Rm ,* :
R~m --+ IR nm are monotone
op~rators, what we assume in the following. Moreover, we require that 9 (but neither
*'
- f nor !) is uniformly monotone with some constant 7]1 > 0 and some exponent
a > 0 in the sense that

holds for all pI, p2 E R~. Also we require that 9 (but neither - / nor !) is Lipschitz *'
continuous with some constant 7]2 > 0 and an exponent f3 in the sense that
134

holds for all p, pI ,p2 E R~.

Now we define the seminorm lui = IIpll for all u = (p,q,x) E Rn+m+nm. Hence
the semicoercivity assumptions (A2) , (A3) are satisfied, since the box constraints
in M give some positive upper bounds 1'0,1'1. Also the assumption (A4) holds for
o < (3 < 1 + Q. With the strong topology coinciding with the weak one in finite
dimensions (A1) holds trivially. Since the kernel space Y = {OJ x IRm x IRnm and the
asymptotic cone ac M = Rn x {OJ intersect trivially, (A5) is satisfied.
For simplicity we assume that p = 0 ,g(O) = 0 ,i(O) = 0 ,*(0) = o. Then we can
choose y = 0 E M such that cp(y,.) == o.

Stability with respect to the Bounds

Now we vary the bounds in the box constraints of the set M and introduce the
convex closed sets
n m n m
MCv):= II[O,oo) x II[O,~v)J x II II[o,xli)J,
i=1 ;=1 i=1 ;=1

where we suppose ~v) > 0, xli) > o. Then jj = y = 0 E MCv) for all v E IN.
This leads to the perturbed variational problem: Find u Cv ) = (pCI') , qCv), xCv» E M(v)
such that
n m m n
L(Yi(pCv» - L xlj»(p. - plv» - L(/i(qCv» - L xli»(q; _ q~v»
.=1 ;=1 ;=1 .=1
n m
+L L(plv) + *i;(X Cv » - qJv»(x.j - xli» ~ 0, (4.8)
.=1 ;=1
Then we have the following result.
Corollary 5 Let cp, M and the sequence {MCv)}vEN be given as above in this section.
Suppose that the bounds q~v), xli) converge to qj, Xij respectively for all i = 1, ... ,n; j =
1, ... ,m. Then there exist.! a .!ub.!equence {uCvk)hEN that converge.! to a solution u of
(4.7). If the solution u of (4.7) is unique, then 1'_00
lim lIu Cv ) - ull = 0 holds.
Proof. By the assumed convergence of the bounds, the sets MCv) clearly Mosco-
converge to M. Therefore the claimed convergence of a subsequence is a direct con-
sequence of Theorem 3. In case of uniqueness, the convergence of the entire sequence
follows from an obvious contradiction argument. 0
To conclude this section let us remark that is not hard to extend this Corollary
5 to include perturbations of the mappings g, i, *.
One considers mappings gCv) that
converge to g in the sense that pCI') ~ p implies gCv)(pCv» ~ g(p) as v ~ 00; likewise for
i Cv ), *Cv). According to Theorem 3 it is enough to assume that the mappings gCv) satisfy
the same uniform monotonicity condition and the same Lipschitz continuity condition
as y, whereas the mappings -1<1'), *Cv) are allowed to be arbitarily monotone. - On
the other hand, if not all demand prices qj or fluxes Xi; are bounded above, then the
seminorm 1·1 has to be modified and also the mappings - i, *
have to satisfy appropriate
uniform monotonicity and Lipschitz continuity conditions. The details are omitted.

5. AN APPLICATION TO ELLIPTIC VARIATIONAL INEQUALITIES

MODELLING STATIONARY CONTACT PROBLEMS

In this section we consider a more concrete elliptic Variational Inequality of the

first kind, respectively of the second kind [18J in an appropriate Sobolev space. By
 135

invoking Green's formula these problems can be seen to be the weak formulation of
a boundary value problem involving the p-harmonic differential operator as a typical
nonlinear monotone elliptic operator and the unilateral Signorini boundary condition,
respectively nonclassical boundary conditions describing given friction on the boundary
[14]. Thus both elliptic Variational Inequalities can be considered as scalar models of
the stationary unilateral contact problem [24] in solid Mechanics without and with
given friction, respectively, where a certain nonlinear behaviour of the material [39] is
admitted.
First we introduce the p-harmonic Signorini boundary value problem and collect
some technical tools. Then we apply our abstract stability theory to its variational
formulation, a semi coercive elliptic Variational Inequality, and investigate its stabil-
ity under perturbations of the coefficients of the differential operator and also of the
boundary obstacles. Finally we discuss a related Variational Inequality of the second
case and apply the abstract stability theory to perturbations in the coefficient in the
differential operator only.

A P-Harmonic Signorini Boundary Value Problem

Let us address the following nonlinear elliptic boundary value problem:

-div (0IiVuIl P- 2 Vu) =f in Q,

u~x,
au
0liVullP-2an ~h,
au
(u-x)(0IiVuIlP-2an -h)=O onr,

where p > 2, Q C 1R2 is a Lipschitz domain with its boundary r

= aQ, IIVull 2 =
(c3tu)2 + (~u)2 and o,X,!,h are given data. Our purpose is here to study stability
with respect to the coefficient 0 and with repect to the boundary obstacle x. But
before we can apply Theorem 3 to settle this stability issue, we have to discuss several
more technical questions, in particular we need three lemmata.
By invoking an appropriate Green's formula (see e.g. [3, Chapter 18]) we can
see that the variational formulation of the boundary value problem considered is the
Variational Inequality (P), where for all u, v E V = Wl,p(Q)

cp(u,v) .- a(u,v - u), a(u,v):= 10 0IiVuIl P- 2Vu. Vvdx, (5.1)

-X(u) .- 10 f(x)u(x)dx +1h(s)u(s)ds, (5.2)

K .- {u E Wl'P(Q) lu ~ X on r}. (5.3)

In virtue of Sobolev's embedding [38, Theorem 3.8, p. 72] we have Wl,p(Q) C C(Q}.
Therefore the restriction of u on r is pointwise defined in the definition of K and
accordingly we assume X E C(r). To render -X a continuous linear form it suffices to
assume that f E L2(Q) and h E Ll(r). Finally to have the nonlinear form a well
defined we assume that the coefficient function 0 E LOO(Q) with 0 ~ 00 a.e. on Q for
some constant 00 > O.
As a continuous seminorm on V we have here

Ivll,p = {fo IIVv(x)IIPdxP/p·

Since
136

it follows
IIvlh,p :5 Ivh,p + IIvllo,p :5 2l1 vlh,p .
Hence and by the compact imbedding WI,p(n) c LP(n) ([38, §6.1], [1, Theorem 6.2])
Lemma 1 (ii) entails that the assumption (AI) is satisfied.
Referring to the proof of [8, (5.3.20), Theorem 5.3.3] we have a positive constant
CI such that for all u, v E V

-{'fl(u,v)+'fl(v,u)} = a(v,v-u)-a(u,v-u)
~ cI(ao)lv - ulr,p ,

and (A2) is satisfied with get) = cl t p - 1 • Since moreover with some positive constant C2

l'fl(u,v) - 'fl(u,w)1 = la(u,v - w)1

:5 c2(lI a ll oo )l u IGI Iv - Wll,p ,
(A3) is satisfied with G(t) = C2tp-1 and with /0 = /1 = o. Thus also (A4) is satisfied.
To characterize the subspace Y we insert the following
Lemma 6 There holds
Y = {v E W 1 ,p(n) : Ivh,p = O} = {v : n -+ IR constant} .
Proof. Consider the mapping T : WI,p(n) -+ [LP(n)]N (here N = 2) given for any
u E W 2,p(n) by Tu = {Daullal = I}. T is linear and continuous with T-I(O) closed.
Since in the subspace
x = COO(n) n Wl,p(n)
the distributional derivatives coincide with the classical ones, it follows that

where TIo is the subspace of order zero polynomials, i.e. of constant functions on
n. Since TIo is isomorphic to R, hence finite dimensional, TIo is a closed subspace in
WI,p(n). Denoting the orthogonal subspace with respect to the duality by ..L. and using
the reflexivity of Wl,p we obtain
TIo = [X n T-I(O)].L.L = {cl [X.L + T-I(O).L]}.L j
since X.L = {OJ by density of X in Wl,p (e.g. [17], Theorem 7.9) finally
TIo = T-I(O).L.L = T-I(O) = Y

is obtained. o
In virtue of this lemma one can prove that the quotient norm of the quotient space
V/Y is equivalent to I . h,p, hence also Lemma 1 (i) guarantees that the assumption
(A 1) is satisfied.
Now we choose y as a large enough positive constant function. Then y E K n Y
with 'fl(Y,·) == o. Also by the preceding Lemma 6, A = Al satisfies (A5), if and only if
the following condition

(*) 10 f(x)dx + 1r h(s)ds < 0

is met. It turns out that not only existence, but also uniqueness of the solution to the
semi coercive Variational Inequality problem (P) (in the settings above) hinges on this
condition (*).
 137

Lemma 7 Suppose, (*) hold3. Then there exists (at most) one solution to the Varia-
tional Inequality (P), where <p, A, K are given by (5.1), (5.2), (5.9), respectively.
Proof. Suppose, there exist two distinct solutions '1£1 and '1£2 to (P). Then (A2) implies
lUI - '1£21 g(IUl - '1£21) ~ -<p( '1£1, '1£2) - <p( '1£2, ud
~ -A(U2 - '1£1) - A(UI - '1£2) = O.
Therefore '1£1 - '1£2 E Y, and by Lemma 6, U2(X) = Ul(X) + r (for a.e. x E n) with some
r E R\{O}. Since by monotonicity, the two solutions '1£1 and '1£2 fulfil

o ~ <P(Ul,U2) - A(U2 - '1£1) ~ -{<P(U2' '1£1) - A(UI - U2)} ~ 0,

and

holds, we obtain
A(U2 - ud = r{fo fdx + 1r hds} = 0,
contradicting (*). o
For our announced stability result we do not only make use of uniqueness, but to
prove Mosco convergence in the unilateral constraints we need also a density result,
already given in the Hilbert space case by Glowinski [18, Lemma 4.2], now to be
presented in the general case 1 ~ s ~ 00. For that latter result we exploit the lattice
structure of W = w 1 .S(n). Namely, since n is supposed to be a Lipschitz domain,
w 1·"(n) is a Dirichlet space ([3, Theorem 5.23], [28, Corollary A.6]) in the following
sense: Let 8 : IR -+ R be a uniformly Lipschitz function such that the derivative 8'
exists except at finitely many points and that 8(0) = OJ then the induced map 8* on
W given by w E W -+ 80 w is a continuous map into W. In particular, the map
w E W -+ w+ = max(O,w) = Hw + Iwl) is a continuous map into W. Note that
w~'"(n) = C8"(n) is also a Dirichlet space without imposing a regularity assumption
on r.
Lemma 8 Let W = w 1·"(n) (1 ~ s ~ (0). Under the above assumptions on n,
k:={wEW:w~O a.e.onn}
satisfies
clw[k n C""(fi)] =k .
Proof. Since n is a Lipschitz domain, for any v E W there exists an extension ([38,
3.7, 3.8]) v E W 1 '"(RN ) = C8"(IRN) (here N = 2) such that vln = v. Since Wl'"(R N)
is a Dirichlet space, Ivl is also an extension of any v E k and coincides with v on n.
Hence for any v E k, we can assume the extension v ~ 0 on RN. Now using Friedrich's
regularization, we introduce

iic(x) = c~ iRf Nw(x-y)v(y)dy

c
(XERN,c>O),

where wE Cr(RN ) satisfies

JR Nw(y)dy = 1, w ~ 0.

Thus we obtain Pc E Cr(RN ) , Pc ~ 0 on RN and limc-+o Pc = v stronly in Wl'"(RN).

Now we define Pc by Pc = Pc Ifi and arrive at Pc E C""(fi), Pc ~ 0 on fi and limc-+o Pc =
v strongly in W 1'"(n). 0
138

Stability with respect to the Boundary Obstacles and Coefficients

Now we vary the coefficient a in the elliptic differential operator and the boundary
obstacle X in the constraint of the set K and introduce

c,ov(u,v) .- av(u,v - u), av(u,v):= 10 avIlVuIlP-2Vu. Vvdx,

Kv .- {u E W 1 ,P(11) I u ~ Xv on r} ,

where Xv E C(r) and a v E LOO(11) with a v ~ 0'0 a.e. on 11. Thus we arrive at the
perturbed Variational Inequality problem (P v). For our convergence analysis we need
extensions X, Xv E V = W 1 ,P(11) of x, Xv' Then we can present the following
Theorem 9 Let a v -+ a in Loo(11) and Xv -+ X in V as v -+ 00. Suppose, the
condition (*) is satisfied. Then the Mlutions U v of the perturbed problems (P v) converge
strongly in Vas v -+ 00 to the solution u of the p-harmonic Signorini problem (P).

Proof. Looking at the preliminaries in the preceding subsection we only have to verify
the hypotheses (HI) - (H3) and apply Theorem 3.
Verification of (HI). Let a sequence {Vn}nEN be weakly convergent to v E V, where
VnE Kiln (Yn E IN). Since V is compactly imbedded in CO(IT) [38, Theorem 3.8J, it
follows for the trace operator, on r that after extracting a subsequence again denoted
by {v n }
(5.4)
Moreover by assumption, using X = ,X, XII = ,XII

(5.5)

(5.4) and (5.5) together imply that

lim IIbvn
n--+oo - Xlln ) - (,v - x)lIv"'(r) = O.
Thus ,VII - XII ~ 0 pointwise on r (Yv E IN) implies that for all t/J E CO(r) , t/J ~ 0

irb v - X) t/J ds ~ 0,
hence ,v - X ~ 0 on r or v E K. This proves (HI).
Verification of (H2). We claim that w- E W~'P(11) for any wE K - X. Indeed, since 11
is a Lipschitz domain, by density [38, §1.3J there exist Wj E Coo (IT) (j E IN) such that
W = limj--+oo Wj strongly in W. On the other hand the trace operator, : W -+ Ll(r) is
continuous [38, Theorem 4.6, Corollary 6.31 and by Egorov's theorem L1-convergence
implies pointwise convergence a.e. for some extracted subsequence which is again
denoted by {Wj}. Hence for any fixed c > 0, there exists jo E IN such that for all j ~ jo
we have
-,(Wj)(x) = -w;{x) :::; c (Yx E r\Nc:) ,
where meas(Nc:) < c. Hence

0:::; wj(x) = max(O, -w;(x» :::; c (Yx E r\Nc:).

By continuity of W -+ W- in W and once more by the continuity of the trace operator

we conclude that ,(wn converges to ,(w-) pointwise a.e. on r, hence 0:::; ,(w-) :::; c.
Since c > 0 is arbitrary, the claim is proved.
 139

Now let v E K be arbitrary. Then w = w+ - w-, where w := v-X. By the

claim proved above w- E W;,p(n), hence there exist pv E C8"(n) (1/ E IN) such that
w- = lim v_ oo pV' According to Lemma 8, there exist U v E coo(IT) (1/ E IN) such that
U v 2: 0 on n and lim v_ oo U v = w+ E k. Now define Vv = +
Xv U v - pv (1/ E IN). Then
lim v_ oo Vv = v, U v - pv E Coo(n) and

hence Vv E Kv (Vv E IN). Thus (H2) is satisfied.

Verification of (H3). Let {vv }VEN , {wv }vEN be sequences in V such that Vv --> v and
Wv ~ w in V. By the uniform boundedness principle, IIwvll :S Co and IIvvll :S Co
(Vv E IN) for some Co > O. By the definition of cp, cpv

ICPv(vv,wv)-cp(v,w)1 :S lav(vv,wv-vv)-a(vv,wv-vv)1
+ la(vv,wv - vv) - a(v,wv - vv)1
+ la(v, (wv - vv) - (w - v))1
< in I [av - a]IIIVvvl p- Vvv . [Vwv - Vvv]ldx
2

+ in I a UIVvvIl Vvv -IIVv Il P- Vv]' [Vwv - Vvv]ldx

P- 2 2

+ Iin a IIVvll Vv . [V(Wv - vv) - V( w - v)] dxl .

p- 2

By the estimate [8, (5.3.21),Theorem 5.3.3] the first summand above is bounded by
C2(2cO)Pllav - all oo and the secondsummand is bounded by Ilallooc2(2co)P-lllvv - vlh.p,
hence both convergent to zero. Since (w v - vv) ~ (w - v) and

gives a continuous linear form, the third summand also converges to zero. Therefore
(H3) is satisfied. 0

Some Remarks. By the proof above we established Mosco convergence of unilateral

convex sets in V = W1,p(n). A related result in W;,p(n) is due to Boccardo and
Brezis [7] that applies to coercive domain obstacle problems, but does not apply to
the semicoercive Signorini problem considered here. - Although the studied Signorini
problem can also be formulated as a minimization problem (involving the functional
Ilzlli,p), we did not use any minimization argument, but only relied on monotonicity
arguments instead. Thus by our convergence analysis even stationary problems in
absence of a potential can be treated where the Variational Inequality formulation gives
a variational formulation of a boundary value inclusion problem (see [14, Chapter 1] for
applications involving semi-permeable media). However, the uniqueness of the solution
may be lost (see [14, §1.7.3.3]). - Other unilateral boundary value problems involving
a monotone differential operator with Loo coefficients or the related pseudo-Laplacian
(see [42]) can be similarly treated. In particular, the analysis simplifies for the obstacle
problem under Dirichlet boundary conditions. If both a lower and an upper obstacle
are present the asymptotic cone of the corresponding K reduces even to {O} (compare
also [28, §III.3] for the theory of "semilinear" boundary value problems).
140

An Application to a Elliptic Variational Inequality of the Second Kind

Now let "Y = r l U r 2, where the open parts rl,r 2 have positive measure and are
disjoint. Here we consider the following elliptic boundary value problem:

-div(adIVuIl P - 2Vu) f in n,
au
d(u) := aliVullp-2 an = h on rl ,

Id(u)l:::=;k,d(u)+klul = 0 onr 2 ,

where as above p > 2, IIVull2 = (al u)2 + (~u)2 and a,J,h,k are given data. Now
hE £l(rl ) and k E Loo(r 2) is given such that k(x) ~ ko > 0 a.e. on r 2 . Note that
the latter equation du + klul = 0 in the nonclassical boundary conditions involving the
conormal derivative d( u) is equivalent to

{ Idl < k =} u =0, }

Idl = k =} u = -pd for some p:::: 0,

what reflects Coulomb's law of friction [14]. These nonclassical boundary conditions
give rise to the positively homogeneous, sublinear, hence convex functional

tft(v) = Jklvlds.
r2

Thus the definition of'P in (5.1), repectively of A in (5.2) modifies to

'P( u, v) .- a(u,v-u) +tft(v) -tft(u),

A(U) ._
in
f f(x)u(x)dx + ir,
f h(s)u(s)ds,

whereas now simply K = V. Thus again the assumptions (AI) - (A4) are satisfied,
and A = Al satisfies (A5), if and only if the condition (*) is modified to

inf f(x)dx + ir,

(*) f h(s)ds < 0
However, here uniquenes is a more delicate question.
Now consider similar as above perturbed functionals 'P" that arise from perturbed
coefficients a". Then (HI) and (H2) hold trivially, whereas the continuity and the weak
lower semicontinuity of tft together with the analysis for the nonlinear forms a", a in the
proof of Theorem 9 show that also (H3) is satisfied. Thus we arrive at the following

Corollary 10 Let a" -+ a in Loo(n). Suppose, the condition (*) is satisfied. Then
there exists a subsequence {U"k hEN of the solutions u" of the perturbed problems (P,,)
that converges weakly in V as v -+ 00 to a solution u of the p-harmonic boundary
value problem (P). If (P) has an unique solution, then norm convergence of the entire
sequence holds.

One can conjecture that the abstract stability theory also applies to perturbations
of the convex functional tft by studying the Mosco convergence of the epigraph sets by
means of arguments similar to the proof of Theorem 9.
 141

REFERENCES
[I] R.A. Adams, "Sobolev Spaces," Academic Press, New York, 1975.
[2] H. Attouch, "Variational Convergence for Functions and Operators," Pitman, Boston, 1984.
[3] C. Baiocchi and A. Capelo, "Variational and Quasivariational Inequalities," Wiley, Chichester 1984.
[4] C. Baiocchi, F. Gastaldi and F. Tomarelli, "Some existence results on noncoercive variational
inequalities", Ann. Scuola Norm. Sup. Pisa Cl. Sci., Ser. IV 13:617-659, 1986.
[5] E.Blum and W. Oettli, "From optimization and variational inequalities to equilibrium problems",
The Mathematics Student 63:123-145,1994.
[6] L. Boccardo and F. Donati, "Existence and stability results for solutions of some strongly nonlinear
constrained problems", Nonlinear Analysis 5:975-988, 1981.
[7] L. Boccardo and F. Murat, "Nouveaux resultats de convergence dans des problemes unilateraux",
in: "Nonlinear Partial Differential Equations and Their Applications," College de France Seminar,
Vol. II, H. Brezis and J.-L. Lions, eds., Research Notes in Math. No. 60, Pitman, pp.64-85, 1982.
[8] P.G. Ciarlet, "The Finite Element Method for Elliptic Problems," North-Holland, Amsterdam,
1978.
[9] S. Dafermos, "Traffic equilibrium and variational inequalities", Transportation Sc. 14:42-54,1980.
[10] S. Dafermos, "Exchange price equilibria and variational inequalities", Math. Programming 46:391-
402,1990.
[11] S. Dafermos and A. Nagurney, "Sensitivity analysis for the asymmetric network equilibrium prob-
lem", Math. Programming 28:174-184,1984.
[12] P. Doktor and M. Kucera, "Perturbations of variational inequalities and rate of convergence of
solutions", Czech. Math. J. 30(105):426-437, 1980.
[13] A. Dontchev and T. Zolezzi, "Well Posed Optimization Problems," Lecture Notes in Mathematics
1543, Springer, 1993.
[14] G. Duvaut and J.-L. Lions, "Inequalities in Mechanics and Physics," Springer, Berlin, 1976.
[15] W. Fenchel, "Uber konvexe Funktionen mit vorgeschriebenen Niveaumannigfaltigkeiten", Math. Z.
63:496-506, 1956.
[16] G. Fichera: "Boundary value problems of elasticity with unilateral constraints", in: "Handbuch der
Physik - Encyclopedia of Physics, Band VI a/2 Festkorpermechanik II," S. Fliigge, ed., Springer,
Berlin pp. 391-424, 1972.
[17] D. Gilbarg and N.S. Trudinger, "Elliptic Partial Differential Equations of Second Order," Springer,
New York, 1977.
[18] R. Glowinski, "Numerical Methods for Nonlinear Variational Problems," Springer, New York, 1984.
[19] J. Gwinner, "Nichtlineare Variationsungleichungen mit Anwendungen," Ph.D. Thesis, Universitat
Mannheimj Haag & Herchen, Frankfurt, 1978.
[20] J. Gwinner, "Discretization of semicoercive variational inequalities", Aequationes Math. 42:72-79,
1991.
[21] J. Gwinner and E.P. Stephan, "A boundary element procedure for contact problems in plane linear
elastostatics", RAIRO Mathematical Modelling and Numerical Analysis 27:457-480, 1993.
[22] J. Gwinner, A discretization theory for monotone semicoercive problems and finite element con-
vergence for p-harmonic Signorini problems, Z. Angew. Math. Meeh. 74:417-427,1994.
[23] P. Hess, "On semi-coercive nonlinear problems", Indiana Univ. Mat". J. 23:645-654,1974.
[24] I. HlavaCek, J. Haslinger, J. Netas and I. Lovisek, "Solution of Variational Inequalities in Mechan-
ics," Springer, Berlin, 1988.
[25] H.G. Jeggle, "Nichtlineare Funktionalanalysis," Teubner, Stuttgart, 1978.
142

[26) A. Kaplan and R. Tichatschke, "Stable Methods for Ill-Posed Variational Problems," Akademie
Verlag, Berlin, 1994.
[27) N. Kenmochi, "Nonlinear operators of monotone type in reflexive Banach spaces and nonlinear
perturbations", Hiroshima Math. J. 4:229-263, 1974.
[28) D. Kinderlehrer and G. Stampacchia, "An Introduction to Variational Inequalities and Their Ap-
plications," Academic Press, New York, 1980.
[29) P. Kosmol,"Optimierung und Approximation," de Gruyter, Berlin, 1991.
[30) J. Lagnese, "Perturbations in variational inequalities", J. math. Analysis Applic. 55:302-328, 1976.
[31) P.L. Lions, "Two remarks on the convergence of convex functions and monotone operators", Non-
linear Analysis 2:553-562, 1978.
[32) A. Maugeri, "Stability results for variational inequalities and applications to traffic equilibrium
problems" , in: Atti del Convegno Celebrativo dell. Centenario del Circolo Matematico di Palermo,
Supplemento ai Rend. Cire. Mat. Palermo, Serie II, 8, 1985.
[33) L. McLinden, "Stable monotone variational inequalities", Math. Programming 48:303-338, 1990.
[34) G.J. Minty, "Monotone (non linear) operators in Hilbert space". Duke Mati•. J. 29:341-346, 1962.
(35) G.J. Minty and W. Oettli, "Variational inequalities for monotone-convex functionals", Universitat
Mannheim, 1974 (unpublished).
(36) U. Mosco, "Convergence of convex sets and of solutions of variational inequalities", Advances in
Math. 3:510-585, 1969.
(37) N. Nagurney and L. Zhao, "Disequilibrium and variational inequalities", J. Comput. Appl. Math.
33:181-198,1990.
(38) J.Neeas, "Les methodes directes en theorie des equations elliptiques," Academia/Masson,
Prague/Paris, 1967.
(39) P.D. Panagiotopoulos, "Inequality Problems in Mechanics and Applications," Birkhauser, Basel,
1985.
(40) A.A. Pankov, "Discrete approximation of convex sets and the solution of variational inequalities"
(in Russian), Math. Nachrichten 91:7-22, 1979.
(41) P.A. Samuelson, "A spatial price equilibrium and linear programming", Amer. Econom. Rev.,
42:283-303, 1952.
(42) F. Scarpini, "Some nonlinear complementarity systems algoritms and applications to unilateral
boundary-value problems". Ann. Mat. Pura Appl. IV 123:185-202, 1980.
(43) M.J. Smith, "The existence, uniqueness and stability of traffic equilibria", Transportation Res.
13B:295-304, 1979.
[44) Y. Sonntag, "Convergence au Sens de U. Mosco: theorie et applications a I'approximation des
solutions d'im;quations", These, Universite de Provence, 1982.
(45) G. Stampacchia, "Variational Inequalities", in: "Theory and Applications of Monotone Operators,"
A. Ghizetti, ed., Edizione Odersi, Gubbio, pp. 101-192, 1969.
[46) F. Stummel, "Diskrete Konvergenz linearer Operatoren I", Math. Ann. 190:45-92,1970.
(47) F. Stummel, "Perturbation theory for Sobolev spaces", Proc. Royal Soc. Edinburgl. 73A:5-49,
1974/75.
[48) R. Wets, "Convergence of convex functions, variational inequalities and convex optimization prob-
lems", in: "Variational Inequalities and Complementarity Problems," R.W. Cottle, F.Giannessi,
and J.-L. Lions, eds., Wiley, Chichester, pp. 375-403, 1980.
[49) F. Wille, "Galerkins Losungsnaherungen bei monotonen Abbildungen", Math. Z. 127:10-16,1972.
A PRIMAL-DUAL PROXIMAL POINT ALGORITHM
FOR VARIATIONAL INEQUALITY PROBLEMS

Iwaoka Kouichiro 1.1, Fukushima Masao F.2 and Ibaraki Toshihide 1

1 Department of Applied Mathematics and Physics

Faculty of Engineering,
Kyoto University
Kyoto 606, Japan
2 Graduate School of Information Science
Nara Institute of Science and Technology
Ikoma, Nara 630-01, Japan
e-mail: fuku@is.aist-nara.ac.jp

Abstract: The proximal point algorithm is an iterative method for finding a zero
of a maximal monotone set-valued mapping, and has served as a fundamental method
for solving Variational Inequality problems and, in particular, convex optimization
problems. In this paper, we propose a new variant of the primal-dual proximal point
algorithm for solving monotone Variational Inequality problems. The proposed method
adopts a Gauss-Seidel-like procedure to solve a subproblem in each iteration, which
updates primal variables and dual variables alternatively. We prove the convergence
of the proposed method and show that our method is particularly effective for Varia-
tional Inequality problems with certain separable structure, such as traffic assignment
problems.

Key Words: Variational Inequality problem, primal-dual proximal point algorithm,

parallel computation, traffic assignment.

1. INTRODUCTION

Let F be a maximal monotone mapping from RN into itself and X be a nonempty

closed convex subset of RN. The Variational Inequality problem can be stated as
follows:

[VI] Find x· E X such that (F(x·), x-x·} 2: 0 for all x EX,

where ( " . ) denotes the inner product in R,N. We suppose that X is defined by

(1.1)

143
144

where g(x) = [gl(X),g2(X), ... ,gM(X)]T with gi: RN -+ R being continuously differen-
tiable and convex, hex) = [h 1 (x),h 2(x),···,h p (x)]T with hj : RN -+ JR being affine,
and S denotes a closed polyhedral convex subset of RN .
Over the years, various solution methods for problem VI have been developed [7].
In this paper, we propose a method for solving problem VI based on the proximal
point algorithm [2,14]. Our method turns out to be particularly effective for problem
VI with certain separable structure, such as traffic assignment problems.
The proximal point algorithm is a method for finding a zero of a maximal monotone
set-valued mapping T :

Find v* such that 0 E T(v*).

This method is an iterative method such that in iteration k one solves the equation
1
oE T(v) + C(k) (v - v(k»),

where v(k) is the current iterate and c(k) is a positive parameter. The proximal point
algorithm can be applied to problem VI in the following manner. Let ,\ E JRM and
1r E R P be Lagrange multipliers vectors associated with the inequality and equality
constraints in (1.1), respectively, and define the mapping T by

T(x,,\,1r) = { (~, (,,.,) E R N +M+P I ~ E F(x) + Vg(x),\ + Vh(x)1r + Ns(x),

( E -g(x) + NRr('\), ,., = -hex) }, (1.2)

where R~ denotes the nonnegative orthant of RM, Ns(x) is the normal cone to S at
x, and NRM('\) is the normal cone to R~ at ,\. It is known [15,16] that the mapping
+
T is maximal monotone. Also, it is shown in [15,16] that any zero of the mapping T is
a solution of problem VI if the Slater's constraint qualification is satisfied, i.e., there
is an x E X satisfying g(x) < o. Then, the proximal point algorithm for the mapping
T is to solve at each iteration the following system of equations:
1
o E F(x)+Vg(x)'\ + Vh(x)1r + Ns(x) + c(k)(x-x(k»), (1.3)

o E -g(x) + NRr('\) + )k)('\ - ,\(k»), (104)

1
o = -h(x) + c(k)(1r-1r(k»). (1.5)

Note that ,\ and 1r satisfying (1.4) and (1.5) are expressed as

A(x , ,\(k) , c(k») , (1.6)

II(x , 1r(k) , c(k») , (1.7)

where A(x,,\,c) and II(x,1r,c) are the vectors which have the following functions as
the components, respectively,

Ai(x,,\,c) max{,\i+cgi(X),O}, i=l,···M, (1.8)

IIj(x,1r,c) = 1rj+chj(x), j=l,···,P. (1.9)

To solve problem VI, Rockafellar [15,16] has proposed an iterative procedure called the
proximal method of multipliers. The proximal method of multipliers is a modification of
the method of multipliers or the augmented Lagrangian method developed for nonlinear
 145

optimization [1,8,12,13) and Variational Inequalities [5,10). This method belongs to a

class of primal-dual proximal point algorithms, which have been applied to convex
programming in various manners [9,15,16,18). In the proximal method of multipliers,
One solves approximately (1.3), with A and 7f' expressed in terms of x from (1.6) and
(1. 7). In other words, one finds x satisfying the equation
1
o E F(x)+V'g(x)A(x,A(k),e(k»+V'h(x)II(x,7f'(k),e(k»+Ns(x)+ e(k)(x-x(k». (LlO)

Unfortunately the equation (1.10) is not separable even though F, g, hand S are sepa-
rable, since V' g(x) A(x, A(k), elk»~ and V'h(x) II(x, 7f'(k) , elk»~ usually involve cross terms.
This difficulty is common to the method of multipliers in nonlinear optimization, for
which a number of modifications have been proposed in order to restore the separabil-
ity [4,9,11,17). In this paper, bearing in mind the applicability to separable problems,
we propose an approach to solve subproblems (1.3)-(1.5) of the primal-dual proxi-
mal point algorithm in a way different from [15,16). This approach, roughly speaking,
adopts a Gauss-Seidel-like procedure to solve approximately (1.3)-(1.5). In other words,
one first updates x for fixed A and 7f', then one improves A and 7f' by using the new data
of x. This procedure thereby enables us to make the most of a separable structure the
given problem may have.
This paper is organized as follows. In Section 2, we state a new variant of the
primal-dual proximal point algorithm. In Section 3, we prove the convergence of the
proposed method. In Section 4, we apply the method to traffic assignment problems.
Section 5 concludes the paper.

2. ALGORITHM

In this section, we propose a variant of the primal-dual proximal point method for
solving problem VI with the constraints defined by (1.1). The method is in essence the
proximal point algorithm for the mapping T defined by (1.2). Let SP(x, A, 7f'; c) denote
the subproblem to be solved for the current iteration point (X,A,7f') (see (1.3)-(1.5».

[SP(x, A, 7f'; c)] Find (x, X, 1f) such that

o E F(x) + V'g(x)A- + V'h(x) 1f 1
+ Ns(x) + -(x
e
- x),
- 1-
o E -g(x) + NRM(A)
+
+ -(A
e
- A),
1
0= -hex) + -(1f - 7f').
C

Note that the solution (x, X, 1f) of subproblem SP(x, A, 7f'; c) is uniquely determined. In
fact, using the mapping T defined by (1.2), (x, X, 1f) can be expressed as (x, X, 1f) =
(I + eTtl(x, A, 7f'). We first state a procedure, called SSP(x, A, 7f'; e, E), to find an ap-
proximate solution (z, Ti, 17) of subproblem SP(x, A, 7f'; c).

Procedure SSP (x, A, 7f'; C, E).

Input: a current iteration point (x,A,7f') and positive scalars c and E.
Output: an approximate solution (2, {l, I) of subproblem SP(x, A, 7f'; c).
Step 1: z(o) := x; 1'(0) := A(x, A, c); v(O) := II(x, 7f', c);
1:= O.
Step 2: Find z(l+l) such that

o E F(z(l+l» + V'g(z(l+1) 1'(1) + V'h(Z(l+l» vel) + NS(z(l+l» + ~(z(l+l) - x).

C
146

Step 3: If the stopping criterion

is satisfied, then go to Step 5.

Step 4: 1'(1+1) := A( Z(1+1), >., C); 1/(1+1):= II( z(I+1), 1T , C);
1:= 1+ 1;
go to Step 2.
Step 5: Z := z(l+1);
jl := 1'(1); v:= 1/(1);
exit.

Note that the functions A and II that appear in Step 4 are defined by (1.8) and (1.9)
in the previous section. Using this procedure, a primal-dual proximal point algorithm
for problem VI can be stated as follows. This algorithm uses two sequences of positive
parameters {elk)} and {elk)} such that

inf elk) > 0,

k

Algorithm PDPPA.
Step 1: Choose a starting point (x(O),>.(O),1T(O»;
k:=O.
Step 2: Obtain an approximate solution (x(k+ 1), >.(10+1), 1T(k+1» of subproblem
SP(x(k), >.0:), 1T(k) ; e(k» by executing procedure SSP(x(k), >.(10), 1T(k); elk) , elk) ).
Step 3: If (X(k+1), >.(k+1), 1T(k+1» = (x(k), >.(10), 1T(k», then stop. Else
k:= k + 1;
go to Step 2.

The termination condition (x(k+1), >.(10+1), 1T(k+1» = (x(k), >.(10), 1T(k» in Step 3 guarantees
that x(k+1) is a solution of problem VI. This fact will be proved in the next section.
Now let us consider a Variational Inequality problem with certain separable struc-
ture. Specifically, suppose that

S = So x ... x Sq x ... x SQ,

where Sq is a closed polyhedral subset of RNq, E~=o N q = N and x denotes the

Cartesian product. Moreover, suppose that

x =[ ~: l' F(x) =[ ~~=:: 1' q

y(x) = L:Yq(Xq),
q=O
q
h(x) = L:Aqxq - b,
q=O
xQ FQ(xQ)

where Fq: RNq --+ RNq are maximal monotone mappings, Yq: RNq --+ RM are mappings
whose components are continuously differentiable convex functions, and Aq and b are
P x N q matrices and an P-vector, respectively. Problem VI with the above structure
will be called a separable Variational Inequality problem (SVI). For problem SVI, the
equation to be solved in Step 2 of procedure SSP(x, >., 1T; e, e) can be decomposed into
(Q + 1) smaller equations as follows:
 147

For each q = 0,1,··· , Q, find z~/+1) such that

Note that these (Q+ 1) equations can be solved independently of each other. Therefore,
Algorithm PDPPA can take full advantage of the separable structure of problem SVI.

3. CONVERGENCE
In this section, we prove the convergence of the algorithm PDPPA proposed in
the previous section. First we establish a finite termination property of procedure
SSP(x,.x, 71"; c, f).

Proposition 1. Let h be represented as hex) = Ax - b. Suppose that 9 is Lipschitz

continuous with constant L9 > 0, that is, I g( x) - g( x') II ::; L9 II x - x'II for all x and
x'. If c is chosen smaller than (L; + p(AT A)tt, then procedure SSP (x, oX, 71"; c, f) halts
finitely, where p(AT A) denotes the spectral radius of AT A.
Proof. Let {z(l)}, {/l(l)} and {y(l)} be the sequences generated by procedure SSP(x,.x, 71";
c, f). Since z(l+1) is the unique solution of the equation in Step 2 and since /l(l) and y(l)
equal A( z(/), .x, c) and II( z(l), 71", c) as determined in Step 4 on the previous iteration, we
have for all z E S

Let (x, X, 'if) denote the unique exact solution of subproblem SP(x,.x, 71"; c). Then, since
X = A(x,.x, c) and 'if = II(x, 71", c), we have for all z
E S

1
(F(x) +\7g(x)A(x,.x,c) + \7h(x) II(x, 71", c) +-(x-x), z -x)
c
~O.

Substituting z = x in the first inequality and z = z(l+1) in the second inequality, and
then adding the two inequalities, we have

!c II z(l+1) - x 112::; -( F(z(l+I») - F(x), Z(l+I) - x)

-( \7g(z(l+I») A(z(l),.x,c) - \7g(x) A(x,.x,c), Z(l+I) - x)
- ( \7 h(z(/+1») II(z(l), 71", c) - \7h(x) II(x, 71", c), Z(l+I) - x). (3.1)

By the monotonicity of F, we have

( F(z(l+I») - F(x), z(l+1) - x) ~ o. (3.2)

By the differentiability and convexity of 9 and the nonnegativity of A, we have

( A(z(l), .x, c), g(x) - g(z(l+1») ) ~ ( A(z(l), .x, c), \7 g(z(l+1»)T (x - z(l+I») ) (3.3)

and
( A(x,.x, c), g(z(l+1») - g(x) ) ~ ( A(x, .x, c), \7 g(xf (z(l+1) - x) ). (3.4)
Since h( x) = Ax - b, we have
(3.5)
148

It follows from (3.1)-(3.5) that

II z(l+I) - x 112 ::; e ( A(x, A, e) - A(z(l), A, e), g(Z{l+l») - g(x) )

+e( II(x,7I',e) - II(z(l),7I',e), A (z(l+I) -x)).

From the Cauchy-Schwarz inequality and the definitions of II and h, we obtain

II Z{l+l) - x 112 ::; eII A(x, A, e) - A(z{l), A, e) II . II g(z(l+I») - g(x) II

+ c2 p(AT A) II x - z{l) II . II z(l+I) - x II . (3.6)

On the other hand, it can be shown from the definition of A that

II A(x,A,e) -A(z(l),A,e) II::; ell g(x)_g(z{l») II. (3.7)

To see (3.7), let

!:l.Ai = Ai(X, A, e) - Ai( z(l) , A, e)

= max{>.i + egi(x),O} - max{Ai + egi(z{I»),O},
and consider the following four cases:
(i) if Ai + egi(x) ~ 0, Ai + egi(z{l») ~ 0, then !:l.Ai = e(gi(x) - gi(Z{I»));

(ii) if Ai + egi(x) ~ 0, Ai + egi(z(l») < 0, then °: ; !:l.Ai ::; e(gi(x) - gi(z(l»));

(iii) if Ai + egi(x) < 0, Ai + egi(z{l») ~ 0, then ° ~ !:l.Ai > e(gi(x) - gi(z{l»));

(iv) if Ai + egi(x) < 0, Ai + egi(z{l») < 0, then !:l.Ai = O.
Thus, in each case, we have I!:l.Ail ::; clgi(x) - gi(z(l»)I, which implies (3.7). By (3.6),
(3.7) and the Lipschitz continuity of g, we have

(3.8)

Since e2 (L; + p(AT A)) < 1 by assumption, the inequality (3.8) implies that {z{l)}
converges to x linearly. Therefore, the stopping criterion in Step 3 of procedure
SSP(x, A, 71'; e, E) must be satisfied in a finite number of iterations. 0

Remark. The inequality (3.8) indicates that the sequence {z(l)} generated by proce-
dure SSP(x, A, 71'; e, E) converges to an exact solution x at a linear rate.
The theorem below justifies the termination condition at Step 3 of algorithm PDPPA.
Theorem 1. Let the assumptions in Proposition 1 be satisfied. Suppose that e{k) are
chosen smaller than (L; + p(AT A)t~ for all k. If algorithm PDPPA terminates at
Step 3, then the last iterate x{k+I) is a solution of problem VI.

Proof. By the termination condition of algorithm PDPPA, we have

(3.9)
Suppose that procedure SSP( x{k), A{k), 71'{k); e{k), E{k») halts at iteration 1. That is to say,
we have
(3.10)
Also we have
(3.11)
 149

from Step 1 of procedure SSP(X(k), >.(k) , 1l"(k); C(k), e(k»). It then follows from (3.9)-(3.11)
that
(3.12)
Now let (x, "X, 7f) denote the exact solution of subproblem SP(x(k), >.(k) , 1l"(k)j elk»). By
(3.8), if z(O) =/: x, then

II z(i+I) - x II < II z(i) - x II < ... < II z(O) - x II .

Therefore, in order for (3.12) to hold, we must have

(3.13)

which together with (3.9) and (3.11) implies that

(3.14)

Moreover, by (3.8), (3.13) implies x = z(1). In other words, procedure SSP(x(k), >.(k) , 1l"(k)j
elk), elk») necessarily terminates at the first iteration, i.e., j = O. Since >.(k+I) = fi and
fi = jl(~ = jl(O) , it follows from (3.14) that
(3.15)

This implies that >.(k+ 1) satisfies (1.4) with x = x(k+ 1). In a similar way, we obtain

(3.16)

which implies that 1l"(k+l) satisfies (1.5) with x = x(k+l). Moreover, from Step 2 of
procedure SSP(x(k), >.(k), 1l"(k)j e(k), elk»), we have

(3.17)

Since i = 0 implies z(1) = x(k+I),jl(O) = >.(k+l) and v(O) = 1l"(k+I), it follows from (3.17)
that (X(k+l), >.(k+ 1) , 1l"(k+ 1») satisfies (1.3). Consequently, if (X(k+l), >.(k+ 1) , 1l"(k+I») = (x(k),
>.(k),1l"(k») holds, it follows from (1.3)-(1.5) that (x(k+I), >.(k+I) , 1l"(k+l») is a zero of the
mapping T, i.e., x(k+ 1) is a solution of problem VI. 0

Now we show that algorithm PDPPA converges to a solution of problem VI, when
it generates an infinite sequence.

Theorem 2. Let the assumptions in Proposition 1 be satisfied. Suppose that elk) are
chosen smaller than (L; + p(AT A)t~ Vk. If problem VI has at least one solution,
then the sequence {(x(k),>.(k),1l"(k»)} generated by algorithm PDPPA with an arbitrary
starting point (x(O), >.(0), 1l"(0») converges to a particular solution (x*, >. *, 1l"*) of problem
VI.

Proof. Consider an arbitrary iteration k of algorithm PDPPA. Suppose that proce-

dure SSP(x(k), >.(k) , 1l"(k)., elk) , elk») halts at iteration j.
Namely
, '
(x(k+ ,
1) >.(k+l) 1l"(k+l») I'S

determined as (Z(i+l), jl(i) , vii») satisfying the stopping criterion

(3.18)
150

For simplicity of notation, we shall write (X(k), A(k), 7r(k»), (X(k+l), A(k+l), 7r(k+ 1 »), elk)
and E(k) as (X,A,7r), (X+,A+,7r+), e and E, respectively. Moreover, let z+, z, /-l and v
stand for z(i+l), z(i), /-l(i) and v(i), respectively. Then, by the definition (1.2) of mapping
T, we have

(I + eT) (z+, /-l, v)

( z+ + eF(z+) + eY'g(z+) /-l + eY'h(z+) v + Ns(z+),
/-l - eg(z+) + NRr({I), v - eh(z+)) . (3.19)

From Step 2 and Step 4 of procedure SSP(x, A, 7r; e, f), we obtain

x E z+ + eF(z+) + eY'g(z+) J1 + eY'h(z+) v + Ns(z+), (3.20)

A E {I- eg(z) + NRM({I), (3.21)
+
7r = V - eh(z). (3.22)

It then follows from (3.19)-(3.22) that

(x, A-eg(z+)+eg(z), 7r-eh(z+)+eh(z)) E(I+eT)(x+,A+,7r+),

namely,

(X+,A+,7r+)=J!j.(x,A-eg(z+)+eg(z),7r-eh(z+)+eh(z)), (3.23)

where Jf denotes the resolvent of T, i.e., Jf = (I + eT)-l. On the other hand, we have

II (X+,A+,7r+) - JT(X,A,7r) II
= II J!j. (x, A - eg(z+) + eg(z), 7r - eh(z+) + eh(z)) - J!j. (x, A, 7r) II
~e II (0, g(z) - g(z+), h(z) - h(z+)) II
= e( II g(z) - g(z+) 112 + II h(z) - h(z+) 112)t
~ e( L; + p(AT A))t II z+ - z II, (3.24)

where the first inequality follows from the nonexpansiveness of Jf and the last inequal-
ity follows from the Lipschitz continuity of 9 and the linearity of h. From (3.18) and
(3.24), we have

i.e.,

I (x(k+l), A(k+l), 7r(k+l») - Jf) (x(k), A(k), 7r(k») II < e(k)(L; + p(AT A))t f(k)
< f(k),

where the last inequality follows from the assumption on elk). Then, by Theorem 1 in
Rockafellar [14], the sequence {( x(k), A(k), 7r(k»)} converges to a particular solution of
problem VI. 0
 151

4. AN APPLICATION TO TRAFFIC ASSIGNMENT PROBLEMS

As mentioned in Section 2, algorithm PDPPA will particularly be effective for prob-

lem VI with separable structure such as problem SVI. In this section, we consider an
application of algorithm PDPPA to traffic assignment problems [3]. First we introduce
the notation for the problem.

• 9 = (AI, A): transportation network

• q = 1,'" ,Q: commodities (OD pairs)

• Xq E RIAl: arc flow vector for commodity q

• Xo E RIAl: vector of total flow Xo = 2:~=1 Xq

• E: node-arc incidence matrix of network 9

• d q E RIAl'I: travel demand vector for commodity q

• Sq C RIAl: set of feasible flows for commodity q,

Sq ={ Xq E RIAl'I I EX q = dq , Xq ~ O}
• Xo C RIAl: set of feasible total flows,

Xo = { Xo ERIAl/ Xo = t
q=l
X q , Xq E Sq, q = 1"" ,Q}
• C: travel cost function from RIAl into itself

We assume that the travel cost function C is maximal monotone. Then we can formu-
late the traffic assignment problem as the Variational Inequality problem:

[TA] Find x~ E Xo such that ( C(x~), Xo - x~} ~ 0 Vxo E Xo.

A solution x~ of problem TA is called a user-optimal flow in the sense that no user can
decrease his/her travel cost by only changing his/her route. Clearly, problem TA can
be reduced to problem SVI with the following identifications:

where A is a IAI X (Q + 1)IAI matrix defined by

A = [I -I -I 1
and I denotes the IAI x IAI identity matrix. Note that g( x) and), need not be considered
since there are no nonlinear inequality constraints in problem TA. Then we may apply
algorithm PDPPA to problem TA. Specifically, procedure SSP(x,7l';C,€) specialized to
problem TA, which we call SSPTA( x, 7l'; C, €), is stated as follows.
152

Procedure SSPTA(x, 71"; c, E)

Input: a current iterate (x, 71") and positive scalars c and E.
Output: an approximate solution (z, v) of subproblem SP(x, 71"; c)
Step 1: z~O) := x q , q = 0,1,··· , Q;
1/(0) := 71" + c(xo - E~=l x q );
1:= O.
Step 2: Find z~'+1) such that

C(z~'+1» + 1/(1) + !(Z~l+l) - xo) = o.

c
Step 3: For each commodity q = 1,···, Q, find z~'+1) such that

oE !(Z(l+l) -
c q q . q
+
x ) - 1/(1) Ns (z(I+1» •

Step 4: IT the stopping criterion

Q
L: II z~'+1) - z!') II ~ E
q=O

is satisfied, then go to Step 6.

Step 1/(1+1) := 71" + c (z~I+1) - E~=l z!'+1») ;
5:
1:= 1+1;
go to Step 2.
Step 6: Zq := z!'+l), q = 0,1,···, Q;
v:= 1/(1);
exit.

Step 2 is solving an ordinary system of strongly monotone nonlinear equations with

respect to total flows. This problem may be solved by using any standard method
for nonlinear equations such as Newton or quasi-Newton methods. Step 3 is solving
symmetric linear Variational Inequality problems associated with each commodity,
which may altematively be written as
1
Il!!n 2c II Zq - Xq 112 -(Zq , 1/(1») s.t. Zq ESq.
These are single commodity minimum cost flow problems with separable quadratic
costs. Since these Q problems are independent of each other, they can be dealt with
in parallel. Note that in order to guarantee the convergence of SSPTA(x, 71"; c, f), the
parameter c has to satisfy c < p(AT A)-t (see Proposition 1). Since p(AT A) =
p(AAT) by [6, p.285] and since AAT = (Q + 1)1, we may conclude that procedure
SSPTA( x, 71"; C, f) finitely terminates, provided that c is chosen such that c < 1/ y'Z'J"TI.

5. CONCLUSION

We have proposed a new method for solving monotone Variational Inequality

problems. The proposed method is a variant of the primal-dual proximal point al-
gorithm, which uses the Gauss-Seidel-like procedure SSP to solve subproblem SP at
each major iteration. In particular, when applied to the separable Variational Inequal-
ity problem SVI, procedure SSP efficiently takes advantage of the separable structure
of the problem and hence is suited for parallel implementation. As an example, we
have shown that the proposed method yields a decomposition algorithm for solving
the traffic assignment problem.
 153

REFERENCES
[1) D.P. Bertsekas, "Constrained Optimization and Lagrange Multiplier Methods", Academic Press,
New York, 1982.
[2) D.P. Bertsekas and J.N. Tsitsiklis, "Parallel and Distributed Computation: Numerical Methods",
Prentice Hall, New Jersey, 1989.
[3) S. Dafermos, "Traffic equilibrium and Variational Inequalities," Transportation Science 14, pp.
42-54, 1980.
[4) M. Fukushima, "Application of the alternating direction method of multipliers to separable convex
programming problems," Computational Optimization and Applications 1, pp. 93-111, 1992.
[5) D. Gabay, "Application of the method of multipliers to Variational Inequalities, " in: M. Fortin
and R. Gorwinski eds., "Augmented Lagrangian Methods: Applications to the Numerical Solutions
of Boundary-Value Problems", North-Holland, Amsterdam, pp. 299-331, 1983.
[6) G.H. Golub and C.F. Van Loan, "Matrix Computations", The Johns Hopkins University Press,
Baltimore, 1983.
[7) P.T. Harker and J.S. Pang, "Finite-dimensional Variational Inequality and nonlinear complemen-
tarity problems: A survey of theory, algorithms and applications," Mathematical Programming 48,
pp. 161-220, 1990.
[8) M. Hestenes, "Multiplier and gradient methods," Journal of Optimization Theory and Applications
4, pp. 303-320, 1969.
[9) S. Ibaraki, M. Fukushima and T. Ibaraki, "Primal-dual proximal point algorithm for linearly con-
strained convex programming problems," Computational Optimization and Applications 1, pp.
207-226, 1992.
[10) T. Ito, M. Fukushima and T. Ibaraki, "An iterative method for Variational Inequalities with ap-
plication to traffic equilibrium problems," Journal of the Operations Research Society of Japan 3,
pp. 82-104, 1988.
[11) J.M. Mulvey and A. Ruszczyriski, "A diagonal quadratic approximation method for large scale
linear programs," Operations Research Leiters 12, pp. 205-215, 1992.
[12) M.J.D. Powell, "A method for nonlinear constraints in minimization problems," in: R. Fletcher
eds., "Optimization", Academic Press, London, pp. 283-298, 1969.
(13) R.T. Rockafellar, "Augmented Lagrange multiplier functions and duality in non-convex program-
ming," SIAM Journal on Control 12, pp. 268-285, 1974.
[14) R.T. Rockafellar, "Monotone operators and the proximal point algorithm," SIAM Journal on Con-
trol and Optimization 14, pp. 877-898, 1976.
[15) R.T. Rockafellar, "Mouotone operators and augmented Lagrangian methods in nonlinear program-
ming," in: O.L. Mangasarian, R.R. Meyer and S.M. Robinson, eds., "Nonlinear Programming 3",
Academic Press, New York, pp. 1-25, 1978.
[16) R.T. Rockafellar, "Lagrange multipliers and Variational Inequalities," in: R.W. Cottle, F. Gian-
nessi, and J .L. Lions eds., "Variational Inequality and Complementarity Problems: Theory and
Applications", Wiley, New York, pp. 330-322, 1980.
[17) P. Tseng, "Application of a splitting algorithm to decomposition in convex programming and
Variational Inequalities," SIAM Journal on Control and Optimization 29, pp. 119-138, 1991.
[18) C. Zhu, "Modified proximal point algorithm for extended linear-quadratic programming," Compu-
tational Optimization and Applications 1, pp. 185-205, 1992.
RELATIONS BETWEEN t-, S-, z-DOMAIN DESCRIPTIONS OF
PERIODICALLY-SWITCHED NETWORKS

Koksal Muhammet

Inonu University
Engineering Faculty
44100 Malatya, Turkiye
e-mail: inonu01@trearn.bitnet

Abstract. When a network, either in engineering or in mathematical sense, is abruptly

changing from one state to another, more clearly, it is switched periodically from one
linear time-invariant state to another, its behavior in time domain is described by both
the state equations defining linear time-invariant system dynamics between switching
instants and the switching equations defining the abrupt changes or discontinuities in
system variables. The use of such systems appears extremely common in the last decade
in the form of switched-capacitor networks in electrical and electronics engineering.
Although enough literature can be found for the time and frequency domain anal-
ysis, as well as for the z-domain analysis of the networks of the concerned type, a clear
and rigorous dealing of the relations between z-domain treatments and the other two
domain-treatments seems to be lacking. Especially different approaches have been used
to define the z-domain transfer functions; and their interrelations as well as connections
to the (s-domain) time-varying system function can hardly be found. The main purpose
of this contribution is to set up these relations and to clarify the position of different
definitions and to identify their differences by careful and exact mathematical formulas.

Key Words. Periodically-switched networks, dynamical systems.

1. INTRODUCTION

It is known [lJ that a linear system, which is time-invariant except at a finite

number (K ~ 1) of instants tn,l < tn,2 < ... < tn,k (tn,k - tn,k-l = Tk) in the n-
th period (nTo,(n + l)ToJ, with tn,o = nTo, tn,K = tn+l,O = (n + l)To, having abrupt
changes in its parameters and/or topology at these instants, can be completely described

155
156

by the following equations:

d
= Akxn,k(t) + L
J..
dt Xn,k(t) Bk,jU(')(t); tn,k-l < t < tn,k (Ua)
i=O

J.+J.+,-6(J.+tl
Xn,1+1(t~,k) = Fkxn,k(t;;,k) + L Gk,iU(i)(t;;,k)+
i=O
(Ub)

i=O

J.
Yn,k(t) = CkXn,k(t) + L Dk,iU(i)(t); t n,k-l < t < tn,k. (Uc)
i=O

These equations are known to be the state, switching, and output equations, respectively;
where n = 0,1,2, ... , k = 1,2, ... , K, and due to periodicity t n+1,k = tn,k + To. Note
that, although at any time the system reduces to a linear time-invariant system, the
whole system is periodically time-varying but still linear. Xn,k is the state vector at
the k-th interval; its dimension and content may vary from one interval to another, but
must be the same after each period To. Yn,k(t) represents the output function yet) in
the k-th interval; its dimension and entries are the same for all intervals. u(t) represents
the forcing (source) function or excitation of the system; it is assumed to be pice-wise
continuous in its Jk-th derivative during the intervals t n ,k-l < t < tn,k; and at the
switching instants tn,k to have left and right derivatives up to Jk + J1+1 - 6(h+d and
hH - 1 orders, respectively. h is the maximum order of the source derivative in the
k-th interval, and 6(J1+1) is zero if J1+1 = 0, if J1+1 ~ 1 it is l. Superscript (i) denotes
the i-th derivative. Finally, Ak, Bk,i, Fk, Gk,i, (h,iCk, Dk,i are constant matrices of
appropriate orders.

Although Eqs. (l.la), (l.lb), (l.lc) are originally derived for electrical networks
containing linear, lumped, bilateral components and periodically operated switches [2],
they may describe any physical system analogous to such networks; and the main concern
in this article is focused on the solution properties of these equations and the transfer
characteristics between the input u(t) and the output yet). Note that no nonlinearity
and time-delay as well as distributed parameters are allowed in the system.

When the input u(t) is an exponential, i.e., u(t) = U ept where U is a complex vector
and p is a complex scalar (frequency), the above equations easily simplify to the new
ones containing no source derivatives; for this case the explicit steady-state and transient
solutions for Xn,k(t) and yet) are already found [3]. With the assumption of piecewise
constant excitation u(t) during each interval t n,k-l < t < t",k, such as the case the
input is fed through a sample (at time instants tn,k-d and hold (for tn,k-l < t < tn,k)
circuit, similar simplifications result; further assumptions which are valid in most of
the switched-capacitor networks are the case of 00 time constants (zero natural system
dynamics Ak = 0), no direct coupling between state vectors and inputs (Bk~ = 0), and
between outputs and inputs (Dk,i = 0) during each interval; on the base of the previous
 157

work [1,2,3], this case is considered by Nacaroglu [4] and the first general solution results
in z-domain (discrete frequency domain) are obtained starting from the time domain
equations in (1.1). In a later work, the assumptions Bk,i = 0, Dk,i = 0, and the picewise
constancy of excitation are removed and more general results are obtained in [5].

In the following, the solution of the equations in (1.1) with

a) Zero natural system dynamics
b) Anyone of the following assumptions
i) Exponential input
ii) Piecewise constant input (constant during each interval)
are handled in time (t-), complex frequency (s-), and discrete complex frequency (z-)
domains, and the interrelation between transfer funtions in different domains are treated.

For the judgment of these assumptions and the material in the sequel, the following
basic knowledge about the periodically time-varying systems must be introduced.

The (unit) impulse response function h( t, r) of a linear system is defined to be the

response to a unit impulse bet - r) occurring at time t = r, and it is zero for t < r for

1:
casual systems [6]. Time-varying system function H(s, t) is defined by [7,8]

H(s, t) = h(t, r)e- 8 (t-r)dr, (1.2)

where s denotes the complex frequency 0'+ jw. The forced response of the relaxed system
to any input u(t) with the Laplace transform U(s) is then given by

yet) 1 /00
= -2' H(s,t)U(s)e·tds = ~Residues of H(s,t)U(s)e 8t at its poles. (1.3)
7rJ -00
For a unit exponential input u(t) = ePt , i.e., U(s) = l/(s - p) where p denotes the real
frequency jw, the last expression yields the following steady-state response due to the
input, or the complete steady-state response if the system is asymptotically stable (all
natural modes die away with time, i.e., no steady-state natural modes exist); in other
words H(p, t) can be defined to be

yet) = H(p, t)e pt , (1.4)

H( t) = steady - state response
(1.5)
p, unit exponential input'
Finally, for periodically time-varying systems (with period To), H(p, t) is periodic in
time and can be expanded into its usual Fourier series as
00
H(p, t) = L H/(p)ei/w0t, (1.6)
/=-00

where Wo = 27r/To, and H/(p)'s are the complex Fourier coefficients [9J. Substitution of
(1.6) into (1.4) yields that
00
yet) = L H/(p)ei(w+/wot). (1.7)
/=-00
158

This equation reveals the important property of the periodically time-varying linear
systems, which states that for a single input frequency w of the input, the output contains
infinitely many frequency components w + lwo of which only one is at the input frequency
appears in the output, i.e., H, = 0 if 1 =1= O. Therefore, Ho(p) is known to be the time-
invariant transfer function of a periodically time-varying linear system. Since the input
frequency component of the output is of concern in most of the applications, Ho(p),
being the average value of H(p, t), and hence through (1.5) the exponential input is very
important to study the behavior of periodically time-varying linear systems.

In many cases a switched-capacitor network is fed through a sample and hold circuit
[10]. This is equivalent to say the input u(t) is piecewise constant. Hence both of the
assumptions made in bi) and bii) are valid practically.

2. SOLUTION

With the assumptions considered in the previous section, to solve Equations in

(1.1), for simplicity but without loss of generality, consider a single-input, single-output
system. Define the following K input and K output components; k = 1,2, ... , K:

Uk(t) = {~(t) tn,k-l :5 t < tn,k (2.1a)

elsewhere

Yk(t) = {~(t) tn,k-l :5 t < tn,k (2.1b)

elsewhere

for n = 0,1, .... Associated with the k-th interval (most frequently referred as phase),
these variables are referred to be the k-th input and output phases, respectively. Note
that the value of each variable at the switching instant tn,k-l is defined to be its value
after the switching at this instant; i.e., for e.g., y(t!,k_l) = y(tn,k-l) = Yk(tn,k-t};
however, y(t~,k_l) may not be equal to y(tn,k-t} and it is denoted as Yk-l(t~,k_l)'
With this assumption, the system in (1.1), without any other assumptions, has a unique
solution for the state and output vectors of which the only discontinuities may occur at
the switching instants, except the output may have additional discontinuities wherever
Jk-th derivative of the input is also discontinuous [6].

To prevent the drawback of using fractional powers of z in the transfer functions

between input and output phases [11], each of the input phases Uk(t) is shifted by an
amount 17k = tn,k-l - tn,o = Tl + T2 + ... + Tk-l to the left on the time-axis so that the
new signal !lk(t) = Uk(t + 17k) appears in the first interval, which is now of length Tk,
instead of TI' Similarly iJk(t) = Yk(t + 11k)' A sample of variations of Uk(t), Yk(t) as well
as those of !lk(t) and iik(t) in the n-th switching period is shown in Fig. 1.
 159

-+-:-::-:-:
t~,o
+--::-«:-:--+----~-+--~-~-:-))--II------I--
t~,l In,2
I
tn,k-l
I
tn,k
I
tn,k-l tn,k tn+1,l

nTO :
I I : tn~1.0
.: ..
I I

:.. ......_=--(_n~,..~l ) TO
!- 1
I I
~I ~I
I '/;'1 : 't'2 I I "k 't'k :
I I I
~1I2 I
I
I
I
I
I
~1'II3 I I
~I 'Ilk

~
~ ) t

Figure 1. The n-th switching period [tn,o, tn,k) and a) samples of variations of Uk(t) and
Yk(t), b) Uk(t) and iMt).

Exponential and Piecewise Constant Inputs

With the exponential input u(t) = ept where p = (J' + jw is the complex frequency
u(C k) = epr'u(t+ k-l)' and with the piecewise constant input u(t) = u(t; k-l) for
tn,kn~l < t < tn,k ;here u(t;;-,k) = U(t;,k_l)' Equations in (1.1) simplify to '

tn,k-l < t < tn,k (2.2a)

(2.2b)

tn,k-l < t < tn,k (2.2c)

where for the exponential input,

J. J.+J.+ 1 -6(Jo+d J.
Bk = LBk,ipi, Gk = L Gk,ipi, Gk = 0, Dk = LDk,ipi, (2.3)
i=O i=O i=O

and for the piecewise constant input

(2.4)

To derive the transfer characteristics between u and x, consider the solution for the i-th
phase due to j-th phase of the input.
For j = i, starting from x(tn,i-J), using Eq, (2.2a) for k = j, we obtain
160

This, together with the switching equation at tn,j (Eq. 2.2b with k = j) yields

Xn,j+I(t;) = Fj [xn,j(t;,j_I) + B j ePTjp-1 u(t;,j_I)] + Gju(t;;)

(2.6)
= FjXn,j(t;,j_I) + [FjBj ePTjp- 1 + GjePTj] u(t;,j_I)

since u(t;) = O. Note that although this equation is derived for the exponential input,
it is valid for the picewise constant input as well; infact if p __ 0, (e PTj - 1)/p __ Tj,
ePTj __ 1 by which the correct results are obtained for the picewise constant input case.
Up to time t;;+I,j_I' u(t) = uAt) is zero in (2.2a), hence x remains constant in each
interval of durations Tj+bTj+2, ... ,Tk,TI,T2, ... Tj-I. Therefore, the only variation in
the state vector x occurs at the switching instants; further, in the switching equations
(2.2b), due to zero input, only the first term becomes effective. Hence,

(2.7)

Finally, using the switching equation at tn+I,j-I and letting u(t;;-+I,j_I) = 0, we obtain

Substituting (2.6) in (2.7), and then (2.7) in (2.8) we obtain

(2.9)

Using the shifted variables for the state vectors (Xn,k(t) = Xn,k(t + 1}k» as well as for
the input as defined before, and remembering that j = i, we can write (2.9) as

Xi[(n + 1)Tol = FaijXi(nTO) + FbijUj(nTo) + FCijUj[(n + 1)Tol (2.10)

where
Faij = Fi-I" .FIFK .. .Fi, (2.11a)

Fbij = Fi-I .. .FIFK .. .FHI [FjBj ePTjp-1 + Gje PTj ] , (2.11b)

Fcij = Gj - I . (2.11c)

For j > i, by following exactly similar steps, we can obtain (2.10) with

(2.12a)

Fbij = Fi-I .. .FIFK .. .Fj+I [FjBj ePTjp- 1 + Gje PTj + FjGj-I] , (2.12b)

Fcij = O. (2.12c)

For j < 1, (2.10) remains the same, but now

(2.13a)
 161

(2.13b)

(2.I3c)

In fact, Eq. (2.10) is a difference equation relating the discrete values of Xi and Uj.

Time-Domain (Transient) Solution

Although Equation in (1.1) can be solved both for Xn,k and then Yn,k(t) starting
from the initial vector XO,l(O+) and using the state equation (lola) and the switching
equation (LIb) sequentially, and continuing this process along any desired number of
periods, this is time consuming if only the solution in the n-th period or the steady-state
solution is required.

To find the transient solution (complete solution is meant) in the n-th switching
period, the initial value of each phase is computed as follows: solving (2.2a) with n = 0
and for to,o < t < to,l we obtain

(2.I4a)

Using (2.2b)

(2.14b)

is obtained. Repeating this procedure further for i = 3,4, .. .K, we obtain the following
general formula:

(2.14c)

This formula can be used for i = 1 and 2 as well; care mut be paid that summations can
be taken to be zero and products be taken to be 1 when the upper limits are smaller
than the lower limits.
Using the headed phase variables, (2.14c) can be written as

(2.15)
162
After computing Xi(O), we now return to the difference equation (2.10). The solution of
this constant-coefficient first order difference equation can easily be obtained as

Xi(nTO) = F,:ijXi(O) + F,:ijl FbijUiO)+

n-l
+ L F;:;jl-I(Fbii + FaijFcij)uj(lTo) + FcijUj{nTo).
(2.16)
1=1

Using this equation, the state at t n,i-l can directly be reached without sequential
treatments of Eqs. (2.2a,b). The contribution of all input phases can be considered by
summing the result in (2.16) over the index j = 1,2, ... , K. This way Xn,k(tn,k-I) =
XA,(nTo) is computed; after each switching instant in the n-th switching period, hence
the solution in the intervals t n ,A,-1 < t < tn,A, can be found by using the integral solution
of (2.2a). Knowing solution for each state vector, the output vector yet) can be computed
from (2.2c) readily.

z-Domain Solution

To deal with the discrete frequency (z-) domain solution, we again return to the
difference equation in (2.10). Using the regular rules for the z-transform, we obtain

Xi(Z) = [I - z-1 Faij] -1 [z-l Fbij + Fcij] UAz) + [I - Z-l Faij] -1 [Xi(O) - FcijUj(O)] ,
(2.17)
where Xi(O) is computed from (2.15), and Uj(z) and Uj(O) are assumed to be known.
When all the phases are present, which is usually the case, Xj(z) is computed by sum-
ming (2.17) over j = 1,2, ... , K since the system is linear; i.e.,
K
Xi(Z) = L [I - Z-l Faij] -1 { [z-1 Fbii + Fcij] Uj(Z) - FcijUj(O) + Xi(O)} (2.18)
j=l

The discrete values of Xi(nTO) and hence Xi(nTO + Tfi) = Xi(t n,i-l) can be computed in
any interval specified by n, for e.g. by the formula

Xi(nTO) = ~fzn-lXi(Z)dz = Residues[zn- l Xi(Z)]. (2.19)

27rJ
As mentioned previously, knowing the discrete values of x at tn,A, 's, Eq. 2.2a can be
used to find any value between these instants. Further, y can be computed readily from
(2.2c).
When the steady-state response characteristics are of concern Xi(O) and Uj(O) in
Eq. (2.17) are deleted. Hence the z-domain transfer function Tij(Z) and the relation
between the j-th input phase and the i-th state phase appears to be

Tij(Z)
-
= [I - Z-1 Faij ] -1 [ Z-1 Fbij + Fcij ] , (2.20a)

(2.20b)

respectively. Transforming Eq. (2.2c) into discrete time domain (t = t n ,k-l), using the
headed (shifted) variables, noting that the j-th input phase is not effective on the i-th
 163

output phase through the matrix Dk, and using the z-domain transformed variables, we
obtain
(2.21)

where 6ij is known to be the Kronecker delta and it is equal to 1 if i = j, otherwise it is

zero. Substituting (2.20) in (2.21) we have

Y;(Z) = HijUj(Z) where, (2.22a)

Hij(Z) = CiT;j(z) + 6ijDj (2.22b)
= Ci [I - z-1 Fa ijr 1 [Z-1 F.ij + Fcij] + 6ijDj; (2.22c)

the last is the z-domain transfer function between the j-th input phase and the i-th
output phase.
Since superposition applies due to the linearity of the system, the state and output
responses due to all input phases can be computed from

K
Xi(z) = LTij(Z)Uj(z) (2.23a)
j=1

K
Y;(z) = LHij(Z)Uj(z) (2.23b)
j=1

for i = 1,2, ... , K. These equations can be written in matrix from as well. When this is
done for the output y we obtain

i!IK(Z)
H 2 K(Z)
1[ i!1(Z)(z) 1
U2
(2.24)
HKK(Z) UK(Z)

The above K x K square matrix is known to be Generalized Phase Transfer Matrix

(GPTM). Each transfer function in this matrix is defined by Eq. (2.22b) together with
Eqs. (2.11), (2.12), (2.13) and (2.3). This completes the basic discussion in z-domain.

s-Domain Considerations

In this section, the time-varying system function defined by Eq. (1.2) (with s is
replaced by p) derived by using the z-domain analysis results and using the definition
in Eqs. (1.4,5).
Assuming an exponential input u(t) = ept , Uj(z) can be computed in the following
steps:
164

Hence, from (2.20b) we obtain

- - Z
Xi(Z) = Ti·(z)eP'Ij
J z _ epTo . (2.26)

Using (2.19) with n = 0, Ii(O) is expressed as

1'... (z )eP'Ij
Ii(O) = E Residues IJ
Z - eP
1',
0
(2.27)

We assume an asymptotically stable system so that all poles of T;j(z) are within the
unit circle and the residues at these poles will have no contribution on the steady-state
value of Ii(O). Hence, to find the time varying system function, the residues at the pole
z = epTo (p = jw and hence I ejwTo I = 1) is considered only. The result

(2.28)

is equal to X(to,i-1) = XO,i(1]i). Using the state equation (2.2a) for the i-th interval, the
solution in this interval can be expressed as

1]i ::; t < 1]i+1· (2.29)

Using the output equation (2.2c), and the relation (2.22b) between TiiCz) and Hij(z),
Yo,i(t) can be obtained to be

(2.30)

which is valid for 1]i ::; t < 1]i+1. Comparison of this equation with (1.4) yields that for
this interval

- (pI',)
Hi,j (p, t ) = [Hij e 0 - Oij (CiBj
-p- + Dj )] ep('I·-
J
t)
+ Oij (Ci
-p-Bj + Dj ) . (2.31 )

When the effect of all input phases are considered, the time domain steady-state so-
lutions can be obtained by summing Eqs. (2.29,2.30) over the index j (superposition).
The results are

(2.32)

H(p,t) = tHij(ePTO)eP(f/j-t) + (CiBi +Di) (l-e P('I;-t» , (2.33)

j=l p

which are valid for 1]i ::; t < 1]i+1, i = 1,2, ... , K. With yet) = H(p, t)e pt and H(p, t +
nTo) = H(p,t), Eq. (2.33) defines yet) for all t 2:: O. Similarly, with x(t) = L(p,t)e pt
and L(p, t + nTo) = L(p, t), where

(2.34)
 165

for '1/i ::; t < '1/i+l, i = 1,2, ... , K, x(t) can be computed for all t ~ o.
Since the transfer function is the average value of the time-varying system function,
averaging H(p, t) over (0, To), it is obtained from Eq. (2.33) to be

Ho(p) = _1_ LK [ K
e- P'l;(l- e- pr;) LHij(ePTO)eP'li+
pTo ;=1 j=1
(2.35)
+(e- pr; + pT; _ 1) (C:i + D) ].
We close this section by noting that the obtained results are valid for multi variable
systems having multiple input and/or output variables. In this case H(p, t) and hence
Ho(p) will be matrices which can be called, for e.g., time-varying system matrix and
transfer matrix, respectively. Similarly, each element of the generalized phase transfer
matrix in Eq. (2.24) will be a transfer matrix as well.

Another point following from the linearity is that when the input e pt is replaced by
U e Pt , where U is a complex scalar or vector, the responses for the state and output vectors
are still valid with ept is replaced by Ue pt , i.e., x(t) = L(p,t)Ue pt , yet) = H(p,t)Ue pt •

3. z-DOMAIN TRANSFER FUNCTION

As it is seen from Eq. (2.33), for B; = 0, D; = 0 (no direct coupling between

the state-derivative and the input, as well as between the output and the input), the
generalized phase transfer matrix in Eq. 2.24 is a good tool not containing the fractional
powers of Z to represent the relations between all input and output phases and hence
between the input and output. In many applications, however, this transfer matrix is
too involved for practical applications and a single transfer function (for single-input,
single-output systems) H(z) is defined and used under the following practically valid
simplifications.

i) If the input u(t) is of a slow-varying type with respect to the periodicity of the
system, which is possible if the highest frequency component of the input is much smaller
than the switching frequency l/To , then it can be assumed that

U(Z)=Uj(Z) for j=1,2, ... ,K since (3.la)

u(nTo) = u(nTo + '1/d ~ u(nTo + '1/2) ~ ... ~ u(nTo + 7/1(), (3.1b)

hence Eq. 2.24 can be written as

Y;(Z) = Hi(Z)U(Z) where (3.2a)

I(

H;(z) = LH;j(z); i = 1,2, ... ,K . (3.2b)

j=1
166

The last sum is the row sum of the generalized phase transfer matrix and it is defined
as the transfer function between the input and the i-th phase of the output. In many
of the first order switched-capacitor filters [12,13] and second order SC biquads [14,15]
H;(z) is the same for all output phases, i.e., Hl(Z) = H2 (z) = ... = HJ(z), therefore it
is usual to define the z-domain transfer function of the filter as
J(
H(z) = H;(z) = LH;j(z) . (3.3)
j=l

ii) In many applications some output phases are effected by the corresponding input
phases only [16,17]' and/or the relation between the same input and output phases is of
concern; in this case H (z) is given by

(3.4)

where i E [1, K] is the appropriate integer.

iii) Sometime by using selector switches only, one input phase is made effective at
the output and only one of the output phases are selected; in this case

(3.5)

i.e., any entry of the GPTM is a transfer function of the whole circuit.

4. CONCLUSIONS

In this presentation a tutorial theory about the solution of piecewise constant but
periodically time-varying linear systems with zero natural dynamics is presented. In
addition to the state equation, switching equations are used to determine the discon-
tinuities between adjacent constant states. The theory finds its basic applications in
switched-capacitor networks which are widely used in integrated electronics in the last
decade due to their small chip area, environmental and constructional insensitivities,
and externally adjustable properties.

The solutions results are discussed in the time, complex frequency and discrete
frequency domains; the results especially relating the z-domain transfer functions to the
s-domain transfer functions, as far as the author's knowledge, are believed to be original
and carries much importance. When the system dynamics exist, which is inevitable when
the nonidealities are to be considered in physical systems, the presented analysis gets
much more difficult; although some exact solution results appear in the literature in
t- and s-domains [3], z-domain treatment similar to the one presented in this article
deserves to be theoretical importance.

Another particular interest would be the study of solution properties with zero
dynamics when the input is or can be assumed to be piecewise linear between switching
 167

instants; which is a case that will naturally yield more accurate results than piecewise
constant approximation and better research pleasure.

It is believed that applications in the industry and real life would not be of much
interest for the mathematicians; therefore electronic circuits applying the dealt formulae
some of which can be found in the reference list not included in this material.

REFERENCES

[1] Koksal M. and Y. Tokad, "State space formulation of linear circuits containing periodically oper-
ated switches," Circuit Theory and Appl., vol. 5, no. 1, pp. 155-170, Jan. 1977.
[2] Koksal M., "Analysis and applications of linear circuits containing periodically operated switches" ,
Ph. D. dissertation, M.E.T.U., Elect. Eng. Dept., Ankara-Thrkey, May 1975.
[3] Koksal M., "On the solution of linear circuits containing periodically operated switches," Proc.
1976 European Conf. Circuit Theory and Design, Genoa, Italy, vol. 1, pp. 77-82, Sept. 7-10,
1970.
[4] Nacaroglu A., "Computer oriented analysis of linear circuits containing periodically operated
switches; Applications to switched capacitor networks", Ph. D. dissertation, M.E.T.U., Elect.-
Electron. Eng. Dept., Gaziantep-Thrkey, Dec. 1989.
[5] Kilci T., "Computer aided investigation of nonideal situations of linear circuits containing peri-
odically operated switches; Applications to switched-capacitor networks," Ph. D. dissertation,
Gaziantep University, Elect.-Electron. Eng. Dept., Gaziantep-Thrkey, Sept. 1993.
[6] Desoer C.A., "Notes for a Second Course on Linear Systems" , Van Nostrand Reinhold, New York,
1970.
[7] Zadeh L.A., "Frequency analysis of variable networks," Pmc. IRE, vol. 38, no. 3, pp. 291-299,
March 1950.
[8] Zadeh L.A., "Time-varying networks I," Pmc. IRE, vol. 49, no. 10, pp. 1488-1503, Oct. 1961.
[9] Pipes L.A. and L.R. Harvill, "Applied Mathematics for Engineers and Physicists", McGraw-Hill
Int. ed., Singapore, 1971.
[10] Farg S.C., Tsividis Y.P. and O. Wing, "SWITCAP: A switched-capacitor network analysis pro-
gram, part I: Basic features," IEEE Circuits and Systems Magazine, pp. 4-9, Sept. 1983.

[11] Lee M.S. "Switched-capacitor filters using floating-inductance simulation circuits," Electron.
Lett., vol. 15, no. 20, pp. 644-645, 27th Sept. 1979.
[12] Clement F.L. and W.K. Jenkins, "Computer-aided analysis of switched-capacitor filters," IEEE
Trans. Circuits and systems, vol. CAS-28, no. 7, pp. 681-691, July 1981.
[13] Martin K. and A.S. Sedra, "Effects of the op-amp finite gain and bandwidth on the performance of
switched-capacitor filters," lEE Trans. Circuits and Systems, vol. CAS-28, no. 9, pp. 822-829,
Aug. 1981.

[14] Bermudez J.C.M. and B.B. Bhattachanya, "A systematic procedure for generation and design of
parasitic insensitive SC biquads," lEE Trans. Circuits and Systems, vol. CAS-32, no. 8, pp.
767-783, Aug. 1985.
168
[15] Hegt J.A., "Contributions to switched-capacitor filter synthesis," Ph. D. Thesis, Eindhoven Univ.
of Technology, Jan. 1988.
[16] Rahim C.F., Copeland M.A. and C.H. Chan, "A functional MOS circuit for achieving the bilinear
transformation in switched-capacitor filters," IEEE J. Solid-State Circuits, vol. SC-13, no. 6,
pp. 906-911, Dec. 1978.
[17] Martin K., "Improved circuits for the realization of SC filters, lEE Trans. Circuits and Systems,
vol. CAS-27, no. 4, pp. 237-244, Apr. 1980.
ON SIDE CONSTRAINED MODELS OF TRAFFIC EQUILIBRIA

Larsson Torbjorn1 and Patriksson Michael 2

Division of Optimization
Department of Mathematics
Linkoping Institute of Technology
S-581 83 Linkoping, Sweden
1 e-mail: tl@math.liu.se
2 e-mail: mipat@math.liu.se

Abstract: The basic model of traffic assignment does not capture complex traf-
fic flow relationships such as interactions among vehicles on different road links, joint
capacities on two-way streets, turning priorities in junctions, etc. Traditionally, such
relationships are introduced implicitly through generalizations of the separable travel
cost function of the basic model, and the refined models are usually stated as varia-
tional inequalities. However, this approach has found very little practical use, seem-
ingly mostly because the generalized travel cost functions are not easily chosen and
calibrated.
An alternative strategy for refining the basic model is to capture supplementary
traffic flow relationships explicitly through the introduction of side constraints. This
explicit approach has the advantage of always leading to optimization formulations,
and the immediate physical interpretations of the side constraints introduced may make
it easy to construct, calibrate and apply the resulting models compared to models with
non-separable cost functions.
We consider a traffic equilibrium model with general side constraints and show that
its optimality conditions correspond to a generalization of Wardrop's user equilibrium
principle. Further, we derive a close relationship between models of traffic equilibria
involving side constraints and generalized travel cost functions, respectively. These
results motivate a further study into the art of modelling traffic assignment problems
through the use of explicit side constraints.

Key Words: Traffic assignment, user equilibrium, side constraints, generalized

Wardrop conditions, variational inequalities.

1. INTRODUCTION AND MOTIVATION

Consider a transportation network g = (N, A) of nodes and directed links. Between

certain pairs of origins and destinations, (p, q) E C, fixed positive travel demands dpq

169
170

are given, and each link a E A is associated with a positive and strictly increasing
travel cost function ta : ~~I >--+ ~++. The user equilibrium principle of Wardrop (1952)
states that for each origin-destination (O-D) pair (p, q), the routes utilized have equal
and minimal travel costs, that is, at an equilibrium flow, the conditions

hpqr >0 ===;.. cpqr = 11'pq, (Ll)

(1.2)
(1.3)

where hpqr is an equilibrium flow on route r E R pq , C pqr = cpqr(h), with h = (hpqr)'

is the travel cost on the route, and 11'pq is the equilibrium travel cost of the least-cost
routes in O-D pair (p, q), are satisfied.
For separable travel cost functions, i.e., where the travel cost on each link is inde-
pendent of the flow on other links, a solution to the Wardrop Conditions (1.2)-(1.3)
can be found by solving the convex network optimization problem (e.g., Beckmann et
aI., 1956, and Dafermos, 1972)

[TAP)

[fa
mm T(f) := L Ju ta(s)ds, (1.4)
aEA 0

subject to

L hpqr dpq , Yep, q) E C, (1.5)

rE'Rpq

hpqr 2: 0, Vr E R pq , Yep, q) E C, (1.6)

L L tipqrahpqr fa, Va E A, (1.7)
(P,q)EC rE'Rpq

where
if route r E Rpq uses link a,
Va E A, Vr E Rpq, Yep, q) E C,
otherwise,

is the link-route incidence matrix, and fa denotes the total flow on link a.
The equilibrium model [TAP) is frequently applied in transportation analysis, and
many algorithms have been developed for its efficient solution. (See, e.g., Patriksson,
1995, for a thorough review of algorithms for [TAP).) The inherent simplicity of the
model, however, makes it inapplicable to more complex traffic problems (e.g., Sender
and Netter, 1970). For instance, it does not capture the interactions between the flows
on intersecting links, or between vehicles of different types. An illustrative example of
a deficiency of the model and its possible consequences is provided by Hearn (1980),
who comments on its property of allowing every road to carry arbitrarily large volumes
of traffic. This deficiency of the model causes that

"the predicted flow on some links will be far lower or far greater than the
traffic engineer knows they should be if all as.'lUmptions of the model are
correct. In practice, the result is that the model predictions are ignored, or,
more often, the user will perturb the components of the model (trip table,
volume delay formulas, etc.) in an attempt to bring the model output more
in line with the anticipated results."
 171

In order to avoid such heuristic tampering with components of the model available,
traffic planners must be supplied with analysis tools whose underlying traffic models
are sufficiently general, reliable and accurate.
We distinguish two approaches for improving the model's ability to accurately de-
scribe, reproduce, or predict a real-world traffic situation.
The traditional approach is to capture additional flow relationships through the
introduction of non-separable, and typically also asymmetric, travel cost functions.
The solution of (1.2)-(1.3) can then, however, not be reformulated into an optimization
problem of the form [TAP], due to the non-integrability of the resulting travel cost
function c : lR~1 I-T lR~!. Instead, the Wardrop conditions are formulated as the
Variational Inequality problem of finding an f* E F such that

[VIP]
cu*fu - f*) ~ 0, "lIEF,
where F = {I E lR 1A1 I I satisfies (1.5)-(1. 7)}. The resulting class of models has
been extensively studied from a theoretical and algorithmical point of view (see, e.g.,
Nagurney, 1993, and Patriksson, 1995). Seemingly, the asymmetric models' popularity
is a consequence of their mathematical elegance and nice interpretations rather than
their applicability, since real-world applications seem to be lacking. A major reason
for this is probably the practical difficulty of choosing and calibrating the asymmetric
travel cost functions.
An alternative-but so far little studied-approach to improve the quality of the
basic traffic equilibrium model is to introduce a set of side constraints to model ad-
ditional restrictions on possible flow patterns. Such side constraints could be used to
describe, for instance, the interactions among vehicles in a junction, joint capacities
on two-way streets and links in intersections and roundabouts, requirements that ob-
served flows on some links should be reproduced in the calculated solution, a traffic
control policy, or dynamic aspects. We believe this approach to be appealing from a
practical point of view, since it is certainly easier for the traffic engineer to identify a
suitable set of side constraints-which may have immediate physical interpretations-
than to estimate proper values of parameters in complex travel cost functions. (In the
example provided by Hearn (1980), the proper improvement of the basic model is the
introduction of link capacity constraints corresponding to the engineer's anticipation of
reasonable levels of traffic flow.) The approach to improve a traffic equilibrium model
by introducing general side constraints was first discussed by Larsson and Patriksson
(1994).
Although this alternative approach is more straightforward than that based on
asymmetric cost functions, it has been given comparatively very limited attention.
We present a general side constrained assignment model and investigate its optimality
conditions; these may be interpreted as a generalization of Wardrop's equilibrium Prin-
ciple (1.2)-(1.3) in the sense that an equilibrium holds in terms of generalized travel
costs. Moreover, we show that the side constrained assignment problem may be equiv-
alently solved as a standard equilibrium model with a well-defined adjusted travel cost
function. This result leads to an interesting relationship between side constrained and
asymmetric models of traffic equilibria, which is one motivation for further studying
side constrained models.
172

2. A SIDE CONSTRAINED ASSIGNMENT MODEL

Let gk : ~~I 1-+ ~, k E K, be convex and continuously differentiable, and define the
side constraints
VkEK.
Here, the index set K may, for instance, consist of the index set of the links, nodes,
routes, or O-D pairs, or any combination of subsets of them. (The constraints are,
without any loss of generality, given as inequalities.)
Consider the general side constrained traffic equilibrium problem

[TAP-SCl

min T(f) = 1: infl. ta( s )ds, (2.1)

"eA 0

subject to

1: hpqr d pq , V(p,q) E C, (2.2)

reRpq
h pqr ~ 0, Vr E 14q, V(p, q) E C, (2.3)
L L
(p,q)eC reRpq
lipqr"hpqr f", VaEA, (2.4)

gk(f) ~ 0, VkEK. (2.5)

We assume that the feasible set of [TAP-SC] is non-empty; in cases where some
functions gk are nonlinear, we also assume that a constraint qualification (e.g., Bazaraa
et al., 1993, Ch. 5) holds. The convexity of [TAP-SCl then ensures the existence of an
optimal solution, which is unique in the link flows and characterized by the first-order
optimality conditions. We next show that the optimality conditions of [TAP-SCl give
rise to a Wardrop equilibrium principle in terms of generalized route travel costs.

Theorem 1. (A generalization of the Wardrop principle) If (h, f) solves the problem

[TAP-SC] and 11" E ~Ici and (3 E lRlK:1 are vectors of optimal Lagrange multipliers for
the Constraints (2.2) and (2.5), respectively, then

hpqr >0 ~ cpqr = 1I"pq, Vr E 14q, (2.6)

hpqr =0 ~ cpqr ~ 1I"pq, Vr E 'R. pq (2.7)

holds for all O-D pairs (p,q) E C, where

Vr E 14q, V(p, q) E C. (2.8)

Proof. Stating the stationary point conditions for the Lagrangean function

L(f, (3) ~ T(f) + 1: (3kgk(f) (2.9)

keK:
 173

subject to (2.2)-(2.4) we obtain, from the convexity of [TAP-SC], that (h, f) 1S a

solution if and only if

hpqr (c pqr - 11"pq) 0, Vr E 'R pq , V(p, q) E C, (2.10)

cpqr - 1I"pq ~ 0, Vr E 'R pq , V(p, q) E C, (2.11 )
E hpqr d pq , V(p,q) E C, (2.12)
rE'Rpq

hpqr ~ 0, Vr E 'R pq , V(p, q) E C, (2.13)

E E
(p,q )Ee rE'Rpq
Spqrahpqr = fa, VaEA, (2.14)

f3kgk(f) 0, Vk E IC, (2.15)

gk(f) < 0, Vk E IC, (2.16)
13k ~ 0, Vk E IC, (2.17)

where cpqr is given by (2.8).

The Condition (2.11), together with (2.10) and (2.12), implies that the multiplier
1I"pq is the minimum generalized travel cost cpqr in O-D pair (p, q), and (2.10) further
states that these costs are equal for all routes utilized in the O-D pair. Hence, the
Conditions (2.10)-(2.11) imply (2.6)-(2.7), and the theorem is proved. 0
Solutions to [TAP-SC] thus correspond to flows satisfying a generalization of the
Wardrop equilibrium conditions, based on the generalized travel Costs (2.8).
One can, in general, not relate the actual travel costs of the unused routes to those
of the used ones; for instance, the cheapest route in an O-D pair may be unused
because its generalized cost is too high. Furthermore, the Wardrop principle for [TAP]
is intimately associated with the Cartesian product structure of its feasible set, and one
can for the side constrained model not state similar optimality conditions in terms of
actual travel costs. However, under the (reasonable) additional assumption that each
constraint function gk is non decreasing in each link flow variable fa, a Wardrop-type
principle in terms of actual travel costs may be established. (This result generalizes
that of Larsson and Patriksson, 1994, for the case of link flow capacity side constraints.)

Theorem 2. (A Wardrop-type principle) Let f be the link flow wlution to [TAP-Sq,

°
and suppose that
agk(f) > Va E A, Vk E JC.
afa - ,
Consider any O-D pair (p, q). Let a route r E 'Rpq be defined to be unsaturated if for
all k E IC and all links a E A on route r,

Assume, without any loss of generality, that in a corresponding route flow solution to
[TAP-SC] the first I routes are actually used, and that among these the first mare
unsaturated. Then,

C pq 1 = ... = Cpqm ~ Cpq ,m+1 ~ ••• ~ Cpq /,

and the unused routes in the O-D pair have generalized route costs that are at least as
large as that of the used routes in the O-D pair.
174

Expressing the route travel costs as

C pqr = E Dpqrata(fa), 'rfr E 'R-pq , 'rf(p, q) E C,

aeA

we obtain from (2.8) that

'rfr E 'R-pq , V(p, q) E C.

Associated with the generalized route travel costs hence are the generalized link travel
costs
ta(f) ~ ta(fa) + E 13k a~~f) , Va E A. (2.18)
keIC Ja

The interpretations of the optimal Lagrange multipliers and the Conditions (2.6)-
(2.7) depend on the form of the constraint functions gk. For example, in the case of
simple upper bounds on the link flows (K = A and ga(f) = fa - U a, U a E [0, +ooJ,
for each a E A), (2.18) reduces to ta(fa) = ta(fa) + f3a, a E A, and the multipliers f3a
may be associated with the equilibrium queueing delays on the saturated links (i.e.,
those with flows on their respective upper bounds), and the multipliers 1f'pq with the
(minimal) sum of total travel cost and queueing delay in each O-D pair; see Miller et
al. (1975), Payne and Thompson (1975), Smith (1987), and Larsson and Patriksson
(1994). (In the case of capacitated traffic assignment, the constraint functions ga are
nondecreasing, and an unsaturated route contains no saturated links.)
The reader should note that the optimal multipliers 13 are not necessarily unique.

3. SOLVING THE SIDE CONSTRAINED MODEL

Whenever side constraints are introduced in a traffic assignment model, traditional

assignment methods, such as the Frank-Wolfe algorithm and its relatives, either be-
come inapplicable or their efficiency is seriously degraded. In addition, the existing
program packages do not possess the ability to take side constraints into account.
However, when considering possible solution principles for the side constrained
model, it is most natural to aim at exploiting the efficient solution methods and pro-
gram packages that are available for the basic model. This immediately leads us to
a classical approach for handling complicating constraints: the pricing strategy (e.g.,
Lasdon, 1970, Ch. 8).
We associate with the side Constraints (2.5) non-negative prices 13k, k E K, which
define the costs of violating these constraints. Given certain values of these prices,
the side constraints are priced-out, i.e., handled implicitly by being included in the
objective function. The resulting optimization problem,

[TAP(f3)]

min L(f, (3) = T(f) + E f3kgk(f),

JeF keIC

is a standard assignment model (with, in general, a non-separable travel cost function

though) and is solvable with most standard methods for the basic model, giving the
unique link flow solution f(f3). (Note that the objective L(·, (3) is strictly convex with
respect to the link flows.)
 175

The solution to the priced-out problem may be characterized as the solution to a

side constrained assignment problem where the right hand sides of the original side
constraints are modified through certain perturbations. This result is stated below; it
follows immediately from Everett's Theorem (e.g., Lasdon, 1970, Th. 8.3).

Theorem 3. (An Everett-type result for [TAP-SCD Let (3 E ~~I be a price vector.
Then the solution f«(3) to the priced-out problem [TAP«(3)] solves the side constrained
traffic assignment problem

[TAP-SC«(3)]

min T(I),

subject to

L: hpqr dpq , V(p,q) E C,

re1lpq

hpqr > 0, Vr E 'R pq , V(p, q) E C,

L: L:
(p,q)ec
re1lpQ
Opqrahpqr fa, VaEA,

gk(l) :S 9k' Vk E!C,

where
_ def {9k(l«(3», if (3k > 0,
gk = max {O, gk(l«(3))} , if (3k = 0.

Noting that the priced-out problem [TAP«(3)] is equivalent to a standard traffic

assignment model with a link travel cost mapping of the form (2.18), a precise rela-
tionship between [TAP«(3)] and [TAP-SC] is obtained when choosing the price vector
(3 equal to a vector of optimal Lagrange multipliers for the side Constraints (2.5).

Theorem 4. (An equivalent standard assignment problem) Let (3 be an arbitrary

vector of optimal Lagrange multipliers for the side Constraints {2.5}. Then the solution
set of the standard traffic assignment model with link travel cost mapping

to = t(.) +\1g(.)(3, (3.1)

equals that of [TAP-SC].

Proof. The strict convexity of T and the discussion following Theorem 6.5.1 of Bazaraa
et al. (1993) yield that r= f«(3). The result then follows since the link travel cost
mapping of [TAP«(3)] is \1L(·, (3) = t(·) + \1g(.)(3. 0
The side constrained assignment model [TAP-SC] may thus be solved as an equiv-
alent, convex, standard traffic equilibrium problem with an appropriately chosen ad-
justment of the travel costs ta; this problem will be referred to as [TAP-A]. Hence,
the link travel cost Mapping (3.1) provides a precise statement of the influence of the
side constraints on the travel cost perception of the users of the traffic network, and
therefore on their route-choice behaviour.
The Variational Inequality problem corresponding to the first-order optimality con-
ditions of the problem [TAP-A] is to find an rEF such that

VfE F.
176

In contrast to the Variational Inequality formulation [VIP] this problem is symmetric,

since its cost mapping is integrable.
The result of Theorem 4 may alternatively be derived from Theorem 3; as the
price vector tends to a vector of optimal Lagrange multipliers, the solution 1«(3) will,
because of the strict convexity of T, tend continuously to r,
so that the right hand
sides Wk of the side constraints of the problem [TAP-SC«(3)] tend continuously to zero
and the problem [TAP-SC«(3)] tends to [TAP-SC]. (Simultaneously, the solution set of
the priced-out problem [TAP«(3)] tends to that of [TAP-SC].)
In order to find (near-)optimal values of (3 one may solve the Lagrangean dual
problem

[TAP-SCD]

max L«(3),
(3>O

where

L«(3) = min
IEF
L(f, (3).

Lagrangean dual problems are typically solved using simple iterative search meth-
ods for (essentially) unconstrained optimization. Within a dual solution procedure for
[TAP-SCD], the result of Theorem 3 may be utilized for monitoring the progress with
respect to the aim of finding a solution to [TAP-SC]. This result also facilitates the
finite termination of the dual algorithm when the solution is near-feasible with respect
to the side constraints. Clearly, near-feasible solutions are often satisfactory consid-
ering the uncertainties in the input data; near-feasibility is also satisfactory when the
side constraints are weak, in the sense that they do not need to be fulfilled exactly.
For a link capacity side constrained equilibrium model, Larsson and Patriksson
(1994) investigate and evaluate an augmented Lagrangean dualization (i.e., nonlinear
pricing) technique for finding optimal values of (3 and establish that it is more efficient
than traditional Lagrangean dualization; moreover, for certain instances of augmented
Lagrangean schemes, the sequence of dual iterates generated can be shown to converge
(at least linearly) although the set of dual solutions is not a singleton in general.

4. CONCLUSIONS AND FURTHER RESEARCH

The foundation for the development of models of equilibria based on the inclusion of
explicit side constraints is the hypothesis that the additional flow relationships which
we want to capture when modelling a real-world traffic problem may be well repre-
sented by a set of side constraints. Under this hypothesis, we may establish a close
relationship between the improvements of the basic traffic equilibrium model through
the introduction of explicit side constraints and generalized travel cost functions, re-
spectively:

(i) In the generalized travel cost approach, the proper cost function to be used has
the form (3.1).

(ii) This generalized travel cost function involves gradients of the constraint functions
gk, which are unknown since they are not formulated explicitly in this approach.
 177

(iii) The proper values of its travel cost parameters (3 are unknown, since the side
constrained problem is not solved in this approach.

Hence, the strategy of using generalized travel costs may be regarded as implicit,
which explains, at least partially, why equilibrium models with generalized travel cost
functions are difficult to calibrate, and therefore difficult to use in practice.
In contrast, the inclusion of side constraints in the equilibrium model constitutes an
explicit approach, in which the physical interpretations of these constraints facilitate
the calibration of the model. The strategy of extending the basic model with side
constraints also allows a large flexibility in the construction of the model, since the
side constraints may be nonlinear as well as non-separable. Thus, the use of side
constraints is a direct and general approach for improving the basic model.
To summarize, if explicit side constraints are utilized in a refinement of the basic
assignment model, the solution of the resulting model [TAP-SC] automatically pro-
duces the travel cost mapping of an equivalent standard traffic equilibrium model.
Hence, through a process in which one or more side constrained models are solved,
one may derive (i.e., determine the appropriate side constraints) and calibrate (i.e.,
find the proper coefficients (3) adjusted travel cost functions for use in existing trans-
portation analysis tools based on traditional equilibrium models, in order to (indirectly
through the cost functions) take into account the additional model components which
are described by the side constraints. The solution of an (augmented) Lagrangean dual
problem may then be viewed as a means for calibrating these travel cost functions.
The many possibilities for modelling traffic interactions with explicit side con-
straints, and the strong relationships to equilibrium models with generalized travel
costs, motivate the further exploration of this modelling strategy for traffic equilibrium
problems. The successful outcome of this exploration relies on cooperations between
operations researchers and users ot today's transportation planning systems.

ACKNOWLEDGEMENTS

The research leading to this report was sponsored in part by the Swedish Transport
and Communications Research Board (Grant TFB 92-128-63). The report was written
while the second author was on leave at the Department of Mathematics, University
of York, Heslington, York, England. The second author wishes to thank Professor
Michael J. Smith of the University of York for his hospitality and many interesting
discussions, and the Swedish Transport and Communications Research Board (Grant
TFB 93-131-63), the Swedish Institute (Grant 303 GH/MLH), and the Royal Swedish
Academy of Sciences for financing the visit.

REFERENCES

[1) M.S. Bazaraa, H.D. Sherali and C.M. Shetty, "Nonlinear Programming: Theory and Algorithms,"
second ed., John Wiley & Sons, New York, NY, 1993.
(2) M. Beckmann, C.B. McGuire and C.B. Winsten, "Studies in the Economics of Transportation,"
Yale University Press, New Haven, CT, 1956.
(3) S.C. Dafermos, "The traffic assignment problem for multicJass-user transportation networks",
Transportation Science 6:73-87, 1972.
178

[4] D.W. Hearn, "Bounding flows in traffic assignment models", Research Report 80-4, Department of
Industrial and Systems Engineering, University of Florida, Gainesville, FL, 1980.
[5] T. Larsson, and M. Patriksson, "An augmented Lagrangean dual algorithm for link capacity side
constrained traffic assignment problems", in: "Proceedings of the 2nd Meeting of the EURO
Working Group on Urban Traffic and Transportation, Paris, France, September 15-17, 1993,"
vol. 38 of Actes INRETS, F. Boillot, N. Bhouri, and F. Laurent, eds., Institut National de Recherche
sur les Transport et leur Securite (INRETS), Arcuei!, France, pp. 163-199. Also as Report LiTH-
MAT-R-93-22, Department of Mathematics, Linkiiping Institute of Technology, Linkiiping, Sweden,
1993.
[6] L.S. Lasdon, "Optimization Theory for Large Systems," Macmillan, New York, NY, 1970.
[7] S.D. Miller, H,J. Payne, and W.A. Thompson, "An algorithm for traffic assignment on capacity
constrained transportation networks with queues". Paper presented at the Johns Hopkins Confer-
ence on Information Sciences and Systems, The Johns Hopkins University, Baltimore, MD, April
2-4, 1975.
[8] A. Nagurney, "Network Economics: A Variational Inequality Approach," Kluwer Academic Pub-
lishers, Dordrecht, The Netherlands, 1993.
[9] M. Patriksson, "The Traffic Assignment Problem: Models and Methods," VSP, Utrecht, The
Netherlands, 1994.
[10] H,J. Payne, and W.A. Thompson, "Traffic assignment on transportation networks with capacity
constraints and queueing". Paper presented at the 47th National ORSA Meeting/TIMS 1975
North-American Meeting, Chicago, IL, April 30-May 2, 1975.
[11] J.G. Sender, and M. Netter, Equilibre offre-demande et tarification sur un reseau de transport,
Rapport de recherche 3, Departement Economie, Institut de Recherche des Transports, Arcueil,
France, 1970.
[12] M.J. Smith, "Traffic control and traffic assignment in a signal-controlled network with queueing".
Paper presented at the Tenth International Symposium on Transportation and Traffic Theory,
Boston, MA, 1987.
[13] J.G. Wardrop, "Some theoretical aspects of road traffic research", Proceedings of the Institute of
Civil Engineers, Part II, 325-378, 1952.
ADVANTAGES AND DRAWBACKS OF
VARIATIONAL INEQUALITIES FORMULATIONS

Marcotte Patrice

Departement d'informatique et de recherche operationnelle

Universite de Montreal
CP 6128 Succursale Centre-Ville
Montreal, Canada H3C 3J7
e-mail: patricem@crt.umontreal.ca

Abstract: In recent years, the Variational Inequality framework has been recognized
as a powerful modelling tool in operations research and economics, whenever equilib-
rium relationships among several agents occur. This interest has prompted researchers
to develop efficient algorithms for solving Variational Inequalities in finite dimensional
spaces. In this presentation we will try to assess the advantages and possible drawbacks
of Variational Inequality formulations, focusing on four problems: oligopoly models,
traffic assignment, bilevel programming, multicriterion equilibrium. Each topic will be
analyzed from the modelling and computational points of view. In particular we will
see whether the Variational Inequality formulations of these models has led to a better
understanding of their inner structure and to efficient solution algorithms.

Key Words: Variational Inequalities, oligopoly, traffic assignment, bilevel program-

ming, multicriterion modelling.

1. INTRODUCTION

In this paper we consider four equilibrium situations inspired from the realm of
Operations Research and Economics, that can be formulated as Variational Inequal-
ities, and discuss whether this framework is adequate from either the modelling or
computational points of view. Our approach is not encyclopaedic but pragmatic, and
biased towards our own research interests. We basically wish to answer ("raise" would
perhaps be more adequate) two fundamental questions:
1. Have Operations Research (or Economics) models benefited from recent advances
in the field of Variational Inequalities?

2. What developments would be required in order to make the Variational Inequality

framework more attractive, computationally?

179
180

The second question has been motivated by the fact that the Variational Inequality
formulation of practical situations frequently leads to nonmonotone mappings, and
that little research has been devoted to nonmonotone Variational Inequalities. One of
the goals of the present paper is to motivate research in this direction.

2. NOTATION AND BASIC DEFINITIONS

Let F be a mapping from the closed, convex set Xc R n into Rn. A vector x* is
a solution of the Variational Inequality V I(F, X) if x* E X and satisfies

(F( x*), x* - x) ~ 0, VxEX, (2.1)

where h .) denotes the inner product in Rn. We say that the mapping F is
- monotone on X if

(F(x) - F(y),x - y) ~ 0, Vx,y E X,

- strictly monotone on X if

(F(x) - F(y), x - y) > 0, Vx,y E X (x:f. y),

- strongly monotone with modulus fl on X if

(F(x) - F(y),x - y) ~ flllx _ y1l2, Vx,y E X,

- co-coercive with modulus 'Y on X if

(F(x) - F(y),x - y) ~ 'YIIF(x) - F(y)1I 2 , Vx,y E X,

- pseudomonotone on X if

(F(x),x - y) ~ 0 => (F(y),x - y) ~ 0, Vx,y E X,

- pseudo co-coercive with modulus u on X iff

(F(x),x - y) ~ 0 => (F(y),x - y) ~ uIlF(x) - F(y)1I 2 , Vx,y E X.

If F is pseudomonotone on X, then x is a solution of V I( F, X) if and only if it lies in

X and if it solves the dual Variational Inequality

(F(y), x - y) ~ 0, VyEX.

A function g defined on X is a gap function for V I( F, X) iff it possesses the following

three properties:

1. g(x) ~ 0, Vx E X.

2. g( x) = 0 iff x is a solution of V I( F, X).

Most gap functions also possess the property:

3. If g( x) > 0, then there exists y in X such that the directional derivative g'( Xj y-x)
is negative.
 181

The simplest gap function is the primal gap function defined, for compact feasible
sets X, as
9(X) = max(F(x),x - y).
yEX

Based on the dual Variational Inequality, one can also define the dual gap function

g(x) = max(F{y),x
yEX
- y).

If the set X is not compact, one can use the projective gap function introduced by
Fukushima [I]:
1 IIx - YII~,
9B{X) = max(F{x), x - Y)B - -2
yEX a
where (" ')B and II . liB denote, respectively, the scalar product and norm associated
with a symmetric, positive definite matrix B, and a is a positive parameter. Several
other gap functions have been proposed in the literature.
The projective gap function 9B is differentiable but strongly nonconvex and difficult
to evaluate, unless the set X assumes a simple form. The primal gap function 9, on
the other hand, is simple to evaluate, weakly nonconvex (if F is affine and monotone,
9 is convex), but nondifferentiable. The dual gap function is convex, weakly nondif-
ferentiable (it is differentiable if F is affine and strongly monotone), but difficult to
evaluate.
Any gap function 9 can be used to transform V I{ F, X) into the mathematical
program
ming(x).
xEX

3. OLIGOPOLY MODELS

Consider a set of n firms competing to sell a product on a single market. Each firm
i E I = {I, ... ,n} is characterized by a production cost Ci{q;) that depends on its own
output qi, and the demand market by an inverse demand function p that relates the
market price p( Q) to the total output Q = L,iEI qi. A Nash-Cournot equilibrium, if
one exists, is a vector q* that satisfies the conditions

q; E arg max !i(qi) ~ qiP(qi

q.~o
+ #i
Lqj) - Ci(qi), Vi E I,

where the "max" operator should be interpreted in the global sense. Standard assump-
tions on the functions Ci and pare:

Assumption Al The function p is nonnegative and decreasing on R+.

Assumption A2 The function Qp(Q) is a proper concave function of its argument

Q.

Assumption A3 The functions Ci are positive, convex and increasing Vi E I.

Assumption A4 For every i in I, there exists a nonnegative quantity qi such that

qiP(7j;) - Ci(q;) is negative.
Assumption A5 All functions involved are continuously differentiable on R +.
182

Under assumptions AI-A4, the solution set of VI(F,X) is nonempty, convex

and compact, and each profit function J; is concave with respect to qi. If Assumption
A5 is satisfied, the function J; achieves, for given {q;}jel,i¢i, its maximum at any point
qi that satisfies the first-order optimality conditions
P(Lqi) + qiP'(Lqj) - C:(qi) ::::; 0,
iel iel
qi[P(L qi) + qiP'(L qi) - C:(qi)] = 0,
iel iel
qi ~ O.
This nonlinear complementarity problem can be reduced to the Variational Inequality
VI(F,R!) if one sets

jel iel
Under assumptions AI-A5, the mapping F is not necessarily monotone. A sufficient
condition for F to be monotone is that the inverse demand function p be convex (Good-
man [2]). If this is the case, traditional algorithms can be implemented. Tiitonnement
processes, for instance the Jacobi scheme, have been suggested by Marcotte [3] and
Harker [4]. They require, at each major iteration k, the resolution of the n single-firm
mathematical programs:
q; E argmax qiP(qi + Lqj) - Ci(qi).
q.~O JEl
,tli

At iteration k + 1 we let qk+ 1 = tqk + (1- t)qk where t (0 < t ::::; 1) is an underrelaxation
parameter. The convergence proof given in [4] requires that the mapping F be strongly
monotone. This condition will be satisfied if p is strongly convex or if the c;'s are
strongly convex. Furthermore, the parameter t must be less than some threshold value
l that depends on the strong monotonicity modulus of F. While an overestimate of t
could lead to nonconvergence of the iterative process, an underestimation could affect
negatively the convergence rate of the method.
Clearly, this is a situation where the variational (or complementarity) formula-
tion hides the salient features of the problem. Computationally efficient algorithms
are more easily designed around the formulation of the oligopoly problem as a one-
dimensional equation. Indeed, if a total equilibrium output Q were known a priori, the
corresponding equilibrium outputs could be easily obtained from the conditions
qi= 0 if P(Q) - c:(O) ~ 0, (3.1)
p(Q) + qiP'(Q) - C:(qi) = 0, otherwise. (3.2)
These conditions are the first-order optimality conditions of the separable concave
program parameterized in Q

m~ P(Q)Lqi+-21 p'(Q)Lq?- LCi(qi).

L iel iel iel
Since we are looking for a solutions of the above program that are compatible with the
total output Q, it is natural to investigate the constrained program

max
q>O
p(Q) ~ qi + -21p'(Q) ~ q; - ~Ci(qi) (3.3)
iel iel iel
subject to ~qi=Q.
iel
 183

If the inverse demand function is strictly decreasing, the solution q( Q) of (3.3) is

unique. Let >.(Q) denote the optimal dual variable associated with the unique con-
straint of (3.3). Murphy, Sherali and Soyster (5) showed that >.( Q) is a continuous and
nonincreasing function of Q and that Q is an equilibrium total output if and only if
>.( Q) = O. Based on these properties, these authors proposed a binary search algorithm
for locating the interval containing the optimal Q-values. They also proved that, if the
functions Ci'S are strictly convex, or if Qp(Q) is a strictly concave function of Q, then
the equilibrium solution is unique.
Another line of approach, which does not require the functions co's and p to be
differentiable, is based on the backward reaction correspondence

B(Q) = Lqi(Q),
iel
where qi(Q) is the set (interval) of optimal solutions of (3.1)-(3.2). It is shown in
Marcotte [6] that any solution Q is an optimal total output only if 0 E I}i( Q) ~
B(Q) - Q. Conversely, if 0 E I}i(Q), any set of outputs {qi(Q)}.eI that satisfies the
compatibility equation Eiel qi( Q) = Q yields an equilibrium. As in [5), a zero of I}i( Q)
can be located by dichotomic search. Contrary to the function >.( Q), the function B( Q)
is not necessarily monotone decreasing on the entire region R+. However, it has been
shown in (6) that one can restrict the binary search to an interval where the function
B( Q) is decreasing. This interval is [Qrnin, Qmax] where Qrnin is defined as

Qrnin = lI)ax{
.eI
q;run}
q;run = argm~qiP(qi)
q,-
- Ci(qi)

and Qmax as the marginal cost total output

Qmax = ""'
L..Jqimax
iel
qffi&X = argmax
q~O
L0
EiEr qi ""'
p(t)dt - L..JCO(qi).
iel
Such results, intimately related to the problem's structure, cannot be derived from its
Variational Inequality formulation.
On a small five-firm example used in [5) our algorithm, coupled with the false
position method in its last iterations, could determine the equilibrium production
vector to 7 significant digits in 6 iterations. This compares favourably with Murphy's
method (7 iterations) and the Jacobi method (20 iterations). In the particular case
where all functions are affine, a finitely convergent greedy algorithm would be still
more efficient (see Marcotte [3)).
The situation is modified altogether when additional constraints are taken into
account. For instance, Marcotte [7) considered the case of firms shipping their products
to demand markets through a congested transportation network. Let us associate each
firm with an origin node i E I of the network, and each demand market k E K with
a destination node. Denoting by v~ the flow from origin (firm) i on arc a E A of the
network, dj. the quantity sold by firm i on market k, Sa the transportation cost function
on arc a and Pk the inverse demand function of market k, the ith firm payoff function,
to be maximized, is given by

/o(v!,d7)aeA,keK =- L V!Sa(LV~) + L d7Pk(Ldi).

aeA jel keK jel
184

Each firm's flow vector must satisfy its own flow conservation equations. The interac-
tion between firms takes place both on the network, through congestion, and at the de-
mand markets, through competition. Let us introduce the vectors Xi = (v~, di.)aEA,kEK,
x = (Xi)iEI and the set X of feasible x-vectors. Assuming that the payoff functions
are concave, a vector x is a Cournot-Nash equilibrium for the network oligopoly model
if and only if it satisfies the Variational Inequality VI(F,X) where the ith block F;
of the mapping F corresponds to the ith firm negative gradient with respect to its
decision variables: Fi(X) = -Vdi(X) Vi in I. A sufficient condition for VI(F,X) to
be monotone is that Goodman's condition be satisfied, i.e. that the demand functions
Pk be convex and that the functions Sa be concave, a rather unusual assumption for
congestion functions!
The structure of this model is characterized not so much by its underlying network
structure (see next section) than by the shape of the payoff functions. It is convenient
to linearize the functions Sa and Pk at the current iterate, since this yields a linear
Variational Inequality involving a gradient mapping, which can be transformed into a
concave quadratic maximization program. One must be aware that this linearization
scheme is not Newton's method, and that one should not expect the convergence
properties of the latter. A better approximation of the cost mapping F is obtained
through a partial linearization scheme that keeps, a la Jacobi, the nonlinear part of
F; associated with firm i, for all i in I. On the set of test problems solved in [7]
the linearization algorithm was preferable, in terms of accuracy and computer time,
to the nonlinear approximation, Jacobi and Gauss-Seidel algorithms. All algorithms
converged, whenever they did converge, to the same equilibrium solutions which, we
conjecture, might be unique.
Let us finally mention that, as the number of firms grows, the equilibrium solutions
converge to a Wardrop equilibrium on the transportation network (see Haurie and Mar-
cotte [8]). Although Wardrop equilibria are not necessarily Cournot-Nash equilibria,
this result shows that they can be obtained as limits of the above network oligopoly
equilibria.

4. THE TRAFFIC ASSIGNMENT PROBLEM

In this section we only consider the fixed demand, static traffic assignment problem
(TAP in short), although most of the discussion could be extended to the VRl'iable
demand case. The static TAP has long served (still does!) as a benchmark for testing
Variational Inequality algorithms. This is due as well to the practical significance of
the problem and the availability of data, as to its specific decomposable structure.
The traffic assignment problem consists in determining flows on a congested trans-
portation network. Let Pod denote the set of paths joining the origin node E 0 °
to the destination node d E D, h;d the flow on a path P E Pod, god the demand for
°
transportation between nodes and d, and C,,( h) the generalized transportation cost
of traveling along path p, given a path-flow vector h = (h;d)(o,d)EOXD,"EPod ' By defini-
tion (Wardrop), an equilibrium h* is reached when flows are concentrated on shortest
paths with respect to the current path-cost vector C(h*), i.e., h* solves the Variational
Inequality V I(C, H), where H denotes the set of feasible path-flow vectors:
H = {h;dl E h;d = gOd, V(o, d) E 0 X D}.
"EPod

The cost of path P is the sum of the costs along the arcs making the path p. Therefore
 185

we have v = Ah and
C(h) = AtF(Ah)
where A denotes the arc-path incidence matrix of the network and F is the arc-cost
mapping.
We will also consider the so-called arc-flow formulation of the TAP where one
searches for a total arc-flow vector v· compatible with a path-flow vector h· in Hand
such that
(F(v·),v· - v) :::; a
for all total arc-flow vector v compatible with some path-flow vector h in H. If the
Jacobian matrix V F( v) is symmetric for all v, then F is the gradient of some function
f, and VI(C,H) can be reduced to the problem of finding stationary points of the
mathematical program
mm f(v)
v,h
subject to v= Ah
hEH.
In this case, the Frank-Wolfe algorithm is the only algorithm able to exploit efficiently
the network structure of the TAP. At each iteration, it solves the linear program
obtained by substituting to f its first-order approximation at the current iterate vi:
mm f(v l ) + (Vf(vl),v-v l )
v,h
subject to v =Ah (4.1)
hEH.
This subproblem can efficiently be solved by computing shortest path trees rooted
at the origin nodes 0 E O. Let wi be the total arc-flow corresponding to one of its
solutions and set
(4.2)
The fact that the multi commodity character of the problem is "hidden" in the shortest
path subproblems makes this method especially attractive. However this algorithm is
difficult to extend to the case where F is not a gradient mapping (see Marcotte [9]
however) and its convergence rate is not good.
Researchers have tried to overcome the slow (typically sublinear) order of conver-
gence of the Frank-Wolfe algorithm by working directly in path-flow space. Since the
number of paths is in general prohibitive, it is convenient to solve, at a given iteration,
a Variational Inequality restricted to the convex hull HR of a restricted number of
extreme points of the set H:

HR = n::: Arhrl L Ar =
rER rER
1, Ar ~ a}.

This leads to the Variational Inequality: Find AR in the IRI-dimensional simplex SR

such that

rER

If the vector AR is not a solution of the original Variational Inequality, then the set R
can be enlarged by incorporating into it the path-flow solution obtained by solving the
linear program
186

A more disaggregated approximation can be performed by generating one "column"

(path) for each set of origin-destination variables, thus working with 101 x IDI convexity
constraints at each iteration. Both schemes work well if an equilibrium solution can be
constructed using few extremal solutions, which is frequently the case in practice. In
order to reduce the computational and memory requirements, it is possible, if certain
conditions are met, to drop some of the extreme points generated (see Lawphongpanich
and Hearn [10] for the details). The convergence of a path-flow sequence {hi} to an
equilibrium can be measured using a gap function such as
g(hl) = max
hEH
(C(hl), hi - h).

Even if no extreme points are dropped, there is no guarantee that the sequence {g(hL)}
be decreasing, nor that the cardinality of R can be limited a priori, although it has
been observed that a very good approximation to an equilibrium solution could be
constructed from the convex combination of less than twenty extreme points. Indeed,
in disaggregated mode, three or four paths per origin-destination pair can be combined
to yield quasi-equilibrium path flows.
There remains to propose a procedure for solving the restricted subproblems. Define
G(>.) = CO::::rER>'rhr). If F is strongly monotone, the mapping C is co-coercive and
the sequence generated by the rule
>.+ = ProjsR(>' - aG(>.»
converges to a solution of the restricted Variational Inequality (see Marcotte and Zhu
[11]) provided that the parameter a is sufficiently small.
In practice, the cost mapping is not strongly monotone, but monotone at best.
Gap-decreasing methods can be adapted to this situation: Newton's direction is a
descent direction for the primal gap function (see Marcotte and Dussault [12]); the
projection
>: = ProjsR(>' - aG(>.»
induces a descent direction X - >. for Fukushima's projective gap, provided that the
parameter a be larger than some threshold value a that can be updated between
iterations (see Marcotte and Zhu [13]).
Since restriction-based algorithms work in path-flow space, a theoretical challenge
consists in devising a limited-memory method operating in arc-flow space. One could
for instance mimic the Frank-Wolfe algorithm and replace the linesearch (4.2) by the
one-dimensional Variational Inequality VI(F, [vl,w l ]), where wi is now the partial v-
solution of the linear program
min f(v l ) + (F(vl),v - vi)
v,h
v =Ah (4.3)
hEH.
If the Jacobian V F is "weakly" asymmetric, this heuristic scheme provides a reasonable
approximation. If V F is strongly asymmetric however, the method could zigzag, or
even cycle. This could be fixed by using a smaller stepsize, for instance:
1
+
VI 1 = Vi + _(wi - vi).
1
This constitutes a generalization of the well-known Brown-Robinson's fictitious play
algorithm (see Robinson [14]) for solving zero-sum matrix games. We conjecture that
it can be used to solve monotone Variational Inequalities as well.
 187

The TAP is by now a well-understood large-scale optimization problem that can be

solved by several known methods that are both numerically efficient and theoretically
convergent. The main challenge is now the development of algorithms for solving
dynamic models of traffic flow. These are best formulated as infinite-dimensional,
highly nonmonotone Variational Inequalities (see Wu [15]) for which only heuristic
procedures have been proposed yet.

5. THE GENERALIZED BILEVEL PROGRAMMING PROBLEM

Bilevel programming problems are two-stage decision problems where a leader op-
timizes her objective, taking into account the reaction to her course of action of a
follower, whose program (objective and/or constraint set) depend on the leader's de-
cisions. Players do not cooperate although, given two equivalent alternatives, the
follower will choose the one most favourable to the leader. Mathematically we have:

mill
r
F(x, y)
subject to (x,y)EX (5.1)
yEarg min J(x,z).
zEY(r)

Denote by S(x) the set of optimal solutions of the lower level program; then (5.1) can
be reformulated as the one-level program

mIn
r,Y
F(x, y)
subject to (x,y)EX
y E S(x).

This is a difficult problem. Even in the linear case, checking the local optimality of a
solution (x, y) is strongly NP-Hard (Hansen, Jaumard and Savard [16]).
It occurs frequently that the lower level variables are controlled by several, if not
infinitely many, agents or players. For instance, the x-variables might represent energy
taxes set by the government, and y the industrial sector's reaction to those tax levels.
The governments's objective function F could embody financial as well as economic,
social or environmental goals. At the lower level, the industries are price takers, and
seek to maximize their own profit functions. Assuming that they behave according to
the Cournot-Nash principle and that certain conditions are satisfied (see section on
oligopoly problems), their equilibrium state can be represented as a Variational In-
equality. We then obtain the generalized bilevel program where S( x) above is replaced
by the solution set S(x) of a Variational Inequality with cost mapping G, i.e., find y
in Y(x) such that:
(G(x,y),y - z) :S 0, Vz E Y(x).
There are several ways to reduce a generalized bilevel program to a single-level program.
One can for instance express S( x) as the set of points for which the primal gap function
is equal to zero:

S(x) = {ylg(x,y) ~ zEY(r)

max (G(x,y),y - z) = O}
to obtain the standard mathematical program:

mill
r,Y
F(x,y)
188

subject to (x,y) E X (5.2)

g(x,y) = O.

Obviously, any other gap function could have been used for that purpose. Even under
strong assumptions, the constraint g(x, y) = 0 in (5.2) is nonconvex, and cannot be
expected to satisfy any constraint qualification. By penalizing the nonconvex term,
one obtains the program
min F(x,y) + Kg(x,y). (5.3)
(X,II)eX

Under suitable assumptions (see Marcotte and Zhu [17]), this penalty is exact, i.e.,
there exists a threshold value K such that, for any K ~ K, a global solution of (5.3)
is a global solution of (5.2).
If F is monotone in y, for fixed x, and the set Y(x) is given explicitly as

Y(x) = {ylhj{x,y)::; 0, j E J},

where the functions h j are convex in y and satisfy some constraint qualification, then
the lower level Variational Inequality can be replaced by its Kuhn-Tucker necessary
and sufficient conditions, and the bilevel program (5.2) written as

mm F(x,y)
(x,y)eX

subject to G(x,y) +:E ujVhj(x,y) = 0 (5.4)

jeJ
Uj ~ 0
(uj, hj(x, y») = 0, j E J.

The complementarity term can be penalized to yield the program

min
("',!I)eX
F(x,y) +KEujVhj(x,y) =0
jeJ
subject to Uj ~ 0 (5.5)
(Uj, hj(x, y») = 0, j E J.

Under certain conditions, for instance if all functions involved are affine, it can be
shown that a solution of (5.5) is globally optimal if and only if it is also globally
optimal for (5.4) (see Luo, Pang, Ralph and Wu [18]).
If the set Y(x) is a compact polyhedron Y = {By::; b} independent of x, one can
write the bilevel program as

mm F(x,y)
(x,y)ex
(G(x,y),y - yi) ::; 0, Vi E I

where {yihel is the (finite) set of extreme points of Y. Based on this formulation,
Marcotte [19] has suggested the following constraint generation scheme
 189

CONSTRAINT GENERATION ALGORITHM

Step 0 E +- 0; i +-1
Step 1 Solve

mm F(x,y) (5.6)
(r,II)EX
subject to (G(x,y),y_yi):50 VieE

and let (x, y) be one of its global solutions.

Step 2 Let yi be an extremal solution of the linear program

min (G(x, y), x - y).
IIEY

Step 3 if (G(x, y), x - yi) :5 f then stop and record the approximate solution
(x, y(x» where y(x) is an optimal lower-level reaction to x.
else go to Step 4 below.

Step 4 E +- E U {yi}; i +- i + 1; return to Step 1.

This method can only be efficient if the number of additional constraints in (5.6) is
small and if the nonconvex program (5.6) can be solved to global optimality. Otherwise,
one must resort to heuristic procedures. If the set X only involves the x-variables, one
such procedure, that works well for "weak" bilevel programs where the leader's and
follower's objectives are not too far apart, is to replace the original bilevel program by
the Variational Inequality: Find (x, y) «x, y) eX, ye Y(x» that satisfy

(5.7)

If F is convex in x and G( x, y) = 'V IIf( x, y) for some function f convex in the variable
y, then (x, y) is a Cournot-Nash equilibrium for a game involving two players having
objective functions F and f. Under suitable assumptions, the Variational Inequality
(5.7) is monotone and can be reduced to a convex optimization problem (see Marcotte
and Marquis [20]).
Bilevel programs are computationally intensive. Exact algorithms, based on the
implicit enumeration of the lower-level optimal faces, have only been tested on small
instances involving linear or quadratic objectives and constraints. Even local optimality
results are difficult to obtain (see Hansen, Jaumard and Savard [16] and Gauvin and
Savard [21]) and the theory of sensitivity analysis for Variational Inequalities is of no
great help, as local optima may be numerous.
In view of the intractability of the general problem, the main challenge will be
the development of efficient heuristic procedures able to solve to near-optimality large
instances of bilevel programs. These will have to be adapted to the specific structure
of the problem they intend to solve.
190

6. A BICRITERION PROBLEM

One shortcoming of the traffic assignment model presented in the next-to-Iast sec-
tion is that it yields nonunique equilibrium path flows, even when arc flows are unique.
This can be remedied by considering stochastic versions of the TAP (see Fisk [22]),
where all paths carry some amount of flow. The larger its cost, the lower will be the
amount of flow carried by a given path. These models have shortcomings of their own:
(i) they can only operate in path-flow space and require, a priori, the knowledge of
all paths, of which there are exponentially many (ii) they cannot discriminate well
between almost identical paths, e.g., paths sharing several arcs.
In this section, we consider a bicriterion extension of the TAP that possesses none
of these drawbacks and that can be implemented in arc-flow space, even when the
cost mapping is not a gradient mapping. This extension is formulated in an infinite-
dimensional space, but all steps of the proposed solution algorithm can be implemented
finitely, without resorting to discretization of any kind. Although it remains valid in a
wider context, we will restrict our analysis to the realm of transportation.
Let the generalized cost of traveling along an arc a of a network assume the form
1
Fa(V) = Ca(V) + -Go
a
where Ca(Vo) represents the traversal time of arc a, which depends on the total arc-
flow vector V, Go represents the out-of-pocket cost of travel along arc a, and a is a
parameter converting time units into cost units. The parameter a is referred to as the
"value-of-time" parameter. Following Leurent [23], we assume that the parameter a
varies continuously across the population of network users, and that its density function
h( a) is continuous and strictly positive on [0,00), with h(O) = O. Let us introduce the
notations:

K: The set of commodities (origin-destination pairs)

A: The set of network arcs
N: The set of network nodes
i+ (i-) The forward (backward) star associated with i E N
dk : The demand for commodity (origin-destination) k E K
v!(a): The flow density of commodity k along arc a with respect to a
Vo(a): The total flow density of class a along arc a: Vo(a) = LkEK v!(a)
v(a): The flow vector v!(a), a E A, k E K
Va: The total flow along arc a: Va = LkEK f;' v! (a ) da = fooo Va (a ) da
Go: The fixed monetary cost of travel along arc a
V = {Vo}oEA
Co(V): The delay (congestion) function associated with arc a
G = {GO}OEA
C = {Co}oEA.
The feasible set n of our problem corresponds to those non-negative flow density
functions v! that satisfy the flow conservation equations

dkh(a), if i = origin of k,
2;: v!(a) - 2;: v!(a) = { -dkh(a), if i =.destination of k,
oE.+ aE.- 0, otherwise.
 191

Throughout the section, we will assume that the mapping C is monotone and con-
tinuously differentiable on n. In our assignment model, equilibrium is reached when,
for every possible value of time a, the flow density vector {v:(a)}aEA is assigned to a
shortest path from the origin to the destination node of k, with respect to the value
a, given the congestion levels Ca(V) on the network, e.g.,

L L(v:(a) - y:(a))(Ca(V) + Ga/a) ~ 0

kEKaEA
for all feasible flow patterns y:(a), a E A, k E K. This is equivalent, modulo a set
of measure zero, (see Dafermos [24)) to finding a vector v in n and its corresponding
total flow vector V = LaEA,kEK v: satisfying the monotone Variational Inequality

(6.1)

for all flow vectors y in n compatible with the feasible total flow vector Y. If the
congestion functions are separable by arc, i.e., Ca(V) only depends on the flow Va, then
(6.1) is equivalent to finding a minimum point of the convex mathematical program

min L
vEO aEA
{l
0
Va
Ca(u)du 1
+ Ga 00 _a_da.
0
v; (a) }
a

This infinite dimensional problem can be solved efficiently using Frank and Wolfe's
linear approximation algorithm, without resorting to numerical discretization or path
enumeration. The convergence of the method follows from a result established by
Auslender and Brodeau [25].
For a current solution vector v and corresponding total flow vector V, the linear
subproblem resulting from the linearization of the objective with respect to the flow
variables is:

min L L(Ca(Va) + Ga/a)y:(a), Va ~O.

yEO kEKaEA
This problem is separable by commodity. For a given commodity k and value of time
a, its solution can be found by computing a shortest path corresponding to k, based
on the current arc costs Ca(Va) + Ga/a. Actually, all these shortest paths can be
computed in one pass by solving a parametric shortest path problem with respect to
the parameter a. This can be performed efficiently using a variant of the network
simplex method. The solution to this problem will take the form

if arc a lies on a the computed shortest path corresponding to commodity k and

value of time parameter a, and y:( a) = 0 otherwise. Let Ya( a) = LkEK y:( a), and
Ya = fooo Ya( a) da be the associated total arc flow. The next iterate is then determined
by solving the one-dimensional problem

mm
tE[O,l)
L 1
aEA 0
Va +t (Ya- Va)
Ca(u)du+Ga
100 -[v,,(a)+t(Ya(a)-Va(a))]da
0
1
a
= rjJ(t) + (1 - t)Q + t R
192

R = L"eA G" Iooo Y(o.)/o. do.. Now let

t* E arg min ¢J(t) + t(R - Q)

te[O,I]

and update
v.,+ = (1 - t*)V" + t*Y;..
It is remarkable that the above update can be realized without explicit reference to
the commodity flow variables, exactly as for the standard traffic assignment problem.
One must only be careful to keep track of the integrals I v" (0. ) / 0. and I Y;. ( 0. ) / 0.. This
is simple; indeed, by linearity:

One can obtain a closed form formula for the latter integrals only if the function h( 0.) / 0.
admits a closed form primitive. If this is not the case, an approximate but closed form
expression can be derived from polynomial quadratures for h(o.). It follows from the
previous discussion that the arc-separable bicriterion traffic assignment problem can
be solved in a similar fashion to the standard problem, with a parametric shortest path
problem being substituted to a standard shortest path problem.
If the delay mapping C is not arc-separable but its Jacobian 'VC is symmetric, the
Variational Inequality (6.1) can still be reduced to a convex optimization problem. If
the Jacobian 'VC is not symmetric, a slightly modified Frank-Wolfe algorithm can be
used to solve it, under the following "nondegeneracy" assumption:
Assumption A: For any two paths PI and Pi joining the origin node to the destina-
tion node of k, one has:
L G"f L G,..
,.ep~ ,.eP:
It is always possible to induce Assumption A through a suitable perturbation of
the cost vector G.
Let vI be the feasible flow vector at iteration 1 and V' the corresponding total flow
vector. Under Assumption A, the functional solution (yl, yl) of the linear program

mIn (C(V') + !.G, Y) (6.2)

yerl,Y 0.

subject to Y=LY"
"eK
is unique, except for possibly a finite number of values of the state variable 0.. This
remark implies that the primal gap function g( v) is continuously differentiable. More-
over, the extremal solution of (6.2) induces the feasible descent direction d' = yl-v l for
9 at vI. The next iterate V ' +1 can be determined by minimizing 9 along the direction
d' or by setting

where

0.1 ~ 0
lim
1-+00
0.1 =0
00

Lo.l = +00.
1=1
 193

Once the solution yl is obtained, all the calculations can be performed with respect to
total flows, and one can forget the functional flows v. The convergence of the method
can be established by adapting a proof of convergence of a gap-decreasing algorithm
for solving Variational Inequalities defined over strongly convex sets (see Auslender
[26)). In both cases the solution of the linear approximation problem (6.2) is unique.
The uniqueness of the correspondence y( v) has other interesting consequences. In
particular, at any equilibrium solution v·, one has that y( v·) = v·. This result is
independent of any monotonicity assumption on the mapping C. Furthermore, given
only the equilibrium arc costs C(V*), a full equilibrium solution can be recovered as
the unique y-solution of (6.2).
The basic model presented in this section can be enhanced to involve flow de-
pendent functions Ga(V) and/or more than two criteria. Unfortunately, the resulting
Variational Inequality formulations are always nonmonotone. It would be interesting to
know whether they satisfy some generalized monotonicity condition, or can nonetheless
be solved efficiently.

REFERENCES
[1] M. Fukushima, "Equivalent differentiable optimization problems and descent methods for asym-
metric Variational Inequalities", Mathematical Programming 53,99-110, 1992.
[2] J. C. Goodman, "A note on existence and uniqueness of equilibrium points for concave N-person
games" ,Econometrica 48, 251, 1980.
[3] P. Marcotte, "Quelques notes et resultats nouveaux sur Ie probleme d'equilibre d'un oligopole",
RAIRO Recherche Opirationnelle 18,147-171,1984.
[4] P. T. Harker, "A Variational Inequality approach for the determination of oligopolistic market
equilibrium", Mathematical Programming 30, 105-111, 1984.
[5] F. H. Murphy, H. Sherali and A. L. Soyster, "A mathematical programming approach for deter-
mining oligopolistic market equilibrium", Mathematical Programming 24,92-106, 1982.
[6] P. Marcotte, "Un algorithme general de calcul de l'etat d'equilibre d'un oligopole", RAIRO
Recherche Opirationnelle 22,225-241, 1988.
[7] P. Marcotte, "Algorithms for the network oligopoly problem" , Journal of the Operational Researcl,
Society 38, 1051-1065,1987.
[8] A. Haurie and P. Marcotte, "A game-theoretic approach to network equilibrium" Mathematical
Programming Study 26, 252-255, 1986.
[9] P. Marcotte, "A new algorithm for solving Variational Inequalities, with application to the traffic
assignment problem", Mathematical Programming 33, 339-351, 1985.
[10] S. Lawphongpanich and D. W. Hearn, "Simplicial decomposition of the asymmetric traffic assign-
ment problem", Transportation Research B 18B, 123-133,1984.
[11] Zhu and Marcotte "Co-coercivity and its role in the convergence of iterative schemes for solving
Variational Inequalities", Publication CRT-926, Centre de recherche sur les transports, Universite
de Montreal, 1993.
[12] P. Marcotte and J.-P. Dussault, "A note on a globally convergent Newton method for solving
monotone Variational Inequalities", Operations Research Letters 6, 1,35--42, 1987.
[13] P. Marcotte and D. Zhu, "Modified descent methods for solving the monotone Variational Inequality
problem", Operations Research Letters 14, 111-120, 1993.
[14] J. Robinson, "An iterative method of solving a game", Annals of Mathematics 154, 296-301, 1951.
194

[15) J. H. Wu, "Dynamic network equilibrium problem formulated as an infinite dimensional Variational
Inequality problem", Publication CRT-952, Centre de recherche sur les transports, Universite de
Montreal, 1993.
[16) P. Hansen, B. Jaumard and G. Savard, "A variable elimination algorithm for bilevellinear pro-
gramming", SIAM Journal on Scientific and Statistical Computing 13, 1992.
[17) P. Marcotte and D. Zhu, "Exact and inexact penalty methods for the generalized bilevel pro-
gramming problem", Publication CRT-920, Centre de recherche sur les transports, Universite de
Montreal, 1993.
[18) Z.-Q. Luo, J .-S. Pang, D. Ralph and S.-Q. Wu, "Exact penalization and stationarity conditions
of mathematical programs with equilibrium constraints", Technical report 275, Communications
Research Laboratory, McMaster University, Hamilton, Canada, 1993.
[19) P. Marcotte, "Network design problem with congestion effects: A case of bilevel programming",
Mathematical Programming 34, 1, 142-162,1986.
[20) P. Marcotte and G. Marquis, "Efficient implementation of heuristics for the continuous network
design problem", Annals o/Operations Research 34,163-176,1992.
[21) J. Gauvin and G. Savard, "The steepest descent direction for the nonlinear bilevel programming
problem" , forthcoming in Operations Research Letters.
[22) C. Fisk, "Some developments in equilibrium traffic assignment", Transportation Researcl. B 14B,
243-255, 1980.
[23] F. Leurent, "Cost versus time equilibrium over a network", European Journal 0/ Operational Re-
search 71, 205-221, 1993.
[24] S. C. Dafermos, "A multicriterion route-mode choice traffic equilibrium model", Unpublished
manuscript, Lefschetz Center for Dynamical Systems, Brown University, 1981.
[25] A. Auslender, and F. Brodeau, "Convergence d'un algorithme de Frank et Wolfe applique 11 un
probleme de controle", RIRO 2, 3-12, 1968.
[26] A. Auslender, "Optimisation: methodes numeriques", Masson, Paris, 1976.
VARIATIONAL AND QUASI-VARIATIONAL INEQUALITIES IN
NETWORK FLOW MODELS. RECENT DEVELOPMENTS IN
THEORY AND ALGORITHMS

Maugeri Antonino

Dipartimento di Matematica
Universita di Catania,
Viale A. Doria, 6
95125 Catania, Italy
e-mail: maugeri@dipmat.unict.it

Abstract: Variational and Quasi-Variational Inequalities have revealed themselves

as very useful tools to derive existence results and various kinds of algorithms for the
network equilibrium problems.
But, in contrary, the necessity to achieve more and more wider theory and efficient
algorithms has contributed to improve our knowledge about theoretical topics and
their mutual connections as Duality and Potential Theory, Gap functions, Separation
Theory, Sensitivity Analysis, Stability, etc.
A survey is presented on some recent developments in the above mentioned topics.

Key Words: Variational Inequality, Quasi-Variational Inequality, traffic network,

gap function, separation theory, paradoxes in equilibrium problems, efficiency of algo-
rithms.

1. INTRODUCTION

The theory of unilateral problems, which in the simplest case, deals with the
problem of finding a solution of the equation
~u = F in n, (Ll)
satisfying the boundary conditions

uau = 0 on an;
a" (1.2)
au
u;:: 0, a" ;:: 0 on an,
where ~ is the derivative of u along the inner normal direction; this problem was
solved in 1964 by Fichera [1], and this work marked the beginning of the theory, the
Variational Inequalities Theory, that received its best abstract setting for bilinear and
non-symmetric forms by Stampacchia ([2J, 1964), and by Lions and Stampacchia ([3],
1967). The same Stampacchia at the page 191 of the survey [4] writes:

195
196

"The first theorem of existence and uniqueness of the solution of Vari-

ational Inequalities wa.5 proved in [2). The first proof of Theorem 2.1 is
contained in [2). The recent proof of the same theorem is contained in [3J. "
Starting from 1964 there was an intensive development of the theory of Varia-
tional Inequalities and numerous applications of that theory have been done to the
free boundary problems (Stefan problem, plasticity, fluid dynamics, etc.).
It happened precisely in a hall of Majorana Center when G. Stampacchia expressed
his opinion that Variational Inequalities in finite dimensional spaces could have had an
important role also in the Mathematical Programming Theory and he wrote the first
work in this direction (see [5]), where he presented a direct method to compute the
solutions of a Variational Inequality.
The opinion of Stampacchia had a confirmation in 1979, when Smith [6J showed in
his work, dealing with existence, uniqueness and stability of traffic equilibrium problem,
that the asymmetric traffic equilibrium problem may be expressed by a Variational
Inequality. Since that year to the present several works have been devoted to this topic.
We may observe that not only the Variational Inequalities framework has been very
useful to achieve important results on traffic equilibrium problem, but viceversa many
theoretical topics (new and important) have been developed as result of attemptions
to solve more general and deep problems.
The aim of the present survey is to give some recent development, trying to make
clear the connections, among different points of view.

2. MODELS

In this section I will present the usual model of traffic network, that in a natural
way leads to Quasi-Variational Inequalities. Subsequently, I will report the general
approach suggested by Giannessi ([7J, [15J, and this Volume). As usually, a set N of
nodes is given, say Nb ... , N p , a set A of n directed arcs, say Ab ... ,An and a set Wof
1 ordered points: Origin/Destination, say WI' ... ' Wi. Hence the network is identified
by the 3-triple (N, A, W).
The flow on the arc Ai is denoted by fi' f = (fl, ... , fn). We assume that the
nodes ofthe origin-destination (in short OlD) pair Wj are connected by rj ~ 1 paths,
whose set is denoted by Pj. PI! ... , p" with m = r1 + ... + rl, are all the considered
paths, and F1 , ••• , Fm are the corresponding flows; F:= (FI! ... , Fm).
We assume that the travel cost on arc Ai depends on t, and denote it by Ci(f),
c(f) = (C1(f), ... ,cn(f)). C.(F) will denote the travel cost on the path R. as a function
of the flows of all paths; C(F) = (Cl(f), ... ,Cm(F)). By means of Kronecker numbers:
I, if Ai E R., i = 1, ... , n,
6i • ={
0, if Ai ft R., s = 1, ... ,m,
the flows on arcs can be expressed in terms of flows on paths:
m

f; = L: 6i .F. , i = 1, ... ,n, (2.1)

.=1
or setting ~ = {6i .}
f=~F.

Also the cost C.(F) on the path R. can be expressed as a sum of all costs on the arcs
of R.:
n
C.(F) = L: 6i.Ci(~F), (2.2)
i=l
 197

or
C(F) = ~TC(~F).
According to the classic Wardrop's definition, a vector H E R+ is said to be an
equilibrium pattern flow iff 'IRq, R. E Pi we have

C.(H) > Cq(H) ::::::} H. = 0, (2.3)

and this for each Wi'

If H is an equilibrium pattern flow, if Rq, R. E Pi and Hq > 0, then the cost
functions satisfy the relations:

= Cq(H), if H. > 0,
C.(H) { .
~ Cq(H), 1f H. = 0.
Then, when the equilibrium pattern flow is established, an equilibrium path cost Ci(H)
is associated to each O/D pair Wi:

Ci(H) = R.EPi
min C.(H). (2.4)

Now, if we denote by Pi the travel demands for the OlD pair Wj, in general each
demand depends on the costs for all O/D pairs:

and, then, they depend on equilibrium pattern flow H.

Dafermos (8) proposed to introduce new 1 variables

c)j, j = 1, ... ,1; (2.5)

and to assume that the function p = p( c)

p:R~ ---+ R~ (2.6)

be continuous and strongly monotone. Then the inverse function c) c)(p) exists;
moreover, taking into account the equilibrium conditions:

c) { = Cr(H), if ,Fr > 0, Rr E Pj,

) :::; Cr(H), if ,Fr = 0, Rr E Pi,
she showed the variational formulation: (H, pO) is an equilibrium pattern iff

H E K(p*): C(H)(F - H) ~ c)(p*)(p - pO), (2.7)

'IF E K(p*), Vp E R~ ,
where
K(p)={FER+: 'PF=p} PER~. (2.8)
In (9) the authors, taking into account that, as a matter offact, the demands depend
on H, proposed to embed the problem into a Quasi-Variational Scheme.
To this end, let me introduce the Kronecker numbers

1, if R. E Pi
'Pi. ={ and 'P = {'Pi.}, j = 1, ... ,1, s = 1, ... ,m
0, if R. ¢. Pi
198

by means of which the so-called flow conservation law can be written as:

cpF = p(H).
In (9] (see also (10], (11]) it is shown that HE K(H) = {F E R~: cpF = p(H)} is
an equilibrium pattern iff

(C(H), F - H) ~ 0, VF E K(H). (2.9)

The paper by De Luca of this Volume provides more information about solvability
of problem (2.9), but it is worth to report the formulation suggested by Giannessi
(see [7]), who remarks that the capacity of arcs must be taken into consideration; as

f;
concerns this question see the paper by Ferrari of this Volume. If we suppose that
every arc flow must satisfy the upper bounds

i = 1, ... ,n,
we have from (2.1)
m
L hirFr ::; 'Yi, i = 1, ... ,n
r=1
or
t:..F::; 'Y
and, therefore, the set of feasible path flows becomes

K(H) = {F E R~: cpF = p(H), t:..F::; 'Y}. (2.10)

Hence the Problem (2.8) is solvable if compatibility conditions on K(H) are fulfilled.

3. GAP FUNCTIONS AND EQUIVALENT DIFFERENTIABLE

OPTIMIZATION PROBLEMS. AN EMPIRICAL APPROACH

Let us consider the Variational Inequality

x* E S: (f(x*), x - x*) ~ 0, \:Ix E S, (3.1)

(we use the usual notations) where S is a closed, convex and non-empty subset of R,n
and
f: S (or Rn) --+ R n
is a vector valued function with properties to be specified in what follows.
It is well known that, if f is differentiable and the Jacobian matrix

of (Of;)
ax = ax· ',J=l,
.. ...,n (3.2)
J

is symmetric, then the problem (3.1) can be considered as the first differentiable con-
dition for an optimization problem. Then, well-known theoretical and computational
results can be applied.
If (3.2) is not symmetric, i.e. we are out of the usual optimization framework,
it is possible to associate to the Variational Inequality (3.1) a non-negative function,
called gap function, that is zero iff the argument is a solution of (3.1) and such that,
if differentiable, its first optimality condition is verified by a solution of (3.1).
 199

Auslender [12] was the first who introduced the function

defined by
g(x) := max(f(x), x - y). (3.3)
liES

He showed that
1) g(x) ~ 0,
2) x* is a solution of the Variational Inequality (3.1) iff g(x*) = minxEs g(x) = O.
But g(x), in general, is not differentiable, and this has led some authors to construct
more refined gap functions.
Fukushima in [13] presents a new kind of gap function whose stationary point solves
the Variational Inequality (3.1). He considers a n X n symmetric matrix G and, for
any xES, the function
1
F(x) = -(f(x),H(x) - x) - 2(H(x) - x,G(H(x) - x») , (3.4)

with
H(x) := Pr S,G(x - G- 1 f(x»,
where PrjsG(x)
,
denotes the (unique) solution of problem min
liES
lIy - xliG' It is shown
that
F(x) ~ 0,
F(x*) = 0 iff x* is a solution of the Variational Inequality (3.1).
The result is achieved taking into account the fact that a solution of (3.1) verifies
the condition
x* = Pr SG(x* - G- 1 f(x*)) = H(x*), (3.5)
and that F(x) can be rewritten as

F(x) = ~{lIx - (x - G-1f(x)lIb -IIH(x) - (x - G-1f(x)lIb}.

Then F(x) = 0 iff H(x) = x. The result is important since the following properties
hold.

Theorem 3.1. If the mapping f: R n -+ R n is continuous, then the function F(x)

is also continuous. If f is continuously differentiable, then F is also continuously
differentiable, and its gradient is given by

of = f(x) _ [af(x) _ G]. (3.6)

ax ax

Moreover, if ~ is positive definite for all xES, and x* is a stationary point of the
problem
minF(x) = F(x*), (3.7)
:rES

that is,
( a~~*) ,x - x* ) ~ 0, Vx E S, (3.8)

then x* solves the Variational Inequality (3.1).

200

The proof is achieved by the fact that, setting

1
hex, y) = (J(x), y - x) + '2(y - x, G(y - x)), (3.9)

h: R n xS ---+ R,

it results

F(x) = min-h(x,y) = -(J(x),H(x) - x) -

yES
~(H(x)
2
- x,G(H(x) - x)).

The gap function provides us also an iterative method to compute the solution of
(3.1). In fact, Fukushima showed the following:

Theorem 3.2. Set {xk} to be a sequence generated by iteration

k = 0,1, ...

where dk are given by dk = H(x k ) - xk, and tk E [0,1] are determined by Armijo-type
rule
F(x k + fidk)::; F(x k ) - a,Bllldk Il 2 •
If S is compact, then the mapping f( x) is differentiable and strongly monotone with a
modulus /-I > 0 on S, and f(x), 8~~) are Lipschitz continuous on S, then the generated
sequence {xk} lies in S and it converges to the unique solution of the Variational
Inequality (3.1) for any starting point XO E S, if the positive constant a of the Armijo-
type rule is chosen sufficiently small so that a < /-I.

It is worth to observe that the generated sequence

(3.10)

is very similar to that generated by the projection method:

(3.11)

We will give a valuation of the convergence rate of projection method in Section 5

and one can get an idea of efficiency of the methods.
The formulation of a differentiable gap function can be extended by replacing the
function
1
hex, y) = (J(x), y - x) + '2(y - x, G(y - x))

with a more general one (see [14]):

h(x,y) = (F(x),x - y) - Q(y,x),

where Q(y,x): S X S -+ R is such that

1) Q(y,x) ~ 0,
2) Q(x,x) = 0, Vx E S,
3) Q is continuously differentiable on S x S and

VQ(x,x) =0, VxES,

4) Q is strongly convex on S with respect to y for all xES.

 201

Considering the gap function

g(x) max hex, y)

yES
=- min {U(x), y - x)
yES
+ !ley, x)} (3.12)

-min{U(x),H(x) - x)
yES
+ !l(H(x),x)},
it is clear that H (x) is well-defined since h is a sum of a linear and a strongly convex
function.
The authors showed the following result:

Theorem 3.3. g( x) is a gap function. If f( x) is continuously differentiable, then 9

is also continuously differentiable, and it results:

og(x) = f(x) _ fOf(x) ,H(x) _ x) _ o!l(H(x),x).

ax \ ax ax
Moreover, by introducing an auxiliary Variational Inequality

(r(w, y) - r(y, y) + F(y), x - w) ~ 0, "Ix E S,

where y is a fixed point in S and r: S x S --+ R n is a continuous mapping, strictly

monotone with respect to w, and continuous with respect to y which has the property
that, if
w = w(y) = y,
then y is a solution of the Variational Inequality (3.1), it is possible to derive various
kinds of gap functions.
The generalization leads to the most general formulation for the gap function. This
will be the topic of the next section.

4. GAP FUNCTIONS. AN IMAGE SPACE FORMULATION

Giannessi proposed (see [7], [5]) to associate to a Quasi-Variational Inequality

an image space and, then, to derive very general gap functions by means of theorems
of the alternative or of separation theorems. Moreover, it seems provide a promising
framework to reach very general connections between gap functions and duality. To
show this technique, let us consider a Quasi-Variational Inequality in the form:

YE K(y): (F(y),x - y) ~ 0, "Ix E K(y),

where
K(y):= {x E X(y): g(y;x) E C} ( 4.1)
and where C is a closed, convex cone with apex at the origin, X(y) is a subset of a
Hilbert space 3, and
g: X(y) x X(x) __ Rm.
This form includes (3.1) and the approach proposed by Giannessi starts with the remark
that y is a solution of (4.1) iff the system (in the unknown x)

u:= (F(y), y - x) > 0

{ (4.2)
v:=g(y;x)EC, x EX(y)
202

is not solvable. The space where (u, v) runs is the image space associated to (4.1) and
the set

K(y) := {(u,v) E R X Rm : u = (F(y),y - x), v = g(y;x),x E X(y)}

is the image of (3.1). Setting 11.:= {(u,v) E R XRffl: u > 0, v E C}, the unsolvability
of system (4.2) is equivalent to 11. n K(y) = 0. Then, considering the particular case

0p = (0, ... ,0) E RP, °: ; p ::; m,

let us set

w = u + (>., 'Y(y, v, w»), u E R, v E R ffl , >. E C·, wEn,

where

is the polar of C, and

= II n.,
ffl

Wi E n, n
.=1
and 'Y, n are such that Vy E X(y)
(4.3)

(weak separation).
Then the function

'l/;w(y):= min max [(F(y),y - x)

.\eO· xeX(y)
+ (>','Y(y,g(y,x),w)j
will be shown to be a gap function and will be called weak gap function. For the proof
see [6] and this Volume.
On the other hand, we may consider the function

s = u - Ii(y, v,w), uER, vERffl , wEn,

where 0: ::: x Rffl xn -+ R is, for each y, such that

(strong separation)

and Ii and n must be such that

3w E n s.t. K(y) n lev=os :f:. 0 (4.4)

11.nK(y):f:.0=>3wEn and (u,v)E11.nK(y): s>O. (4.5)

Then the function

'I/;.(y) = max max [(F(y),y -

wen xex(y)
x) - o(y,g(y,x),w)j

will be proved to be a gap function and will be called strong gap function.
 203

It is worth to remark that choosing

WV, if v ~ 0
~(y,v,w)= {
+00, if v < 0
and
g(y,x) = (x - y, G(x - y»)
(G is positive definite square matrix of order n), or

g(y,x) = Q(x,y)

(the same of Zhu-Marcotte), we recover the gap function considered by Fukushima

and Zhu-Marcotte.
The approach given above makes us confident that:
1) More general conditions for the solvability of the Quasi-Variational Inequality
(3.1) can be found.
2) More efficient algorithms can be constructed.
3) The dual Variational and Quasi-Variational Inequalities can be defined under
non-restrictive conditions.

5. ON THE CONVERGENCE RATE OF PROJECTION METHODS

Dafermos [17] presented a method to compute the solution of a Variational In-

equality. This method has a significant analytical value and reporting the valuations
of its rate of convergence, due to Zanni (see [18]), we want to elevate a warm homage
to her memory.
Let us consider the Variational Inequality:

HE K: r(H)(P - H) ~ 0, VP E K,

where K is a closed, convex and non-empty subset of R n and r is a mapping

Dafermos supposed that the function r be continuously differentiable on K and

strongly monotone, i.e.

vPI, p 2 E K, (5.1 )

a being a positive constant. We remember that an equivalent condition for (5.1) holds
is that the Jacobian matrix ;~ is positive definite VP E K.
The construction of a sequence {Pi} c K that converges to the solution of the
Variational Inequality is developed in the following way:
- firstly, one fixes an arbitrary po E K, a symmetric positive definite matrix G, a
positive parameter 0 < p < ~, where

v = max
K {maxAi' l<i<m:
- - Ai is an eigenvalue of (u~pr )T G- 1(~pr)
u }., (5.2)

- at the i-th step, one calculates the vector

(5.3)
204

and solves the Variational Inequality

(5.4)

where
(5.5)
r
We observe that, being the Jacobian of i - 1 a positive definite symmetric matrix,
the problem (5.4) is equivalent to the problem of finding the minimum, over K, of the
quadratic, strictly convex function

(5.6)

Thus, one obtains a sequence {Fi} C K and it is possible to prove that this sequence
converges to the solution H.
Further, we have the following estimate

IIF" - Hila :5 ~ III - pOlio, (5.7)

where IIFlia = (pGF)1/2 and

A= (1- p(2a; IIp)f/2, (5.8)

p being the greatest eigenvalue of G.

Choosing the maximum of the function

F(p) = p(2a - lip)

P
on (0, 2:), one obtains p = ;: and, from (5.8):

(2)1/2
A= ( 1 - - , (5.9)
pll

that represents, by (5.7), the best value for the rate of convergence.
The following theorem is proved by Zanni [18]:

Theorem 5.1 If J is a non-singular m x m matrix, G a symmetric positive definite

matrix, then one has
II > 1IJ1I2
- P ,
where II and p are defined above, and II . II is the usual matrix norm.

Applying this theorem to (5.9) one has, in the affine case r(F) = J F + b:
2 ) 1/2
A ~ ( 1- 11 ;11 2 (5.10)

Now, denoting by f3 the greatest eigenvalue of the matrix J.

observing that it results
 205

from (5.10) it follows:

2) 1/2
oX> ( 1-~ (5.11)
- ~2

For the nonlinear case, denoting by a the minimum eigenvalue of a~<p, and by Ii
the maximum eigenvalue of the symmetric part [a~)f)l. of a~l{l, it results
a
YFEK: oX> 1-~.
- ~

.) 1/2
The graph of ( 1- ~ on ~/o. shows that, independently of G, the constant oX
takes values close to 1 also for values of ~ / 0. very close to 1.
Now, using the above results, we show a comparison between the projection method,
as developed by Dafermos, and the direct method.
We consider as an example the model of a circular highway exhibited by Bertsekas
and Gafni (see Fig. 1, and [19)).
In this case, the convex K is
I
F;+F;+s = 10' i = 1, ... ,5}
r is strongly monotone, and the components of r are quadratic functions of F:
r 1 (F) = 34Fl2 + 42FlF2 + 20Fl F3 + 20FI F4 + 42FlFs + 34Fl + 21F; + 20F2F3
+20F2Fs + 21F2 + 10F; + 10F3 + lOF; + 20F.Fs + lOF. + 21Fl + 21Fs + 34
r2(F) = 21Fl2 + 42FlF2 + 20Fl F3 + 20F1 Fs + 21Fl + 34F; + 42F2F3 + 20F2F4
+20F2Fs + 34F2 + 21F; + 20F3F4 + 21F3 + lOF; + lOF. + 10Fl + 10Fs + 34
r3(F) = lOFl2 + 20Fl F2 + 20Fl F3 + lOFl + 21F; + 42F2F3 + 20F2F. + 21F2 + 34F;
+42F3F. + 20F3FS + 34F3 + 21F; + 20F.Fs + 21F. + lOFl + 10Fs + 34
r.(F) = lOF; + 20Fl F. + 20FlFs + lOFl + lOF; + 20F2F3 + 20F2F. + lOF2 + 21F;
+20F3FS + 21F3 + 42F3F4 + 34F; + 42F4FS + 34F. + 21Fl + 21Fs + 34
rs(F) = 21F; + 20Fl F2 + 20F1 F. + 42FlFs + 21Fl + lOF; + 20F2Fs + lOF2 + lOF;
+20F3F. + 20F3FS + lOF3 + 21F; + 42F4FS + 21F. + 34Fl + 34Fs + 34
r6(F) = 23F; + 10F; + lOF;o + 20F6 F7 + 20F6FlO + 23F6 + 10FT + lOFlO + 23
r7(F) = 23F; + 10F; + lOF; + 20F Fs + 20F6Fr + 23F
T 7 + lOFs + lOF6 + 23
rs(F) = 23F; + 10F; + lOF; + 20FgFs + 20FsFr + 23Fs + lOFg + lOF7 + 23
rg(F) = 23F; + lOFl2o + lOF; + 20FgFlO + 20FsFg + 23Fg + lOFlO + lOFs + 23
rlO(F) = 23F;o + 10F; + lOF; + 20F6F lO + 20FlOFg + 23FlO + lOF6 + lOFs + 23.
Applying the direct method, proposed by [20] one finds, at the third step, the
unique solution

H = (0,0,0.018807,0.135947,0.114582,0.1,0.2, 0.281193, 0.264053, 0.385418).

It is remarkable to observe that the normalized measure of convergence
206

is equal to zero for the solution H.

By returning to the projection method, it is not difficult in this case, thanks to the
particular form of r, to obtain an estimate of .A by theorem 5.l.
One finds an approximation of Ot / (3 calculating ~~ in the point
P = (0,0,0,0,0,0.1,0.2,0.3,0.4,0.5):

(%f ~ 0.003

and then, for (5.11),

.A :2: 0.997.
We recall that, in order to calculate the i-th step, it is necessary to consider the
vector (5.3)
H i - 1 = pr(Fi-l) - GFi-t,
and hence it is necessary to fix p E (0, ~).
Choosing p = 0.2, G = I, the algorithm does not converge. Then the author
assigns to p values inferior than 0.2 and tries to find those values which yield the fastest
convergence. Initially, he decreases the parameter p of 0.005 at each application. The
first value which yields convergence is p = 0.02. From this value he decreases p of
0.001 at each application. The best result is obtained at p = 0.018 . In this case, after
38 steps, the following approximate solution is obtained:

fI = (0,0,0.0188074,0.1359459,0.1145822,0.1,0.2, 0.2811925, 0.2640540, 0.3854177).

The graph on Fig. 1 shows how the number of iterations depends on p.
300,--r--------------------------------~

250

200

150

100

50

0.005 0.01 0.015 0.02 0.025

Figure 1

The conclusion is that, except for some particular cases, the constant .A does not
assume those values that guarantee a rapid convergence of the method.

6. SENSITIVITY ANALYSIS. PARADOXES

We conclude this presentation showing some problems where the analysis of sensi-
tivity of the equilibrium solution, with respect to changes in input data, leads to results
that seem counter-intuitive, and hence have been exhibited as paradoxes (see [21], [22]).
 207

Even if several works are devoted to the sensitivity analysis, however these counter-
intuitive situations go out of the known analysis.
We present a generalization of Fisk's paradox [21] and, by means of the direct
method due to [20], we show that these situations may often occur.

p3.L-____________----------~P2

Figure 2

Let us consider the network in Fig. 2 in which

N {PI, P2,P3},
A W = {WI = (P3,P2), W2 = (P3,Pl ), W3 = (P2,Pl )},
R {RI,R 2,R3,R4}

with

Rl (P3, P2)
R2 (P3,Pd
R3 (P2 , PI)
R4 (P3,P2) U (P2,Pd.

Let us suppose that the costs are:

Cl(f) 11
C2(f) h +a
C3(f) h
where a is a non-negative constant. Since

it F l +F4
12 F2
h = F3+ F4
the path costs are:

Cl(F) = Fl +F4
C2(F) = F2 +a
C3 (F) F3+ F4
C4 (F) Fl +F3 + 2F4 •
208

Then we must solve the variational inequality:

4
C(H) (F - H) = E Ci(H) (Fi - Hi) ~ 0, (6.1)
i=l

where
K = {F E R!: FI = Ph F2 + F4 = P2, F3 = P3},
Pi being constant positive travel demands.
Applying the method of [201, we find that the solution of the system:

CI(H) C2 (H)
HI = PI
H3 = P3
H2+H4 = p2
i.e.,
H_ (
- Ph
2P2 + PI + P3 -
3
a,P3, P2 - PI 3- P3 + a)
is the solution of (6.1) if:

Now, let us increase the travel demands related to the pair WI, from PI to PI + dO,
(d" > 0). The new solution is the vector H" with
H"I = HI +d"
d"
H; H2+-
3
HO3 = H3
d"
H"4 = H4 - -
3'
provided that
a - 2P2 ~ PI + P3 + d" ~ a + P2.
Then, if we calculate the total costs CO(HO)H" and C(H)H, we find:

COCH")HO = CCH)H + ~d"C2PI + P2 - P3 + i + d").

Now, if it results
(6.2)
we get
C"CH")HO < CCH)H
although PI + d" > Pl.
Hence, we obtain by (6.2) a set of parameters in which paradoxical results can be
occur, namely, to an increase of the travel demands corresponds a decrease of the total
travel cost.
As a particular case, we find the example of Fisk, if PI = 1, P2 = 20, P3 = 100,
d" = 3, a = 90 (note that for this values C6.2) is satisfied).
An other interesting situation was presented by Braess (see [22]), and here we will
present a generalization of its paradox.
 209

~~----------------~~

"'-__________---------' Pl

Figure 3

Let us consider the network in Fig. 3 where

N = {Pt ,P2 ,Pa,P,,},

A {(P}, P2 ), (Pt , Pa), (P2 , Pa), (P2, P4 ), (Pa, P4 )},
W {w = (P},P4 )}.

The paths for w are

Rl = (Pt,Pa)U(Pa,P,,)
R2 (PI, P2 ) U (P2 , P4 )
Ra (Pt , P2 ) U (P2 , Pa) U (Pa , P4 ).

Let us suppose that the link costs are:

CtC!) = 10ft
C2C!) c:t+/2
caC!) {3+ /3, 0<c:t<{3
C4C!) = c:t+h
csC!) 15
with the path flows F; given by:

ft = F2+Fa
h Ft
/3 F3
14 F2
15 Ft + Fa.
Then, the path costs are

Ct = 11Ft + 10Fa + c:t

C2 llF2 + lOFa + c:t
Ca 10Ft + 10F2 + 21Fa + {3.
 210

The variational inequality is the following:

3
C(H) (F - H) = :E C;(H) (F; - H;) ~ 0, VFeK (6.3)
;=1

with
K = {F e R 3+: Fl + F2 + F3 = p},
p being the travel demand for w.
Once again, using the above method, we obtain that the solution is given by the
system:

C1 (H) C2(H)
C2 (H) - C3(H)
HI +H2
~
°
P (6.4)
H3
HI
=
~
° H2 0, ~ 0.

Since it results
P
HI = H2 = 2' H3 = 0,
9
C2(H) - C3 (H) = -"2 P + a - {J < 0,

the vector H = (i, i,O) is the solution of (6.3) and the total cost is:

Now, we improve the total cost C3, considering a positive parameter {J* < a, instead
of {J, i.e.:

C;*(F) = Ci(F), i = 1,2

C;(F) C3(F) + {J* - {J.
We obtain, for the new equilibrium distribution, the vector:

H* = 113 (llp - (a - {J*), IIp - (a - {J*), 2(a - {J*) - 9p)

provided that
(6.5)

The new total cost is

C*(H*)H* = I~P(31P + 22a - {J0)

and it results C(H)H < C*(H*)H* iff

9 a* 31
"2 P +1-' < IS a . (6.6)

Thus, we find a paradoxical result for the class of problems where the parameters a,
{J, {J* and p are such that (6.5) and (6.6) are satisfied.
The case considered by Braess with a = 50, {J = 100, {J0 = 10, p = 6, is a particular
case of this class.
 211

REFERENCES
[1) G. Fichera, "Problemi elastostatici con vincoli unilaterali: il problema di Signorini con ambigue
condizioni al contorno," Memorie dell'Academia Nazionale dei Lincei , pp. 91-140, 1964.
[2) G. Stampacchia, "Formes bilineares coercives sur les ensembles convexes," Compt. Rend. Acad.
Sci. Paris, pp. 4413-4416, 1964.
[3) J.-L. Lions and G. Stampacchia, "Variational inequalities," Comm. Pure Appl. Math. 22, pp. 493-
519, 1967.
[4) G. Stampacchia, "Variational inequalities," Proceedings of NATO Advanced Study Institute, Venice,
June 17-30, pp. 101-191,1968.
[5) O. G. Mancino and G. Stampacchia, "Convex programming and variational inequalities," J. Optim.
Theory Appl. 9, pp. 3-23, 1972.
[6) M. J. Smith, "The existence, uniqueness and stability of traffic equilibrium," Transportation Res.
138,pp. 295-304, 1979.
[7) F. Giannessi, "Theorems of the alternative, quadratic programs and complementarity problems" .
In "Variational inequalities and complementarity problems", R.W. Cottle et al. eds., J. Wiley, New
York, pp. 151-186, 1980.
[8) S. Dafermos, "The general multi modal network equilibrium problem with elastic demand," Network
12,pp. 57-72, 1982.
[9) M. De Luca and A. Maugeri, "Quasi-variational inequalities and application to equilibrium prob-
lems with elastic demand," in: F. H. Clarke et al. eds., "Nonsmooth optimization and related
topics", Plenum, New York, pp. 61-77, 1989.
[10) M. De Luca and A. Maugeri, "Discontinuous quasi-variational inequalities and application to equi-
librium problems," in: F. Giannessi ed., "Nonsmooth Optimization. Methods and Applications",
Gordon and Breach, U.K., pp. 70-75, 1992.
[11) M. De Luca and A. Maugeri, "Quasi-variational inequalities and application to the traffic equilib-
rium problem; discussion of a paradox," J. Compo and Appl. Math., 28, pp. 163-171,1989.
[12) A. Auslender, "Optimization: Methodes Numeriques", Masson, Paris, 1976.
[13] M. Fukushima, "Equivalent differentiable optimization problems and descent methods for asym-
metric variational inequality problems," Math. Progr. 53, pp. 99-110, 1992.
[14] D. L. Zhu and P. Marcotte, "A general descent framework for monotone variational inequalities,"
J. Optim. Theory Appl., 1993.
[15] F. Giannessi, "On some connections between combinational and continuous optimization," In "An-
nals of Operations Research", North-Holland. To appear.
[16] F. Giannessi, "Theorems of the alternative and optimality conditions." Jou. Optim. Th. Appl., Vol.
42, N. 3, pp. 331-365, 1984.
[17] S. Dafermos, "Traffic equilibrium and variational inequalities," Transport. Sci. 14, pp. 42-54,1980.
[18] L. Zanni, "On the convergence rate of two projection methods for variational inequalities in R n ,"
Calcolo, 1993.
[19] D. P. Bertsekas and E. M. Gafni, "Projection methods for variational inequalities with application
to the traffic assignment problem," Math. Progr. Study 17, pp. 139-159, 1982.
[20] A. Maugeri, "Convex programming, variational inequalities and applications to the traffic equilib-
rium problem," Appl. Math. Optim. 16, pp. 169-185, 1987.
[21] C. Fisk, "More paradoxes in the equilibrium assignment problem," Transport. Res. 13 B, pp. 305-
309, 1979.
[22] D. Braess, "Uber ein Paradoxon der Verkehrsplanung," Unternehmensforschung 12, pp. 256-268,
1968.
A VARIATIONAL INEQUALITY ASSOCIATED TO A SCALAR
CONSERVATION LAW WITH UNILATERAL CONSTRAINTS

Murthy Venkatesha M.K.

Department of Mathematics
University of Pisa
Via F. Buonarroti 2, 56127 Pisa, Italy
e-mail: murthy@gauss.dm.unipi.it

Abstract. We consider an inequation of hyperbolic type associated to a first order

quasi-linear equation, generalizing the equation describing a scalar conservation law,
of the form

BtU + div(f(x, t, u)) + g(x, t, u) = 0 f = (ft,···, fn), fi = fi(x, t, u)

in a cylindrical domain QT = n x (-T, T), on a bounded open subset n in lR n with

initial and boundary conditions, and a unilateral obstacle condition. It is shown that
there exists a unique weak - entropy solution belonging to LOO( QT) n EV( QT) for such
a problem. The proof is based on an approximation by solutions of a sequence of vari-
ational inequalities of parabolic type and such an approximation procedure generalizes
the well known method used in the study of solutions of conservation laws by viscosity
solutions as the viscosity parameter tends to zero and these in turn are obtained as
limits of solutions of penalised problems associated to the corresponding quasi-linear
parabolic equation.

Key Words: Quasilinear hyperbolic equation, weak-entropy solution, functions of

bounded variation, unilateral constraint, vanishing viscosity method

1. INTRODUCTION

We present a brief survey of some results on weak--entropy solutions of first order

quasilinear equations of hyperbolic type of the form

E(u) = BtU + divf(x,t,u) + g(x,t,u) = 0

where f(x,t,u) = (ft(x,t,u),···,fn(x,t,u)). A basic model for the study of such
equations is the classical equation describing a scalar conservation law. Our present a-

213
214
tion closely follows the papers of Kruzkov [6] and Bardos-Leroux-Nedelec [1], a recent
note of Levi [9] and some results obtained independently by the author.
Starting from the fundamental work of E. Hopf [1] the central problem of the
weak or generalized solutions of conservation laws consists in describing the classes
of functions in which there is existence and eventually uniqueness for the solution of
the associated Cauchy problem or the mixed Cauchy-Dirichlet problem. The unique
solution, if it exists, should correspond to the physically relevent solution of the con-
servation law considered.
Ever since the fundamental paper [4] of Hopf the basic method used to study the
scalar conservation law equation

with bounded measurable initial values u(x,O) = uo(x) has been the so called van-
ishing viscosity method which consists in studying the family of solutions Up of the
corresponding parabolic equation

OtU + oxf(u) = p.o"""u

with initial data Uo and then a passage to the limits as the viscosity parameter J.l tends
to zero. Hopf considered the equation

OtU + UO",U = p.o"""u

and the method introduced by him allows not only to prove the existence of general-
ized solutions but also to determine additional conditions on the class of generalized
solutions which characterize the uniqueness.
We shall consider quasilinear scalar hyperbolic equations of first order on a domain
in the n-dimensional Euclidean space IR= of the form

E(u) = OtU + divf(x, t,u) + g(x, t, u) = 0

where f(x, t, u) = (h(x, t, u), ... .In(x, t, u». We shall formulate the exact regularity
assumptions on Ii and g required in each context seperately. We have the following
natural definition of a weak or generalized solution: A function u E Loo is said to be
°
a weak solution of the equation E( u) = if it satisfies the integral identity

ff [UOt<P + L Ii(x, t, u)O"'i <P - g(x, t, u)<pl dx dt =0

for all finite, smooth functions <p.
Some other examples of quasilinear equations of the above form are:
 215

Remark 1. The function g(x, t, u) - E(oxj Ii )(x, t, u) is also referred to as the source
term.
In the one dimensional case, for feu) a convex function of u and for initial values
uo E Loo the problem is most studied. For a non convex feu), again in the one
dimensional case the problem was studied by Gelfand, Oleinik and others who also
determined a condition for the uniqueness among piecewise smooth solutions of the
Cauchy problem.

1.1 Functions of bounded variation and weak solutions

Since we shall be interested in weak solutions which admit traces for t = and °
also on the boundaries of domains a natural space of functions (generalizing the class
of piecewise smooth functions) is the class offunctions of bounded variations. We shall
briefly recall below the basic facts on functions of bounded variation.
Let !1 be an open set in JRn with a piecewise smooth boundary and T > 0. We shall
denote by Qr the cylinder Qr = !1x(O, T) and by Er = o!1x(O, T) its lateral boundary.
Similarly Zr = JRn x (0, T) denotes the infinite cylinder. (x, t) = (Xl'···' x n , t) denotes
the standard coordinate system.

1.2 The space BV of functions of bounded variation

We denote by BV(!1) the space of functions of bounded variation on n. That is,

BV(n) is the space of functions v E Lloc(!1) such that OXj v for j = 1,··· n are Radon
measures on!1.
Similar definitions hold for the spaces of functions of bounded variation on Qr or
Zr, denoted respectively by BV(Qr) and BV(Zr).

Lemma 1. If v E BV(Qr) then

(i) the trace IOV = Vlt=o is defined and belongs to LOO(n).
(ii) the trace IETV = ViET E LOO(Er) and can be approached by convergence in Ll.
Further,
(iii) if hE C 1 (JR) then IETh(v) = h(IEV) almost everywhere on Er
In fact, in (i) vex, t) -t IOV in LI(n).
For (ii), choosing the canonical basis {e1' ... ,en} we can define the trace ijV at each
point (x, t) E Er by passing to the limits along the direction ej not tangent to the
boundary at (x, t). Since OXj v are Radon measures we have, from the Green's theorem
that, for 1> = (1)1,···, 1>n) with 1>j E CI(n),

This allows us to approach IEr in LI norm.

Finally (iii) follows from this convergence in L 1 .
We need in the following the sign function sgn defined by
216

1, forr>O
sgn(r) = { 0, forr=O
-1, forr<O
and this can be approximated by the continuous function sgn>. defined, for A > 0, by

1, forr > A
sgn>.(T) = { 0, for Irl = A
-1, for r < A
In order to use suitable test functions starting from the weak solution of the equation
we introduce also the primitive of the function sgn>.:

riel
I>.(e) = 10 sgn>.(r') dr' for eE JRn
Then it is easily seen that
e--+ I>.(e) is a convex function on JRn and that

It is some times convenient to use the fact that a function w belongs to BV(JRn )
if, for every relatively compact open set wand for all h E JRn, we have

L Iw(x + h) - w(x)ldx ~ const·lhl

If v(x,t) E BV(JR n +1 ) then at every point of discontinuity (except at most for a
set of n-dimensional Hausdorff measure zero) the discontinuity is of the first kind and
the set of discontinuities admits a normal at almost all its points.

2. THE CAUCHY PROBLEM FOR FIRST ORDER QUASI LINEAR

EQUATIONS ON THE WHOLE SPACE

Consider the Cauchy problem: Given Uo E LOO(JRn ) to find a generalized solution

{ °
u E BV(JRnH )
E(u) = in JRnH
u(x,O) = uo(x)
We have the following global existence result:

Theorem 1 (Conway and Smoller) [2]. If Uo E BV(JRn ) then there exists a weak
solution of the Cauchy problem

{ u E BV(JRnH )
Otu+divf(u)=O 10 JRn+l
u(x,O) = uo(x)
such that, for all t = const, x 1-+ u(x, t) belongs to BV(JRn ).
 217

The proof is based on the method of vanishing viscosity and the method is justified
only when the initial data uo is sufficiently smooth.
An existence result without uniqueness for the general quasilinear equation E( u) =
o was proved by Kuznecov in 1967.
In order to obtain existence and uniqueness in the Cauchy problem Kruzkov in-
troduced a some what different notion of generalized solutions in the class of functions
Loo in [6].
Assume
f : ZT X IR --+ IRR and g: ZT X IR --+ IR

{ f(x, t, u) = (... , fj(x, t, u),···) with hE Cl(ZT

9 E C1(ZT X IR)
Uo E Loo(IRR) with lIuoll = Mo
X IR)

Definition 1. (Kruzkov). A function u(x, t) E Loo(ZT) is a generalized solution of

the Cauchy problem in ZT if
(1) for every constant k E IR and for all smooth functions ¢>(x, t) ~ 0 in ZT with
supp.¢> c (ZT)O we have

JlZT[ [Iu(x, t) - klat ¢> + sgn( u(x, t) - k) l)h(x, t, u(x, t)) - hex, t, knox; ¢>-

-sgn( u(x, t) -kHL: ax; hex, t, k) + g(x, t, u(x, t))}¢>]dx dt ~ 0 (*)

(2) there exists a measurable subset E of(o,T] with meas (E) = 0 such that "It E [0, T]\E
the function x f-t u(x, t) is defined almost everywhere in IRA and Vr > 0 we have

lim [ lu(x, t) - uo(x)1 dx = 0

VtE[O,T]\E, t~O llxl~r

Remark 2. For k = ±suplu(x, t)1 the inequality (*) implies the original weak formu-
lation.

Remark 3. This definition also characterizes the set of permissible discontinuities of

the generalized solution, In fact, if the generalized solution u is piecewise smooth then
we find on integration by parts using Gauss theorem that, in some neighbourhood of
a point of discontinuity (*) implies the following:
Denotimg by v(x,t) the oriented normal to the surface of discontinuity of u at (x,t)
and by u± the one sided limit of u as (x',t') tends to (x,t), namely,

u±(x, t) = p~±O
lim u«x, t) + pv(x, t))
we have, for all constants k,

lu+(x, t) - klcos(v(x, t), t) + sgn( u+(x, t) - k )[Jj(x, t, u+) - hex, t, k )]cos(v(x, t), x j)
218

::; lu-(x, t) - kjcos(v(x, t), t) + sgn( u-(x, t) - k )[/j(x, t, u-) - /j(x, t, k )]cos(v(x, t), Xj)

In order to state the principal results of Kruzkov we introduce the following no-
tation: Let

N = NM(R) = {x,t)E{lxl=:;R}
max [~)ou/j)2(x,t,uW/2
lul=:;M j

let /C denote the cone ({x,t); Ixl ::; R - NT, 0 ::; t::; To = min(T,~)} and
ST = /C n {t = T} for T E [0, To]

Theorem 2 (Kruzkov) [6]. Suppose

(i) /j,g E Cl(ZT X IR) and
(ii) the derivatives ox; /j and ot!j are Lipschitz continuous with respect to u on compact
sets. If u and v are generalized solutions with initial data Uo and Vo such that
lu(x, t)l, Iv(x, t)1 ::; M a.e. in {Ixl ::; R} x [0, T]
then, for almost all t E [0, To], we have

f lu(x,t) - v(x,t)ldx::; e"Yt f luo(x) - vo(x)ldx

1S T 1so
where I = max{x,t)EK,lul=:;M[-(Oug)(x,t,u)].
As a consequence we can deduce easily the uniqueness of solutions of the Cauchy
problem and also a comparison theorem.

Theorem 3 (Existence of generalized solutions) (Kruzkov) [6]. Suppose (i) /j(x, t, u) E

C 3 (ZT X IR)
(ii) (ou/j)(x,t,u) are uniformly bounded in DM = ZT x [-M,M]
(iii) F(x,t,u) = 'L,(oxJj)(x,t,u) + g(x,t,u) E C2(DM)
(iv) SUPZT IF(x,t,O)I::; const. and SUPZT l(ouF)(x,t,O)1 ::; const.
(v) Uo E LOO(IRR) with luo(x)1 ::; Mo
then there exists a generalized solution of the Cauchy problem in ZT.
The proof is essentialy based on the following steps: If Uo E W 3,OO(lRn ) then
under the assumptions made there exists a unique classical solution u. of the following
parabolic equation:

E(u.) = Otu. + divf(x,t,u.) + g(x,t,u.) = E~U.

with the viscosity parameter E. Then one proves that {u.(x,t)}, {Otu.(x, t)} and
{ox;u.(x,t)} are bounded In ZT, and finally that {Otu.} and
{ox;u.} are uniformly Holder continuous. These imply that u. tends to a limit u
a.e in ZT which is a generalized solution of the problem.
 219

3. THE CAUCHY-DIRICHLET PROBLEM FOR FIRST ORDER QUASI

LINEAR EQUATIONS ON A BOUNDED CYLINDRICAL DOMAIN
AND VISCOSITY SOLUTIONS

In this section we follow closely the paper [1] of Bardos et al. We consider the
following Cauchy-Dirichlet problem for the quasilinear hyperbolic equation:

E(u) = 8t u + div/(x,t,u) + g(x,t,u) = °

{ in the bounded cylindrical domain QT satisfying
the initial condition u(x,O) = uo(x)
u(x, t) = hex, t) on a part EI of the lateral part of the boundary E
where
(i) 1= (h,'" ,In) with /j = /j(x,t,u) and g(x,t,u) E G2(n X [O,T] X IR)
(ii) g, 8"';/i Lipschitz continuous with respect to u uniformly in QT and
(iii) Uo E G 2(n) and h E G2(E).
Writing v = u - h where h is a G 2 (n X [0, T]) extension of h we may assume u = °
on E I .

The vanishing viscosity method - Consider the parabolic problem associated to

the equation E( u) = ° E(u.) = €~u. in QT
{ u.(x,O)=uo«x) inn
u.(x, t) = °
on EI
Then there exists a classical solution u. (see the book of Friedman [3] for a proof).

Theorem of Compactness (Bardos, Leroux, Nedelec) [1]. The set of solutions

{u.}.>o is a bounded set in WI,I(QT). For this we estimate the norm of u. in
WI,I (QT). First of all, by the maximum principle, we have

where, denoting the Lipschitz norm of a function with respect to the variable u by
II· II Lipu , we have
C = IIgllLipu + L 118"'i/jIlLipu
i
Next we multiply the equation E(u.) = €~u. by sgn(8t u.), integrate by parts over n
°
and using the initial condition u.(x,O) = we obtain an estimate of the form

118t u.(x, t)IIL1(O) :5 C + c'IIuo(x)IIw2,l(O)

Similarly we take the scalar product of the vectors grad(E(u.) - €~u.) and
8E; h(grad(u.» and integrating over n we find an estimate of the form

IIgradu.(x,t)IIL1(O) :5l1uoll +C3t+

+C41t {ligradu.(x,s)IILl(O) + 118t u.(x,s)IIL1(O)}ds

220
Putting these three estimates together and using Gronwall's lemma we finally obtain
the estimate

(Ko is a constant) which proves the required assertion.

It follows from this that the set {U., f > O} is sequentially bounded Ll(Qr) and
that {u., f > O} is uniformly bounded in the space BV(Qr). Then we see that, if
{u<m} converges in Ll(Qr) to a function u then u E BV(Qr) and

lIu(x, t) - UO(X)IIL1(O) ~ Kot

Remark 4. It can be shown that if u = Ll(Qr) -lim U<m then u E BV(Qr) and
u(x,O) = uo(x) but u does not satisfy the homogeneous boundary condition on ~r.
This can be seen easily in the case of the well known Burger's equation

with u(x,O) = 1. An exact solution, given by the method of vanishing viscosity, for
t < lis
u(x, t) = {xlt
for x < t
1 forx>t
The difficulty consists in the fact that the speed of propagation depends on term
involving f(x, t, u) (which in the case of Burger's equation is ~~ = u.
These considerations suggest that the boundary condition should be reformulated
suitably. The following formulation of the problem was introduced in the paper of
Bardos, Leroux and Nedelec [1)

E(u) = 0 in Qr
{ u(x,O) = uo(x) in n
minkEJ('n::,u,O)(sgn('YE,u)[f(x,t,'n::,u)-f(x,t,k»),II(x,t» =0 on ~l

where J( 'YE, u, 0) denotes the closed interval [mine 'YE, u, 0), max( 'YE, u, 0»). We now have
the following definition of generalized solutions of the mixed problem.

Definition 2 (Bardos, Leroux, Nedelec). A function u E BV(Qr) is said to be

a weak solution of the Cauchy-Dirichlet problem if for all k E IR and for all ¢> E
C~omp(n x (0, T»
with ¢> 2: 0 we have

JkT [lu(x, t) - klat ¢> + sgn(u(x, t) - k) L)Ji(x, t, u(x, t» -Ji(x, t, k)}axj ¢>

-sgn(u(x,t) - k){2:0xJi(X,t,k) + g(x,t,u(x,t»}¢>]dxdt

+ JiT (sgn(k)[f(y,t,'YEU) - f(y,t,k)],II(y,t»)¢>(y,t)du(y)dt ~0
and u(x,O) = uo(x) a.e. in n.
 221

This definition generalizes that of Kruzkov for the Cauchy problem on the whole
space and hence without the boundary condition. The weak solution is the so called
entropy solution which characterizes the physically relevent solution among those ob-
tained by the vanishing viscosity method.
As we shall formulate the corresponding parabolic mixed problem (the associated
viscosity problem), taking into consideration of the above definition, for a more general
situation we shall postpone the details to the next section.
We have the following main result:

Theorem 4 (Bardos, Leroux and Nedelec) [1]. The Cauchy Dirichlet problem has a
unique weak solution given by the vanishing viscosity method.
The proof is again based on the vanishing viscosity method after an appropriate
weak formulation. We shall call a solution of the mixed problem in the sense introduced
by Bardos et al. a weak-entropy solution.

4. FIRST ORDER QUASI LINEAR EQUATIONS ON A BOUNDED

CYLINDRICAL DOMAIN WITH UNILATERAL CONSTRAINTS

In this section we shall come to our main problem of the inequation associated to
the first order quasilinear hyperbolic equation. We consider a unilateral regular obsta-
cle condition which is compatible with the boundary condition. After a translation we
may assume this to be zero.
Let Kdenote the cone {u E LOO(QT) n BV(QT)i u(x, t) :::: 0 a.e.in QT} and let
Uo : n -+ lR with Uo :::: 0 be given.
We shall be concerned with the following

Problem. To find (u, E'), with u E K and E' a subset of the lateral boundary ET of
QT, such that
E(u) = atu + divf(x,t,u) + g(x,t,u) = 0

{
on Q~ = {(x, t) E QTi u(x, t) > O}
u(x,O)=uo(x) a.e.on n
u(x,t) = 0 on E'
where we assume the following hypothesis:
(i) f E C2(n X [0, t] X lR, lR n ), f = (II,'" ,In)
(ii) 9 E Cl(n X [0, t] X lR)
(iii) u 1-+ g(x, t, u) and u 1-+ ax; hex, t, u) are Lipschitz continuous in u uniformly
with respect to (x,t)
(iv)uo E LOO(n) n BV(n) with Uo :::: 0 a.e. in n.
We introduce the following definition, a slight modification of the definition of
weak solutions introduced in the previous section.

Definition 3. A function u(x, t) is said to be a weak-entropy solution of the problem

of inequation if u E K such that for all real k :::: 0 and for all ¢ E C1omp(n X (0, T))
222

with <I> ~ 0 we have

f kT [lu(x,t) - kl8t <l> + sgn(u(x,t) - k) L{!i(x,t,u(x,t)) - fj(x,t,k)}8xj <l>

-sgn(u(x, t) - kHL 8x Jj(x, t, k) + g(x, t, u(x, t))}<I>]dx dt

+ ~T f (sgn(k)[f(y,t,iEIU) - f(y,t,k)],v(y,t»)<I>(y,t)dO"(y)dt ~0
and u( x,D) = uo( x) a.e. in n. where v denotes the exterior normal vector field on I;T
defined (dO" x dt) - almost everywhere.
We have the following main theorem due to Levi [9] (a similar result was obtained
independently also by the author).

Theorem 5 (Levi) [9]. Given Uo E LOO(n) n EV(n) with Uo ~ 0 a.e. in n there

exists a unique entropy solution u E LOO( QT) n EV( QT) of the inequation and can be
obtained as a strong limit in LP( QT), 1 ::; p < +00 of weak-entropy solutions u., of a
family penalized problems as TJ ~ 0+.

5. IDEA OF THE PROOF

Here we give a very brief sketch of the idea of the proof. The proof is based on the
method of penalisation to reduce to a problem associated to a quasi linear equation of
the type studied by Bardos, Leroux and Nedelec in [1]:
To find, VTJ > 0 a solution u.,(x, t) and I;' C I;T satisfying

E(u ., ) - lu-.,.,
= 8tu., + div f(x, t, U.,) + g(x, t, U.,) - lu;;- =0
{ .
III
QT
U.,(x,O) = uo(x)
U.,(y, t) = 0 on
a.e. on n
I;'
.,

that is, V<I> E C~omlIT x (0, T)) with <I> ~ 0 we have

J~T [lu'l- kl8t <l> + sgn(u'l- k) L{!i(x,t,u'l) - !i(x,t,k)}8xj <l>

-sgn(u'l - kHL 8xj !i(x, t, k) + g(x, t, u'l)}<I>]dx dt

+ JhT (sgn(k )[J(y, t, iE' u'l) - fey, t, k )], v(y, t») <I>(y, t)dO"(y)dt =0
Then we have the corresponding parabolic problem (with the viscosity term added),
namely,

E(u'l .• ) - ~u;;-.•
= 8tu'l •• + divf(x,t,u'l .• ) + g(x,t,u'l .• ) - ~u;;-.• = f/~.U'l.f
{ in QT
u'l .• (x,O) = uo(x)
u'l .• (y, t) = 0 on
a.e. on
I;'
n
 223

The weak form of this equation is formulated in the same way as above. By the result of
Bardos et al. (section 3) there exists a unique entropic solution u" E LOO( QT )nBV( QT)
obtained by the method of vanishing viscosity: let U",E be the unique viscous solution
of the parabolic problem (e > 0 is the viscosity parameter).
One proves that the family {U",E, e > 0, '7 > O} is bounded in the space Wl,l( QT)n
LOO(QT)' By using the monotonicity of the penalising operator (3(u) = - U - this
bounded set is seen to be independent of the penalisation parameter '7. It then follows
that, first passing to the limits as e -+ 0+ the family of solutions of the penalised
hyperbolic problems belong to a fixed bounded set in the space LOO(QT) n BV(QT).
Finally, using the compactness of the inclusion Ll(QT) n BV(QT) in L1(QT) we can
pass to the limits as '7 -+ 0+. By the result of Bardos et al. this limit is the required
unique weak-entropy solution.

REFERENCES

[I] C. Bardos, A.Y. Leroux and J.C. Nedelec, "First order quasilinear equations with boundary
conditions", Comm. Partial Differential Equations 4, 1017 - 1034, 1979.
[2] E. Conway and J. Smoller, "Global solutions of the Cauchy problem for quasi-linear first order
equations in several space variables", Comm. Pure Appl. Math. 19), 95 - 105, 1966.
[3] A. Friedman, "Partial differential equations of parabolic type", Prentice Hall, New York, 1969.
[4] E. Hopf, "The partial differential equation Ut + UU., = Jlu:u", Comm. Pure Appl. Math. 3,
201 -230, 1950.
[5] O.A Ladyzenskaja and N.N. Ural'ceva, "Boundary problems for linear and quasilinear parabolic
equations, I - II", Iw. Akad. Nauk SSSR, 26, 2 - 52, 1964'.
[6] S.N. Kruzkov, "First order quasilinear equations with several independent variables", Math.
USSR - Sh., 10,217 - 243,1970.
[7] S.N. Kruzkov, "Methods for constructing generalized solutions for the Cauchy problem for a
quasilinear equation of the first order", Uspehi Mat. Nauk 20, 112 - 118, 1965.
[8] P.O. Lax, "Shock waves, increase of entropy and loss of information", Publ. Math. Sci. Inat.,
2, 129 - 243, 1984.
[9] L.Levi, "Equations quasi lineaires du premier ordre avec contrainte unilaterale", C.R. Acad. Sci.
PaN, 317, 1133 - 1136, 1993.
[10] J .-L. Lions, "Quelques methodes de resolution des problemes aux limites non lineaires", Dunod,
Gauthier - Villars, Paris, 1969.
[11] F.Mignot and J.P. Puel, "Inequations variationelles et quasi variationelles hyperboliques du
premier ordre", J. Math. Pure. Appl. 55, 353 - 378, 1976.
[12] A.I. Vol'pert, "The spaces BV and quasilinear equations", Math. USSR Sb., 2, 225 - 267, 1967.
CONTINUITY OF THE SOLUTION SET
OF HOMOGENEOUS EQUILIBRIUM PROBLEMS
AND LINEAR COMPLEMENTARITY PROBLEMS

Oettli Wernerl and Yen Nguyen Dong

1 Universitat Mannheim
Fakultat fiir Mathematik und Informatik
68131 Mannheim, Germany
2 Hanoi Institute of Mathematics
P.O. Box 631 Bo Ho
Hanoi, Vietnam

Abstract: Denote by S(M, q) the solution set of the linear complementarity prob-
lem

z~O, Mz+q~O, (z,Mz+q)=O,

where ME R,nxn and q E lRn. M is called an Ro-matrix iff S(M,O) = {O}. Jansen and
Tijs have proved that if M is an Ro-matrix, then the map S is upper semicontinuous at
(M, q) for every q E lRn. We prove that this property is characteristic for Ro-matrices.
Part of our results extends to homogeneous equilibrium problems of the type

z E K, J(z, y) + (q, y - z) ~ 0, Vy E K.

Here K C lRn is a closed convex cone and J : K x K -+ lR is such that

J(>..x, >.y) = V+l J(x, y) "Ix, y E K, V>' ~ 0, where p > 0 is a fixed constant.

Key Words: Linear complementarity problem, solution map, upper semicontinuity,

nonlinear equilibrium problem.

1. INTRODUCTION AND PRELIMINARIES

Linear complementarity problems with Ro-matrices will be considered, and it

will be shown that this class of problems admits a generalization in the context of
homogeneous equilibrium problems.

The linear complementarity problem LCP(M,q), defined by a matrix M E lR nxn

225
226

and a vector q ERn, consists in finding z ERn such that

z;:::O, Mz+q;:::O, (z,Mz+q) =0. (Ll)

Here z ;::: a means that z belongs to the nonnegative orthant R+ of nn, and (.,.)
denotes the inner product in nn. The set of all z satisfying (1.1) is denoted by SCM, q).

Linear complementarity problems play an important role in the theory of Opti-

mization (see [3] and [13]). A great deal of work has been devoted to studying the
solution set SCM, q) as a multivalued mapping of the variables M and q (see [3],
[4], [6] - [12], [14], [15], and references therein). If M is fixed, there exists a con-
stant 1 > 0 such that for each q ERn there is a neighborhood U of q such that
SCM, q') ~ SCM, q) + lllq' - qll Bn for every q' E U, where Bn denotes the unit ball
in nn. This upper Lipschitz continuity property was obtained by Robinson (see [15],
Proposition 1). If M is a copositive matrix (that is (x,Mx) ;::: a for every x E R+)
and {z E SCM, 0) : (q, z) ~ a} = {a}, then for each open set n :J SCM, q) there exists
6 > 0 such that S(M', q') c n for every copositive matrix M' and vector q' satisfying
11M' - Mil < 6 and Ilq' - qll < 6. This result is due to Gowda (see [6], Theorem 3).
Together with Theorems 4 and 8 from the same paper, it generalizes the well-known
Theorem 2 in [14].

Definition 1. M is said to be an Ro-matrix iff S(M, 0) = {a}. The set of all Ro-
matrices of order n is denoted by Ro. (In [4], [5], and [10], the same set is denoted by
E*(O). In [8] Ro-matrices are called also pseudo-regular matrices.)

Definition 2. The solution map S is said to be upper semicontinuous (u.s.c.) at

(M, q) iff for each open set n :J SCM, q) there exist neighborhoods V of M and U of
q such that S(M',q') en for every (M',q') E V x U.
A result by Jansen and Tijs ([10], Theorem 5.6) says that, if M E R o, then for
any q E R n the map Sis u.s.c. at (M, q). Upper semicontinuity of the solution map at
a point is a good property for numerical solution. Note that there are other stability
concepts relying on lower semicontinuity of the map S.

In Sect. 2 we consider a class of homogeneous equilibrium problems, for which an

analogy of the result by Jansen and Tijs mentioned above is valid. In Sect. 3 we show
.. l.~ .. ;r ~;o" a .. "t (M q) for each q ERn, then M E Ro.

2. HOMOGENEOUS EQUILIBRIUM AND COMPLEMENTARITY

PROBLEMS

First of all we consider a somewhat more general class of problems, namely ho-
mogeneous equilibrium problems.

We fix a nonempty, closed, convex cone K eRn, with polar cone

K* .- {e E R n : (x,e) ;::: 0, Vx E K}. We assume that K is pointed, i.e.
 227

K n (-K) = {O}. Then there exists a linear form (I,·) on R n such that

(I, x) > 0, 'Ix E K, x "10. (2.1)

In fact, if K is pointed, then int K* "10, and every IE int K* has property (2.1). From
now on, some I which satisfies (2.1) is fixed. We set B := {x E K : (l,x) S; I}. Denote
by F the space of all continuous functions f: K x K - t R such that f(O,y) ~ 0 for
all y E K, and
f(>..x, >..y) = >..p+l f(x, y), 'Ix, y E K, V>" ~ 0, (2.2)

where p > 0 is a fixed constant. The norm on F is defined as

max If(x, x + y)l·

Ilfll:= (x,y)EBxB (2.3)

For f E F and q E R n we consider the following equilibrium problem [1]:

zEK, f(z,y)+(q,y-z)~O, VyEK. (2.4)

Let S(J, q) denote the solution set of (2.4). We consider the multi valued mapping S(·,·)
defined on F x R n with values in Rn. Using (2.2), (2.3), (2.4) it is easy to see that S
has closed graph. The latter means that if!k - t fin F, qk - t q in R n , Zk E S(Jk, qk),
and Zk - t Z in R n , then Z E S(J, q). Note that S(J,O) is a nonempty cone. We define

Fo:= {J E F: S(J,O) = {O}}. (2.5)

S is said to be upper semicontinuous at (J, q) E F x R n iff for each open set n :::> S(J, q)
there exist neighborhoods V of f in F and U of q in R n such that S(J', q') c n for
every (J', q') E V x U.

Theorem 1. Let f E F. Then the following properties hold:

(a) If f E F o, then for every q E R n the set S(J' q) is bounded.
(b) If f E Fo, then for every q E R n the map S is upper semicontinuous at (f, q).
(c) Fo is open in :F.
(d) Assume that q ERn is such that S(f,q) is bounded (may be empty). If S(',q) is
upper semicontinuous at f, then f E Fo.

Proof: (a) Assume that f E Fo, q ERn, but S(f,q) is unbounded. Then there exist.s
a sequence {zd eRn such that IIZkll - t 00 and, for all k,

zkEK, f(Zk,y)+(q,y-Zk)~O, VyEK. (2.6)

Let Vk := zk/llzkll. There exists a subsequence of {vd, denoted again by {vd, such
that Vk - t v "I O. Taking account of (2.2) we get from (2.6) that

y 1 y
Vk E K, f(vk, IIZkll) + IIZkll p (q, IIZkll - Vk) ~ 0, Vy E K,
228

and therefore
1
Vk E K, VTJ E IIzkll K = K.
Passing to the limit, we have

vEK, f(v,TJ) "20, VryE/{.

Then v E S(f,O). Since f E Fo, this contradicts with v =I o.

(b) Let f E Fo and q E JR n. If S is not u.s.c. at (f,q), then there ex-
ist an open set 0 :J S(f,q) and sequences Uk} C F, {qk} c JRn such that
!k -+ f, qk -+ q, Zk E S(fk, qk) and Zk f/. 0 for every k. If there exists a subsequence
of {Zk} converging to some z, then Z E S(f,q) CO, because S has closed graph. On
the other hand Z f/. 0, a contradiction. Thus we can assume that IIzkll -+ 00. Without
loss of generality we can assume that Vk := zk/llzkll -+ v =I o. Then from

it follows, as from (2.6) in the proof of part (a), that v E S(f,O). Since f E F o , this
contradicts with v =I O.
( c) Assume that Fo is not open in F. Then there exists 1 E Fo and a sequence
Ud C F such that fk -+ 1 and !k f/. Fo, i.e., S(fk, 0) =I {O}. Since S(fk, 0) is a
nonempty cone there exists for every k some Zk E S(fk, 0) with IIzk II = 1. We may
assume that Zk -+ z =I o. Then z E Sa,O), since S has closed graph. Together with
Z =I 0 this gives 1 f/. F o, a contradiction.
(d) Ab absurdo, let us suppose that S(·,q) is u.s.c. at /, and / f/. Fo. Then there
exists z =I 0 such that
z E K, fez, y) "2 0, Vy E K. (2.7)

Clearly (I, z) > o. Now for all t > 0 let

(l,X})P _
ft(x,y):=J(x,y)- ( t(I,Z) (q,y-x). (2.8)

We have ft()..x, )..y) = )..p+l ft(x, y) for all)" "2 0, so that ft E F. Moreover, the following
holds:
ft(Zt, y) = f(zt, y) - (q, y - Zt).
Thus for all y E K we have

1
ft(Zt, y) + (q, y - Zt)= f(zt, y) = tp+l fez, ty) "2 0,
from (2.7) and since ty E K. Therefore Zt E S(ft,q) for all t > O. Observe that ft -+ f
as t ! O. Since the set S(f, q) is bounded we can choose a bounded open set 0 :J S(f, q).
Since S(·,q) is u.s.c. at f we should have Zt EO for all t > 0 sufficiently small. But
this is not possible since II Zt II -+ 00 for t ! O. This completes the proof. 0
 229

It can be shown (this is a particular case of [1], Theorem 2) that for (f,q) E Fx R,n
problem (2.4) has a solution if I is convex in the second argument, I(x,x) = 0 for all
x E K, and moreover the following conditions hold:
(i) I is copositive in the sense that l(x,O) :::; 0 for all x E K.
(ii) I is q-regular in the sense that {z E S(f,O) : (q, z) :::; O} = {OJ.
These same conditions (i), (ii) ensure also that S is u.s.c. at (f, q), provided we restrict
the first argument of S to copositive functions. To make this precise, let V denote the
set of all I E F such that I is copositive. Then we have the following extension of
Gowda's result quoted in Sect. 1.

Theorem 2. Let (f, q) E V x R,n. Then the following properties hold:

(a) If I is q-regular, then S IvxRn is upper semicontinuou3 at (f, q).
(b) Assume that q E -K* and that S(f, q) is bounded (may be empty). If S(" q) Iv is
upper semicontinuous at I, then I E Fo.

Proof: (a) Assume, for contradiction, that S IVxRn is not u.s.c. at (f, q). Then
there exists an open set n :::> S(f,q) and sequences {fd c V, {qd c R n such that
Ik -+ I, qk -+ q, Zk E S(!k, qk) and Zk f/. n for every k. If there exists a subsequence of
{Zk} converging to some z, then we obtain simultaneously Z E S(f, q) and Z f/. n, which
is impossible. Thus we can assume that IIZkll -+ 00 and Vk := zk/llzkll -+ v =I- O. From
Zk E S(!k,qk) it follows, as in the proof of part (b) of Theorem 1, that v E S(f,O).
Moreover it follows from Zk E S(fk, qk) that

Since h, is copositive this implies (qk, -Zk) ~ 0, and in the limit (q,v) :::; O. Since I is
q-regular we must have v = 0, which contradicts v =I- O.
(b) Ab absurdo, suppose that S(·,q) Iv is u.s.c. at I, but I f/. Fo. Then there exists
Z =I- 0 such that z E S(f,O). For all t > 0 let Zt and It(x, y) be given by (2.8). As
in the proof of Theorem 1, part (d), it follows that It E F and Zt E S(ft,q) for all
t > O. Moreover, since I is copositive and q E -K* we have It(x, 0) :::; 0 for all x E K.
Hence It E V for all t > O. Since S(f, q) is bounded we can choose a bounded open
set n :::> S(f, q). Since It -+ I for t 1 0 and since S(" q) Iv is u.s.c. at I we should have
Zt En for all t > 0 sufficiently small. But this is impossible since IIZtll -+ 00 for t 1 O.
This completes the proof. 0

We observe that part (b) of Theorem 1 yields simultaneously the upper semiconti-
nuity of S(·,q) in an entire neighbourhood of the given element I E Fo, since Fo is
open by part (c) of Theorem 1. Likewise part (a) of Theorem 2 yields without further
assumptions the upper semicontinuity of S IVxRn in an entire neighbourhood of (f, q)
in V x R n. In fact, if I is q-regular, then there exists a neighbourhood U of (f, q) in
F x R n such that f' is q'-regular for all (f', q') E U. The proof of the latter fact is
like the proof of part (c) of Theorem 1.
230

We turn now to homogeneous complementarity problems. Let F c :F denote the

space of all functions I : J{ x J{ - t R such that

I(x,y) = (M(x),y -x), (2.9)

where M(·) : J{ -t RR is continuous and positively homogeneous of degree p, i.e.,

M(Ax) = APM(x), Vx E J{, VA ~ O. (2.10)

Here p > 0 is fixed, and is the same as in the definition of:F. For I E F and q E RR
problem (2.4) becomes the homogeneous complementarity problem

zEJ{, M(z)+qEJ{*, (z,M(z)+q) =0, (2.11)

since for every fixed z E J{ and ~ E RR one has the following equivalence:

((y-z,~)~O, VYEJ{){=:}(~EJ{*, (z,~)=O).

For IE F and q E lRR let BU, q) denote the solution set of (2.4) or, which is the same,
of (2.11). Let Fo := {J E F: BU,O) = {O}}.
Let M be the space of all continuous mappings M(·) : J{ - t R n which satisfy
(2.10), and provided with the norm IIMII:= max
(x,y)EBxB
I(M(x), y)l. Clearly we may
identify F with M, and 11111 = IIMII if I and M are related via (2.9). For M E M
and q E RR we let S(M,q) denote the solution set of (2.11). Moreover we define
Mo := {M EM: S(M, 0) = {On. Clearly Mo can be identified with Fo via (2.9).

Theorem 3. Let M E M. Then the following properties hold:

(a) If MEMo, then for every q E lR n the set S(M,q) is bounded.
(b) If MEMo, then for every q E lRR the map S is upper semicontinuous at (M,q).
(c) Mo is open in M.
(d) Assume that q E lRR is such that S(M,q) is bounded (may be empty). If S(',q) is
upper semicontinuous at M, then MEMo.

Proof: We identify M E M with I E j via (2.9), and take into account the equiva-
lence of (2.4) and (2.11) under this identification. Then the proof is a word-by-word
replica of the proof of Theorem 1, where we have to replace throughout:F by F,:Fo by
Fo, and S by S. Concerning the proof of part (d) we note that if f(x, y) = (M(x), y-x)
with M

Mt(x) := M(x) - (t ~~: ;~ r

E M, and It is given by (2.8), then It(x,y)
q. Then M t
= (Mt(x),y
EM, and we have it E j.
- x) with

Rearranging the content of parts (a), (b), (d) of Theorem 3 we obtain the following
characterization.

Corollary 1. Let ME M. Then MEMo if, and only if, for some (equivalently: for all)
qERR the set S(M,q) is bounded and the map S(',q) is upper semicontinuous at M.
 231

There exists q E R n such that SCM, q) is bounded if, for some A > 0, the set
M(K \ AB) is convex and

M(K \ )"B) n int K* = 0. (2.12)

In fact, it follows from (2.12) and the separation theorem ([16], p. 97) that there exists
a nonzero vector /-' ERn such that

(/-"M(x)) ~ inf (/-',0, VXEK\AB. (2.13)

{El(·

Since K* is a cone, (2.13) implies that /-' E K** = K and

(/-"M(x))~O, VxEK\)"B.

Therefore, setting q := -fl, we have

and so M(x) + q ~ K*. Consequently x ~ S(M,q) for all x E K \ AB. Hence

SCM, q) C AB is bounded.

3. LINEAR COMPLEMENTARITY PROBLEMS

We turn now to the linear complementarity problem LCP(M, q), see (1.1). To this
end we have to assume that K := JR+, and that p = 1 in (2.10). R nxn can be con-
sidered as a subspace of M, by identifying the matrix M E R nxn with the linear
mapping M(·) E M. For consistency we have to provide R nxn with the seminorm
IIMII:= max I(Mx,y)l. If M E R nxn and K = R+, then problem (2.11) coin-
(x,y)EBxB
cides with the linear complement8.l·ity problem (1.1). So does problem (2.4), if we set
f(x,y) := (Mx,y - x). The solution mapping S(M,q) is now defined on R nxn x JR n ,
and 'Ro , as introduced in Sect. 1, coincides with M o n JR nxn .

Theorem 4. Let M E JR nxn . Then the following properties hold:

(a) If ME 'Ro , then for every q E JR" the set SCM, q) is bounded.
(b) If ME 'Ro, then for every q E R" the map S is upper semicontinuous at (M,q).
(c) 'Ro is open in Rnxn.
(d) Assume that q ERn is such that SCM, q) is bounded (may be empty). If S(·, q) is
upper semicontinuous at M, then M E 'Ro.

Proof: The proof is completely analogous to the proof of Theorem 1, where we replace
F by R nxn and Fo by 'Ro. 0
232

Parts (a)-(c) of Theorem 4 have already been proved by Jansen and Tijs [10]. Gowda
[6] removed the hypothesis SCM, q) -10.

For LCP there is an additional fact (which probably fails to hold for the more general
cases considered in Sect. 2), namely:

Lemma 1. Let M E lR nxn and K = lR~. Then there exists q E lRn such that SCM, q)
is bounded.

Proof: Let q E lR n be such that S(M,q) is non empty and unbounded. Then there
exists a sequence {zk} C S(M,q) with IIzkll-+ 00. Let v k := Mzk + q. Then for all k
we have
Zk 2: 0, v k 2: 0, (zk, v k ) = O.
Let J(zk) := {j : zJ > OJ. Without any loss of generality we may assume that
J(zk) =: J is independent of k. Let I denote the complement of J. We have for
all k that
q=_MZk+V k .

Since zJ = 0 for j ~ J, and vJ = 0 for j ~ I, this can be written as

q= I>j(-Mj ) + LvjIj, (3.1 )

jEJ jEl

where Mj denotes the j -th column of the matrix M, and Ij denotes the j -th column
of the (n x n) unit matrix I. Now II(zk,vk)ll-+ 00, and without any loss of generality
we may assume that (zk,vk)/lI(zk,vk)ll-+ (z,v) -I (0,0). Dividing (3.1) by II(z\vk)11
and passing to the limit we obtain

(3.2)

From (3.1) it follows that q is an element of the convex conical hull of the vectors
-Mj (j E J) and Ij (j E I). From (3.2) it follows that these n vectors are linearly
dependent. Thus, their convex conical hull has dimension less than n, and it is closed.
The union of finitely many closed convex sets in lR n, each having dimension less than
n, cannot equal all of lR n (as a consequence of Baire's Lemma, see [2], p. 15). In fact,
its complement is open and dense in lRn. Hence there must remain some q E lR n such
that S(M,q) is bounded (may be empty). 0

Combining Lemma 1 with parts (b) and (d) of Theorem 4 we obtain the following
characterization:

Theorem 5. Let M E lRnxn. Then M E Ro if, and only if, S(·,q) is upper semicon-
tinuous at M for all q E lR n.

In addition to part (c) of Theorem 4 we have the following result:

 233

Theorem 6. 'Ro is dense in lRnxn.

Proof: Assume that M E lR n x nand M rfc 'Ru. Then there exists zf 0 such that

z 2 0, Mz 2 0, (z, Mz) = o.

Let v := Mz. Let J := {j : Zj > O} and J the complement of J. Since ZjVj = a for all
j, we have

0= -Mz+v= LZj(-Mj ) + LVjIj, (z,v) f (0,0),

jEJ jEJ

where Mj and Ij are as in the proof of Lemma 1. Hence the (n x n)-matrix MJ

which consists of the columns -Mj for j E J and Ij for j E J, is singular. Therefore
det M J = O. Since J f 0 the determinant det M J is a nonzero polynomial in the
elements mij of M. The set II( J) := {M E lR n x n : det M J = a} is closed and has
empty interior, hence is nowhere dense in lRnxn. Then, by Baire's Lemma (see [2], p.
15),
U {II(J) I JC {I,··· ,n},Jf0}
is nowhere dense in JR nxn . So the complement of this set is dense in lRnxn. Since this
complement is contained in 'Ro, 'Ro is dense in JR nxn . o

REFERENCES
[1] E. Blum and W. Oettli, "From optimization and variational inequalities to equilibrium problems",
Math. Student 63, 123-145, 1994.
[2] H. Brezis, "Analyse fonctionnelle", Masson, Paris, 1983.
[3] R.W. Cottle, J.-S. Pang, and R.E. Stone, "The Linear Complementarity Problem", Academic
Press, New York, 1992.
[4] R.D. Doverspike, "Some perturbation result.s for the linear complementarity problem", Math. Pro-
gramming 23, 181-192, 1982.
[5] C.B. Garcia, "Some classes of matrices in linear complementarity theory", Math. Programming 5,
299-310, 1973.
(6) M.S. Gowda, "On the continuity of the solution map in linear complementarity problems" , SIAM
1. Optimization 2, 619-634, 1992.
[7] M.S. Gowda, "Applications of degree theory to linear complementarity problems", Math. Oper.
Res. 18, 868-879, 1993.
(8) M.S. Gowda and J .-S. Pang, "On solution stability of the linear complementarity problem", Math.
Oper. Res. 17,77-83, 1992.
[9) C.D. Ha, "Stability of the linear complementarity problem at a solution point", Math. Program-
ming 31, 327-338, 1985.
[10] M.J.M. Jansen and S.H. Tijs, "Robustness and non degenerateness for linear complementarity
problems", Math. Programming 37, 293-308, 1987.
234

[11] O.L. Mangasarian and T.-H. Shiau, "Lipschitz continuity of solutions of linear inequalities, pro-
grams and complementarity problems", SIAM J. Control Optim. 25, 583-595, 1987.
[12] K.G. Murty, "On the number of solutions to the complementarity problem and spanning properties
of complementarity cones", Linear Algebra Appl. 5, 65-108, 1972.
[13] K.G. Murty, "Linear Complement.arity, Linear and Nonlinear Programming", IIeldermann- Verlag,
Berlin, 1987.
[14] S.M. Robinson, "Generalized equations and their solutions, Part I: Basic Theory", Math. Pro-
gramming Study 10, 128-141, 1979.
[15] S.M. Robinson, "Some continuity properties of polyhedral multifunctions", Math. Programming
Study 14, 206-214, 1981.
[16] R.T. RockafeUar, "Convex Analysis", Princet.on University Press, Princeton, 1970.
TENSOR APPROXIMATIONS OF SMOOTH NONLINEAR
COMPLEMENTARITY SYSTEMS

Rapcs8.k Tamas

Computer and Automation Institute,

Hungarian Academy of Sciences
Lagymanyosi utca 11,
1111 Budapest, Hungary.
e-mail: h632rap@ella.hu

Abstract In order to analyse the structure of nonlinear Complementarity Systems

with twice continuously differentiable functions, first-order and second-order tensor
approximations based on Taylor expansions are introduced. The usefulness of this
approach is shown by suggesting a class of penalty methods based on globally conver-
gent variable metric algorithms along geodesics for solving nonlinear Complementarity
Systems.

Keywords: Nonlinear Complementarity Systems, Riemannian manifolds, tensors,

penalty functions, variable metric methods along geodesics.

1. INTRODUCTION

Despite the interest in finite-dimensional smooth nonlinear complementarity sys-

tems (in short, NLCS) and Variational Inequalities (in short, VI), an adequate treat-
ment of the theory, algorithms and applications of these two problems has not been
adapted yet. At the beginnig, most effort was concentrated on the existence of solu-
tions and development of algorithms. The paper deals with the structure of NLCS.
The approach given here is a coordinate-free description with the notion of tensor and
tensor calculus. In theoretical physics, differential geometry and several applications
of mathematics, the concept of tensor proved to be instrumental, which motivated the
use of this notion within the frame of complementarity systems (in short, CS).
The aim is to produce a unified framework for both theoretical and algorithmical
aspects of NLCS invariant under nonlinear coordinate transformations, so that the
theoretical and numerical parts of the results should be separated, the theoretical
aspects could be better emphasized without including the technical details related to

235
236
coordinate representations and a new view concerning the structure of NLCS could
be shown. This approach develops further the results in papers (Rapcsak, 1989, 1991;
Rapcsak and Csendes, 1993; Rapcsak and Thang, 1994).
In Section 2 the problem will be set, in Section 3 a new notion, the tensor field
CS will be introduced defining a generalization of NLCS. Section 4 contains a tensor
formula with first-order and second-order tensor approximations of smooth functions
defined on a Riemannian manifold, based on Taylor expansions, by using covariant
derivatives and some consequences of this approach will be considered. In Section 5
the usefulness of this notion will be demonstrated by suggesting a class of penalty
methods based on globally convergent variable metric algorithms along geodesics of a
Riemannian manifold for solving NLCS.

2. SMOOTH NONLINEAR COMPLEMENTARITY SYSTEMS

Let 9 : R n -+ R n be a twice continuously differentiable function, and consider the

following NLCS:

ht(x) = xTg(x) = 0, g(x) ~ 0, x ~O, (2.1)

where we assume that h~(x)T = g(x) + g'(x)T x =F 0 if ht(x) = 0, x E R. n , n ~ 2; h~

and g' denote gradient and Jacobian matrix, respectively. In the paper, the gradients
are always row vectors. Let us introduce the following notations:

M[htl = {x E Rnl ht(x) = O}, (2.2a)

Ai' = {x E M[htll g(x) ~ 0, x ~ O}, (2.2b)
h~(x) T
= IhHx)12 g j(x),
I

I'j(x) j=1, ... ,n, (2.2c)

where h~(x) =F 0, x E M[htJ. The equality in (2.1)) identifies an (n - 1)-

dimensional Riemannian manifold M[htJ, where the Riemannian metric is induced
by the Euclidean one of Rn. Consider a more general NLCS as follows:

g(x) ~ 0, g(x) ~ 0, (2.3)

where 9 and 9 : R n -+ R n are twice continuously differentiable vector-valued

functions. Assume that h~(x)T = g'(x)T g(x) + g'(x)Tg(x) =F 0 if h2(X) = 0, x E R.n
and n ~ 2. Introduce the following notations:
 237

M[h2J = {x E R"I h2{X) = O}, (2.4a)

A; = {x E M[h2J1 g{x) ~ 0, g{x) ~ O}, (2.4b)
h~(x) T
i=l, ... ,n,
,I
v;(x)=lh;(x)12g;(x), (2.4c)

h;(x) T
= = 1, ... ,n,
I

r/j(x) Ih;(x)1 2gj(x), j (2.4d)

where h;(x) =I 0, x E M[h2J.

CS can be considered one of the traditional approaches for solving equilibrium
models in nonlinear optimization, showing the importance. The solution set of a
complementarity system is, in general, the union of disjoint sets (often of points),
therefore, both from theoretical and numerical points of view structural questions
should be of interest.

3. TENSOR FIELD COMPLEMENTARITY SYSTEMS

First, the notion of the tensor is recalled, thereafter the tensor and tensor field
CS will be defined. Let M be a C 2 n-manifold and m a point in M. The tangent
space TMm at m is an n-dimensional vector space. Let TM:;' be the dual space of
TMm, TM:;' endowed with its natural vector space structure. Now, the theory oflinear
algebra can be applied to define tensors (e.g., Spivak, 1979).

Definition 3.1 A p-covariant tensor at m E M (for p > 0) is a real-valued p-linear

function on TMm x TMm X .•. X TMm (p-copies).
A q-contravariant tensor at m E M (for q > 0) is a real-valued q-linear function
on TM:;' X TM:;' X •.• x TM:;' (q-copies).
A p-covariant and a q-contravariant tensor at m E M is a (p+q)-linear real-valued
function on (TMm)P x (TM:;.)q.
A tensor is symmetric if its value remains the same for all possible permutations
of its arguments. A tensor field on M (or a subset A of M) is a mapping that assigns
a tensor at m to each m in M (or A). A O-covariant tensor at m is a real number.

Definition 3.2 A tensor field CS is defined on A ~ M if the problem can be given in

the form of

Tj(m'Vl' ... 'Vp,V~, ... ,V;) ~ 0, j = 1, .. . ,n,

T;(m,vl, ... ,vp,v~, ... ,v;) ~ 0, i = 1, ... ,n,
n
(3.1)
LT;(m, VI,···, vp, v~, ... , v;)T;(m, VI, ... , vp, v~, ... , v;) = 0,
i=l
238

where T;, T;, i = 1, ... , n are p-covariant and q-contravariant tensor fields on A, respec-
tively. If all the tensor fields are considered at a fixed point mEA, a tensor CS is
obtained at m E M.
Since the inequality constraints of problems (2.1) and (2.3) are O-covariant tensor
fields on R n , thus problems (2.1) and (2.3) are special tensor field CS on Rn. Moreover,
the only equality constraint in (2.1) and (2.3) results in a manifold, thus the inequality
constraint functions can be considered on this manifold. Tensor field CS can be origi-
nated, e.g., by characterizing the equilibrium positions of a system of material points
on which active forces have an effect, mainly if the constraints depend on the veloc-
ity and the system is scleronomous and non-holonomous (e.g. Rapcsak and Szenthe,
1990). Tensor field optimization problems fitting better to the structure of a non-
linear optimization problem class whose equilibrium systems may be also formulated
as tensor field CS can be similarly defined (Rapcsak and Csendes, 1993). Since the
objective function of a general optimization problem defining a constraint set A ~ M
in a Riemannian manifold is a O-covariant tensor field on A, thus these optimization
problems are unconstrained tensor field optimization problems on A ~ M.
It turns out from Definition 3.2 that neither the values of the tensors nor the
solutions of problems (3.1) change by a nonlinear coordinate transformation, thus this
problem class can become an adequate tool to study the structure of problems (2.1)
and (2.3). Consider an arbitrary coordinate representation of the constraints manifold
M[hd or M[h2J in a neighbourhood of any point m in problem (2.1) or (2.3). Then,
a tensor CS assigned to a given point m and to a coordinate representation can be
considered an NLCS with a special structure, i.e., all the functions are multilinear.
This property may be useful for elaborating efficient algorithms.
To build and study tensor field CS the operations of tensor algebra (addition,
subtraction, multiplications and contraction) and of tensor analysis (covariant dif-
ferentiation) can be applied, as well as other operations which preserve the tensor
character.
In the paper, covariant differentiation (e.g., Spivak, 1989; Gabay, 1982) will be
the most important tool. In a system of local coordinates, the coefficient functions
of covariant differentiation (affine connection) r:~ '
2' 11, 12 , 13
n
= 1, ... , n define the co-
variant derivative for all the tensor fields. Let vr = E Vi3r:~'2' 11 ,12 = 1, ... , n,
1.=1
= 1, ... , n are the component functions of a covariant vector field, and let
where Vi., 13
n
vr = E V" r:~'2' 12 , h = 1, ... ,n, where V", 11 = 1, ... , n are the component func-
1,=1
tions of a contravariant vector field. The following result is well-known in differential
geometry (e.g., Spivak, 1979):

Proposition 3.1 On a covariant vector field V, the covariant derivative, denoted by

D, is equal to

DV= JV- vr, (3.2)

 239

while on a contravariant vector field

DV = Jv+vr, (3.3)

where JV denotes the Jacobian matrix of the corresponding vector field and vr is
the multiplication of the vector field and the 3-dimensional matrix r at each point
of an arbitrary coordinate neighbourhood. For an arbitrary tensor field the covariant
derivative forms a tensor field. If the tensor field is scalar (i.e., a smooth function on
M), then the covariant derivative is equal to the gradient.

Definition 3.3 If

r'a','2 -- r'a'2', forall /t,12,h=1, ... ,n, (3.4)

in every system of local coordinates, then the affine connection is symmetric.

In Optimization Theory, generally, the Riemannian metric is given (e.g., induced
by Euclidean ones). The fundamental theorem of Riemannian manifolds shows the
one-to-one correspondence between the Riemannian metrics and the symmetric 3-
dimensional matrices r (e.g., Spivak, 1979):

Theorem 3.1 Let G be a symmetric matrix function defining the metric on a Rie-
mannian manifold M in any system of local coordinates. Then, there exists a unique
symmetric connection such that

g"'2)(G- 1 )'a o
ax, + ag'2
r'a1,'2 = ~1/2(ag"O ax, aaxo
0 _
~ (3.5)
0'=1 2 1

for all 11 ,12,13 = 1, ... , n.

Here, the coefficient functions r::",
11,12, h = 1, ... , n uniquely determined by
the Riemannian metric are called the second Christoffel symbols.
If the Riemannian metric is Euclidean, then r::'2 = 0, 1 ,1
1 2, h = 1, ... , n.

Definition 3.4 A second-order covariant tensor is positive semidefinite at a point

m E M if the corresponding matrix is positive semidefinite on TMm x TMm in any
coordinate representation. A second-order covariant tensor field is positive semidefinite
on A ~ M if it is positive semidefinite at every point of A.

4. TENSOR APPROXIMATIONS OF SMOOTH FUNCTIONS ON

RIEMANNIAN MANIFOLDS

We have seen above that NLCS (2.1) and (2.3) can be considered 2n O-covariant
tensor field inequalities on M[hIJ and M[h2J, respectively. Let D f and D2 f denote the
first-order and second-order covariant derivatives of an arbitrary smooth function f on
 240
a Riemannian C2 k-manifold M (with respect to the Riemannian metric), respectively.
In order to draw conclusions for the local structure of smooth NLCS, the function f
and the manifold M will be investigated in a neighbourhood of an arbitrary point in
the form of

f(x(u», (4.1)

where f, X; E C2, i = 1, ... ,n and U ~ RR is an open set. It should be remarked that

this form is not unique, moreover, it can be different in any coordinate representation.
The same idea can be used in the case of smooth optimization problems.

Definition 4.1 A curve of M is a geodesic if its tangent is parallel along the curve. A
set A ~ M is geodesic convex if any two points of A are joined by a geodesic belonging
to A ~ M.
The definition of geodesics coincides with the classical one in the case of Rie-
mannian metrics (e.g., Spivak, 1979). Consider an arbitrary coordinate representation
of the manifold M in any geodesic convex neighbourhood of M. Then, every two
points of this neighbourhood can be joined by a unique geodesic given in the form of
x( u( s», s E [0, bl, where this function is twice continuously differentiable, s means the
arc length parameter and b is the length of the geodesic between the two points.

Theorem 4.1 In any geodesic convex coordinate neighbourhood of M and for every
geodesic x( u(s» joining two arbitrary points x( u(b» and x( uo) = x( u(O», s E [0, bl,

where e fulfills the equality lim.....o e(O; s)/ s2 = 0 and u~ is the tangent of the geodesic
at uo. IT x(uo) is a stationary point, then

1
f(x(u(s») = f(x(uo» + '2(u~fD2 f(x(uO»u~s2 + e(O;s). (4.3)

Proof. By using Taylor's expansion for the single variable function

f( x( u( s))), s E [0, bl we have that

f(x(u(s))) = f(x(uo» + f~(x(uo»Jx(uo)u~s+

1
+'2( u~)T (J x( uo)T f:(x( uo»J x( uo) + f~(x( Uo »x"( uo)] u~S2+ (4.4)
f~( x( uo»J x( Uo )u~ s2 + e(O; s),
where x"(uo) is an n x k x k matrix with

f '"'( X (Uo »"( ~ af(x(uo»

x Uo ) = L...J ax' x"(u
; 0) (4.5)
;=1 I
 241

and for which c fulfills the equality lims_o c(O; s)/ S2 = o.

As the curve x( u( s)), s E [0, bJ is a geodesic, we can substitute the following system
of differential equations for u~ in (4.4):

u"(s) = -u'(sfr(u)u'(s), (4.6)

where the k X k x k matrix r( u) contains the second Christoffel symbols, u'( s), s E [0, bJ
are tangent vectors and (4.6) means that ui'(s) = -u'(s)Tri(u)u'(s), i = 1, ... , k.
Thus,

f(x(u(s))) = f(x(uo)) + f~(x(uo))Jx(uo)u~s+

1
+2(u~f[Jx(uof f~'(x(uo))Jx(uo) + f~(x(uo))x"(uo)- (4.7)
f~(x( uo))J x( uo)r( uo)] u~s2 + c(O; s),
where the multiplication of a 2-dimensional and a 3-dimensional matrices is defined
by the rule related to the multiplication of a row vector and a 3-dimensional matrix,
applied consecutively for every row vector of the 2-dimensional matrix. Taking into
consideration that Df(x(uo)) = f~(x(uo))Jx(uo) and

D2 f(x(uo)) = Jx(uo)T f~(x(uo))Jx(uo) + f~(x(uo))x"(uo) - f~(x(uo))Jx(uo)r(uo),

statement (4.2) is obtained. At a stationary point Uo the first covariant derivative

Df(x(uo)) = f~(x(uo))Jx(uo) = 0, so

f(x( u(s))) = f(x( uo)) + ~(u~f D2 f(x( Uo ))u~s2 + c(O; s). (4.8)
o

The equalities in (2.1) and (2.3) identify the (n - I)-dimensional Riemannian

manifolds M[hlJ and M[h2J. The second-order covariant derivative can be formulated
by using the property that M[hlJ and M[h2J are immersed submanifolds of R n with
the Riemannian metrics induced by the Euclidean one of R n.

Theorem 4.2 By setting 'f/i = f'h?[h:h?J- 1 , i = 1,2,

D2 f = [I" - Jiihi']ITM[h;J' i = 1,2, (4.9)

where IT M[ hi] as index denotes restriction to the tangent spaces T M[ hi], i = 1,2.

Proof. The space R n should be considered at every point of M[h 1 ] or M[h2] the direct
sum of the tangent space and the normal space with respect to the Euclidean metric
spanned by the column vectors of J Xi(U), i = 1,2 and h:(Xi(U))T, i = 1,2, respectively.
Thus, it is possible to decompose any vector of R n at every point of M[hlJ or M[h2J
242
into a tangential and a normal component. By using these notations for the column
vectors of the n x (n - 1) x (n - 1) matrices x:' (uo), i = 1,2, the well-known Gauss
equation (e.g., Spivak, 1979) can be formulated as follows:

x~'(u) = Jx;(u)f;(u) + h:(x;(u)fB;(u), u E U ~ Rk, i = 1,2, (4.10)

where f; are the (n - 1) x (n - 1) x (n - 1) matrices of the second Christoffel symbols of

M[h;), i.e., the tangential components of the column vectors of the n X (n -1) x (n -1)
matrices x:', and B; are the 1 X (n - 1) X (n - 1) matrices consisting of the normal
components of the column vectors of the 3-dimensional matrices x:'.
The multiplication
of a vector and a 3-d~mensional matrix is defined in (4.5). The multiplication of a
2-dimensional and a 3-dimensional matrices is defined in (4.7). Now, calculate the
second-order covariant derivative:

D2 f(x;(u)) =
Jx;(u)T f~(x;(U))Jxi(U) + f~(Xi(U))X:'(u) - f~(x;(u))Jx;(u)fi(U) =
= Jx;(u)T f~(x;(u))Jx;(u) + f~(x;(u))(x:'(u) - JXi(u)f;(u)) =
by Gauss equation,
(4.11)
= Jx;(uf f~(Xi( u))J x;(u) + f~(x;(u))h:(Xi(U))T Bi(U) =
= Jx;(u)T f~(x;(U))Jxi(U) + (f~(Xi(U))T + f~(X;(U))N)h:(x(u))TBi(U) =
= JXi(uf f~(Xi(U))Jxi(U) + f~(Xi( u))Nh:(Xi(U))T B;( u),
uEU~Rn-l, i=1,2.

As
h;(x;(u)) = 0, u E U ~ Rn - l , i = 1,2 (4.12)

and differentiating (4.12) twice by u, we have

i = 1,2. (4.13)

Multiplying both sides of equations (4.13), respectively, by 7)1(XI(U)), 7)2(X2(U)) and

taking into consideration (4.10), we obtain

i = 1,2 (4.14)

from which the statement follows. o

Theorems 4.1 and 4.2 have some consequences in nonlinear optimization:
(1) In a neighbourhood of a stationary point, the composite function (4.1) can be
approximated by a second-order covariant tensor.
(2) In the case of a local optimum, this tensor is necessarily a positive semidefinite
matrix in any coordinate representation, i.e., a positive semidefinite tensor. If it
 243

is a positive definite tensor and the first order condition is fulfilled, then the strict
local optimality is guaranted in a neighbourhood.
(3) If we can extend continuously the positive second-order covariant derivative at a
stationary point (second-order optimality conditions) for a coordinate neighbour-
hood of the manifold as a second-order positive definite covariant tensor field, then
a Riemannian metric can be introduced there. In nonlinear optimization, by using
first-order and/or second-order information almost every method is based on an
approximation of such a Riemannian metric. The approximations of the Hessian
matrix of the Lagrangian function in a neighbourhood of a local optimum point
(e.g., Newton and Newton-like methods, continuous methods of second-order de-
gree, sequential quadratic programming methods, interior point methods, reduced
gradient methods, penalty function methods, etc.), in general, are not tensor ap-
proximations. The main idea of methods is to choose a decreasing direction v at
any iteration point such that the inner product of a projection of /' to the tangent
space of the constraint manifold and v should be minimized in the Riemannian
metric generated by this tensor field (Rapcsak and Thang, 1994).
Now, the second-order tensor approximations of functions g(x) and x in NLCS
(2.1) will be determined on the basis of formulas (4.2) and (4.9). These statements
are true in the case of any coordinate representation of the complementarity manifold
M[h l ], so we choose a convenient one. Let us fix a point Xo and since h~(xo) -I 0,
suppose that the nth component is not zero. Let

i = 1, ... ,n - 1, U = (UI,' .. ,un-d E U c lRn-l, (4.15)

i.e., the first (n -1) coordinates give the parameters of the complementarity manifold
in a neigborhood of Xo. By applying the implicit function theorem, we obtain the
complementarity manifold M[hd = {x E Rnlhl(x) = O} in a neighbourhood of Xo as
follows:

x(U) = ( 4.16)

where Xn is a C 2 function. In general, we do not know this surface explicitly, but

we can compute the main characteristics, i.e., the matrices of the first and second
fundamental forms in a neighbourhood by using the gradient vector and the Hessian
matrix of the function hI. The n X (n -1) Jacobian matrix
244

o
1 o
o ) ,
1 (4.17)
~i!!.!:l
- 8U n _l / aXn

(UI,U2, ... ,Un-I) E U eRn-I.

The second-order tensor approximations of functions g( x) along a geodesic
x( u( s)), s E [0, bJ can be determined by using formula (4.2) as follows:

gj(x(U(S))) ~ gj(x(uo)) + gj(x(uo))Jx(uo)u~s+

( 4.18)
~(u~)T[gj(x(uo)) - J.lj(x(uo))h~(X(UOmTM[h11U~S2, j = 1, ... ,n.
In the case of functions x, we have

Xi(U(S)) ~ Xi(UO) + u:os - ~(u~)T[h~(x(uo));/lh~(x(uo))12h~(x(uO))JITM[hdU~S2,

i = 1, ... ,n,
(4.19)
where h~ (x( UO))i, i = 1, ... ,n is the ith component of the vector h~ (x( uo).
We remark that [h~1lTM[hl1 is equivalent to the matrix of the second fundamental
form of M[hd given explicitly, e.g., in Rapcsak (1994b). The same technique can be
applicable for NLCS (2.3) by using the Lagrange multipliers Vi and "'Ii, i = 1, ... , n
defined in (2.4).
It follows from tensor approximations (4.18) and (4.19) that a system of 2n in-
equalities defined in an open domain of R n - I with smooth functions can be considered
instead of NLCS (2.1) in a neighborhood of Xo. The efficiency of optimization methods
may depend on coordinate representations. The technique, how to change linear or
nonlinear coordinate representations in optimization theory, is applied in simplex and
reduced gradient methods.

5. AN OPTIMIZATION APPROACH

The history of algorithms for solving the finite-dimensional VI and NLCS is rela-
tively short. From the beginning it has been recognized that NLCS are special cases
of VI, still the research directions were different: VI have been considered in infinite-
dimensional metric spaces and NLCS in finite-dimensional Euclidean spaces. In the
early years of study on CS most effort was devoted to LCS due mainly to the applica-
bility of Lemke-type algorithms. Cottle (1966) developed the first method for NLCS
by extending the principal pivoting algorithms for LCS. This idea led to the class of
methods referred to as fixed-point and homotopy algorithms which are not employed
 245

in the case of VI. Mangasarian (1976) transformed NLCS into a system of nonlinear
equations to put classical algorithms into forefront which seems to be a promising
approach. Nonsmooth equation formulations of NLCS led recently to a robust itera-
tive algorithm published by Pang and Gabriel (1993). The method NE/SQP solves
a sequence of nonnegatively constrained quadratic programs of the least-squares type
based on a robust SQP algoritm. By tensor approximations, a new smooth robust SQP
type approach may be introduced. The reformulation of NLCS (or VI) as optimization
problems has also the advantage to use classical methods for solving NLCS. In the case
of (2.1), the usual form is

minxTg(x) s.t. g(x)~O, x~O, xER n , (5.1)

where the solution point x· must meet the requirement of g(x*)T x* = O. In general, the
feasible region of (5.1) is not convex, but in the case of LCS, problem (5.1) is quadratic.
To solve the optimization problem, among other things, the direct use of software
MINOS, cutting plane based and continuation methods were suggested. Another idea
based on gap functions is to cast NLCS as an unconstrained minimization of a smooth
function derived from (5.1). Fukushima (1992) formulated asymmetric Variational
Inequality problems as differentiable optimization problems which are involved in an
implicit Lagrangian function approach published by Mangasarian and Solodov (1993).
The most general framework for gap functions seems to be introduced by Giannessi
(1994).
The fixed-point and optimization approaches have advantages and disadvantages,
namely, they lack either the computational efficiency necessary for solving large-scale
equilibrium problems or the generality. A detailed description of results on finite-
dimensional NLCS and VI can be read in a state-of-the-art paper by Harker and Pang
(1990).
A general approach for solving NLCS, VI, optimization problems and systems of
nonlinear equalities consists of linear or nonlinear approximations of problem func-
tions. Our approach emphasizes that tensor approximations invariant under nonlinear
coordinate transformations (representations) may be essential. In this part, NLCS will
be reformulated as optimization problems defined on a Riemannian manifold, which
ensures a convenient structure to study theoretical and numerical questions.
First, see NLCS (2.1) which can be considered 2n O-covariant tensor field in-
equalities on M[hlJ. A well-known step to reformulate NLCS (2.1) into optimization
problems different from (5.1) is as follows:

min-t s.t. g(x)-t~O, x-t~O, xEM[hlJ, tEIR, (5.2)

where 9 : R n -+ R n is a twice continuously differentiable function and M[hlJ is the

C2 (n - I)-dimensional Riemannian manifold defined above. This is a smooth opti-
mization problem over a differentiable manifold studied in optimization theory, e.g.,
by Gabay (1982); Rapcsak and Thang (1994). If instead of 2n inequality constraints,
246

only one is formed in (5.2) by using the operation' max' for all the functions, then we
lose the property of smoothness, but efficient techniques of global optimization may be
applied. To ensure the geodesic convexity property of the feasible region of (5.2), some
additional requirements are needed (Rapcsak, 1994a,1994b,1994c), thus to preserve
a more general and flexible treatment of the nonconvex character, a penalty func-
tion approach is introduced. To solve unconstrained optimization problems a general
framework for globally convergent variable metric methods along geodesics can be in-
troduced, containing a generalization of the steepest descent, Newton, Quasi-Newton,
gradient projection, reduced gradient, projected gradient, SQP methods, etc., for the
optimization problem given in the form of

minf(x,w) s.t. x E M[hd, wEn ~ R', (5.3)

where f(x,w) means a suitable penalty function with 1 penalty parameters for opti-
mization problem (5.2). A general variable metric method along geodesics for solving
(5.3) in the case of a fix value wEn proceeds as follows:
Starting from an initial feasible solution Xo, let Xk be the feasible solution of the
kth iteration step, Wk the penalty parameters at the kth iteration step, Dk an n x n
matrix which is positive definite on TM[hlJ",. (the tangent space of M[hlJ at Xk)
defining a linear map Dk : TMx. -+ TMx., G 1 the Riemannian metric of R.n , G the
induced Riemannian metric of M[h l ]; furthermore, Dk and the metric G commute at
the iteration points.
Step 1. Compute the direction Pk as follows: Pk = Dkfb(Xk,Wk)T, where
fb(Xk,Wk)T is the projected gradient of f with respect to the Riemannian metric
G 1 to the tangent space.
Step 2. Let Xk+! = f",.(tk,Pk), where f",.(tk,Pk) is the arc ofthe geodesic starting
from Xk with tangent Pk, and the stepsize tk is determined by an exact geodesic search

(5.4)

In numerical realizations, tensor approximations (4.18) and (4.19) can be utilized

by setting u' = Pk at the iteration points, and the Armijo principle for determining
the stepsize.
The global convergence of this class of methods was proved under more general
assumptions in Rapcsak and Thang (1994). Let Wk denote the connected component
containing the Xk of the level set {x E M[hdlf(x) ::; f(Xk)}.

Theorem 5.1 If f is continuously differentiable, Wo is compact and geodesic convex,

the sequence {xd is generated by the above algorithms, and the linear mapping DkXk
satisfies the Lipschitz condition at the iteration points, then, the sequence {x k} is
either finite terminating at a critical point, or infinite, and every accumulation point
is critical.
 247

By the assumption of Theorem 5.1, I is a continuously differentiable function. In

penalty function approach, there is a continuously differentiable subclass for inequality
constraints (e.g., Bertsekas, 1982), where the convergence analysis is developed for the
convex case. A general image problem approach to penalty methods was published
by Pappalardo (1990) based on the relationships between the weak and strong separa-
tion functions and the penalty methods discussed by Giannessi (1984). The following
penalty function with two parameters suggested by Forgo (1969) is a modification of
the exponential one:

I(x, a", b,,) = -t + lib" L e-a~(gi(z)-t) + lib" L

n n
e-a.(Zi- t ),
;=1 ;=1
lim
" .... ex>
a" = +00, 0 < b" < bk+I, lim b" =
" .... ex>
+00, (5.5)

lim a"/log(b,,) = +00.

" .... ex>

It is also a weak separation function for every pair of parameters. In the case of
convex optimization problems with unbounded feasible region, nonempty and bounded
solution set, the algorithm converges to an optimal solution, and moreover, if the
interior of the feasible domain is nonempty, then a pair of parameters can be selected
such that an f.-approximate solution is reached after one global minimization (Rapcsak,
1975). It follows from the proof that if the global minimums of this penalty function
exist for every pair of parameters, then they converge to a solution of problem (5.2).
By extending the f.-exactness property to our case, the combination of this result and
Theorem 5.1 should result in a globally convergent algorithm for determining an f-
approximate solution of problem (5.2).

6. CONCLUDING REMARKS

In the paper, the structure of smooth NLCS and a class of penalty methods are
studied by using tensors. Some advantages of this approach are as follows:
(1) an attempt for a unified and global description of theoretical and algorithmical
results of NLCS,
(2) the independence of the theoretical results from the imbed dings of the comple-
mentarity manifold and the representations of the Riemannian metrics and
(3) a challange to use a new view in NLCS.

ACKNOWLEDGEMENTS. The present research has been supported in part by

the Hungarian National Research Foundation, Grant No. OTKA-2568 and the Project
"TRASPORTI" of Italian National Research Council (CNR).
248
REFERENCES

[1] D.P. Bertsekas, "Constrained optimization and Lagrange multiplier methods", Academic Press,
New York, London, 1982.
[2] R.W. Cottle, "Nonlinear programs with positively bounded Jacobians", SIAM Journal on Ap-
plied Mathematics 14, 147-158, 1966.
[3] F. Forgo, "A method for solving nonlinear programming problems approximately", Szigma 1,
67-75, 1969. (in Hungarian)
[4] M. Fukushima, "Equivalent differentiable optimization problems and descent methods for asym-
metric Variational Inequalities", Mathematical Programming 53, 99-110, 1992.
[5] D. Gabay, "Minimizing a differentiable function over a differentiable manifold", Journal of Op-
timization Th.eory and Applications 37,177-219,1982.
[6] F. Giannessi, "Theorems of the alternative and optimality conditions" , Journal of Optimization
Theory and Applications 42, 331-365, 1984.
[7] F. Giannessi, "Separation of sets and gap functions for Quasi-Variational Inequalities". This
Volume.
[8] P.T. Harker, and J.S. Pang, "Finite-dimensional Variational Inequality and nonlinear comple-
mentarity problems: a survey of theory, algorithms and applications", Mathematical Program-
ming 48, 161-220, 1990.
[9] O.L. Mangasarian, "Equivalence of the complementarity problem to a system of nonlinear equa-
tions", SIAM Journal on Applied Mathematics 31, 89-92, 1976.
[10] O.L. Mangasarian, and M.V. Solodov, "Nonlinear complementarity as unconstrained and con-
strained minimization", Mathematical Programming 62, 277-297, 1993.
[11] J.S. Pang, and S.A. Gabriel, NE/SQP: "A robust algorithm for the nonlinear complementarity
problem", Mathematical Programming 60,295-337, 1993.
[12] M. Pappalardo, "Image problem approach to penalty methods" , Journal of Optimization Theory
and Applications 64, 141-152, 1990.
[13] T. Rapcsak, "An exterior point algorithms for solving convex nonlinear programming problems" ,
Alkalmazott Matematikai Lapok 1, 357-364, 1975. (in Hungarian)
[14] T. Rapcsak, "Minimum problems on differentiable manifolds", Optimization 20, 3-13, 1989.
[15] T. Rapcsak, and J. Szenthe, "On the connection between mechanical force equilibrium and
nonlinear programming", ZAMM 70,557-564, 1990.
[16] T. Rapcsak, "Geodesic convexity in nonlinear programming", Journal of Optimization Theory
and Applications 69, 169-183, 1991.
[17] T. Rapcsak, and T. Csendes, "Nonlinear coordinate transformation for unconstrained optimiza-
tion, II. Theoretical background", Journal of Global Optimization 3, 359-375, 1993.
[18] T. Rapcsak, "On the connectedness of the solution set to linear complementarity systems",
Journal of Optimization Theory and Applications 80, 501-512, 1994a.
[19] T. Rapcsak, "On the connectedness of the solution set to nonlinear complementarity systems",
Journal of Optimization Theory and Applications 81, 619-631, 1994b.
[20] T. Rapcsak, "Geodesic convexity on Rn", in: "Generalized convexity", Lecture Notes in Eco-
nomics and Mathematical Systems 405, Koml6si, S., Rapcsak, T. and Schaible, S. (eds.),
Springer-Verlag, Berlin, Heidelberg, New York, 91-103, 1994c.
[21) T. Rapc:Ak, _d T.T. Thang, "On nonlineu coordinate representations of smooth optimization
problems", J_raal oj Optimization Tlt.tw'll and Appliccatioru. To appear.
[22) M. Spivak, A comprehensive introduction to differential geometry I-V., Publish or Perish,
IDe. Berkeley, 1979.
BASIC EXISTENCE THEOREMS FOR GENERALIZED VARIATIONAL
AND QUASI-VARIATIONAL INEQUALITIES

Ricceri Biagio

Department of Mathematics
University of Catania
Viale A.Doria 6
95125 Catania, Italy
e-mail:ricceri@dipmat.unict.it

Abstract. We deal with the following problem: given a Hausdorff real topological
vector space E, a closed convex set X ~ E, a multifunction F : X --+ 2x and a
multifunction ~ : X --+ 2E ' , find (x, 'P) E X x E* in such a way that

x E F(x), 'P E ~(x), sup('P,x - y) =5 o.

yEX

Our aim is to report those results on the subject in which the basic assumption on ~
is the following: for each y E X - X, the set

{x EX: inf ('P,Y) =5 o}

opE+(x)

is closed. Particular emphasis is put on the case where F is constant as well as on the
(difficult) open problems.

Key Words. Variational Inequalities, Quasi-Variational Inequalities.

We deal with the following problem: given a real Hausdorff topological vector
space E, a closed convex set X ~ E, a multifunction F : X --+ 2x and a multifunction
~: X --+ 2E ', find (x,'P) E X x E* in such a way that

xEF(x), 5pt'PEI}(x), sup ('P,x-y) =50.

YEF(x)

This problem, as one says, is the generalized Quasi- Variational Inequality associated
with X, F,~. We denote it by GQVI(X, F, ~). It was first introduced, in a finite-
dimensional setting, by Chan and Pang in [3]. When F(x) = X (resp. I}(x) is a

251
252

singleton) for all x E X, our problem reduces to the generalized Variational Inequality
associated with X, ~ (GVI(X, ~» (resp. Qu~i- Variational Inequality associated to
X,F,~ (QVI(X,F,~))). Finally, when, at the same time, F(x) = X and ~(x) is a
singleton for all x E X, we have the classical Variational Inequality associated with
(X,~) (VI(X,~».

Most papers devoted to our problem (in its various specifications) involve some
suitable notion of monotonicity for the operator~. The work done in that direction
is mainly related to contributions by Browder, Brezis, Hess. There are, however, some
writings where the conditions imposed to ~ concern merely the kind of its continuity.
Here we deal with this latter category of papers.
To focus better the essential things in the development of the theory as well as in
the problems still open, I start with the case of Variational Inequalities.
It is well known that one of the basic results is Hartman-Stampacchia's theorem,
stated below (in a more general form) as Theorem 1.
Theorem 1 ([61, Lemma 3.1).- Assume that X is compact and finite-dimensional, and
that ~ is single-valued and weakly-star continuous. Then, VI(X, ~) has some solution.
A question naturally arises: what happens when dim(X) = 00 ? At that point,
the situation becomes definitively more composite and delicate. Continue to suppose
that X is compact and that ~ is single-valued and weakly-star continuous. Denote by
U the family of all finite-dimensional linear subspaces of E meeting X. Consider U as
a directed set, with the set-theoretic inclusion. For each S E U, thanks to Theorem 1,
we get a point x sEX n S such that

sup (~(xs),xs - y) ::;: O.

IIEXns

Since X is compact, the net {XS}SEU admits some cluster point in X, say Xo. How
can we show that Xo is a solution of VI(X, ~) ? To this end, let {x .. } be a sub-net of
{XS}SEU converging to Xo. Fix y E X. Then, from the obvious equality

(~(x .. ),x .. - y) = (~(x .. ),x .. - xo) + (~(x .. ),xo - y)

we readily see two ways to achieve our goal. Indeed, since

limsup(~(x .. ),x .. - y) ::;: 0

01

and

..
lim(~(x .. ),xo - y) = (~(xo),xo - y) ,
it is enough to know that
lim(~(x .. ),x .. - xo)
01
= O.
To get this latter information, there are just two natural ways. Namely, either we
suppose more on the kind of continuity of ~, requiring that it is continuous with
respect to the topology, on E*, of uniform convergence on the compact subsets of E
 253

([1), Theorem 6), or we suppose more on the topology of E, requiring that the weakly-
star compact subsets of E* are equicontinuous ([2), Proposition 15). This latter fact
happens, for instance, if E is a Baire space.
When we are out of the two above-mentioned circumstances, Theorem 1, in gen-
eral, is no longer true without the condition dim(X) < 00. In this connection, we have
the following very interesting result by M.Frasca and A.Villani:

Theorem 2 ([5], Theorem 2). - Let (E, h,) be any infinite-dimensional Hilbert
space. Then, for each closed ball X in E, there exists some {strongly} continuous
affine operator ~ : E -+ E such that, for every x EX, one has

sup(~(x),x - y) > O.
,EX

It is clear how Theorem 2 serves to our purpose: we consider E endowed with the
weak topology and identify E* with E. Thus, X is weakly compact and, at the same
time, ~ is weakly continuous, being affine and strongly continuous.

On the basis of what we have now seen, the following question arises: is there some
general case in which VI(X, ~) has solutions, though ~ is only weak-star continuous
and, at the same time, no restriction is made on the topology of E? An answer to
this question is provided, for instance, by the following result of mine:

Theorem 3 ([8]) - Assume that the relative interior of X {that is the interior of X in
its affine hull} is non-empty, and that ~ is weakly-star continuous. Moreorer, let K, Kl
be two non-empty compact subsets of X, with Kl ~ K and Kl finite-dimensional, such
that, for each x EX \ K, one has

sup (~(x),x - y) > O.

IIEKI

Then, VJ{X, ~} has some solution lying in K.

Here is a first open problem.

Problem 1. - In Theorem 3, can we drop the finite-dimensionality assumption on the

set Kl ?

Actually, Theorem 3 is a very particular case of the following more general result
which is also the basis of our further considerations.

Theorem 4 ([8], Theoreme 1). - Let X,K,K1 be as in Theorem 9, and let ~(x) be
convex and wealcly-star compact for all x E X. Moreover, assume that:
(a) for each y E X -X, the set {x EX: inft;>E+(x){<P,y) ::; O} is compactly closed (that
is, its intersection with any compact subset of E is closed);
(b) for each x EX \ K, one has

sup inf (<p, x - y) > O.

IIEKI t;>E+(x)
254

Then, GVI(X, eli) has some solution lying in K X E*.

For the sake of simplicity, from now on, we denote by VeX, E*) the class of all
multifunctions III : X ~ 2Eo, with non-empty, convex, weakly-star compact values,
satisfying condition (a) of Theorem 4, that is to say such that for each y E X - X, the
set {x EX: infcpEot(x)(!f', y) ~ O} is compactly closed.
It is easy to realize that any weakly-star upper semi continuous multifunction III :
X ~ 2Eo, with non-empty, convex, weakly-star compact values, belongs to the class
V(X,E*).
Passing now to generalized Quasi-Variational Inequalities, let us recall the basic
existence result of D.Chan and J.S.Pang.
Theorem 5 ([3], Corollary 3.1). - Assume that: dim(E) < 00, X is compact, F is
continuous, with compact convex values, and eli is upper semicontinuous, with compact
convex values. Then, GQVI(X,F,eli) has some solution.
There is an infinite-dimensional extension of Theorem 5 which is perfectly analo-
gous to the already mentioned Browder extension of Hartman-Stampacchia's theorem.
Precisely, we have the following result by W.K.Kim.
Theorem 6 ([7]). - Assume that: E is locally convex, X is compact, F is continuous,
with compact convex values, and eli is strongly upper semicontinuous, with strongly
compact convex values. Then, GQVI(X,eli) has some solution.
Very recently, Theorem 5 has been improved, in several directions, by P.Cubiotti.
His result is as follows.
Theorem 7 ([4], Theorem 1). - Assume that: dim(E) < 00, F is lower semicon-
tinuous, with closed graph and convex values, and eli belongs to the class VeX, E*).
Moreover, assume that there is some non-empty compact set K ~ X such that, for
each x EX \ K, with x E F(x), one has

sup inf (!f',x-y) >0.

!/EF(x)nK cpEot(x)

Then, GQVI(X,F,eli) has some solution lying in K x E*

Observe that Theorem 7 is also the extension to the quasi-variational case of the
finite-dimensional version of Theorem 4.
Just a few time ago, N.D.Yen re-obtained Cubiotti's result by means of a com-
pletely different proof based on a beautiful and delicate characterization of the class
V(X,E*) (always when dim(E) < 00), established by Yen himself. Here is such a
characterization.
Theorem 8 ([9], Theorem 2.2). - Let E = RR. Denote by C the orthogonal comple-
ment of span( X - X) and, for each z ERR, by 71"( z) the orthogonal projection of z on
span(X - X).Let III : X ~ 2Rn a multifunction with, non-emtpy, compact, convex val-
ues. Then, III belongs to the class V( X, R R) if and only if the following two conditions
are satisfied:
 255

(a) the 8et D:= {x EX: w(x) nc t= 0} is closed;

(f3) the multifunction G : X \ D -t 2Rn defined by putting
1T(Z)
G(x) = {II 1T(Z) II : Z E w(x)}
for all x E X \ D, is upper semicontinuous.
At this point, we can conclude pointing out the two other main open problems in
the theory we are considering.
Problem 2. - Can Theorem 4 be extended to the quasi-variational case along the
same lines as that of Theorem 7 ? In other words, one has to see whether the following
statement is true.
Conjecture. - Assume that: «l> belongs to the class V(X, E*), F is lower semicontinu-
ous, with closed graph and, for each x EX, F( x) is a convex set whose interior in the
affine hull of X is non-empty. Moreover, let K, Kl be two non-empty compact subsets
of X, with Kl ~ K and Kl finite-dimensional, such that, for each x EX \ K, with
x E F(x), one has
sup inf (r.p, x - y) > O.
yEF(x)nK, <pE9(X)

Then, GQVI(X, F, «l» has some solution lying in K x E*.

Problem 3. - Find a characterization of the class V(X, E*), without the restriction
dim(E) < 00. In particular, can Theorem 8 be extended to an arbitrary real Hilbert
space?

REFERENCES

[1] F.E. Browder, "The fixed point theory of multi-valued mappings in topological vector spaces",
Math. Ann., 177, 283-301, 1968.
[2] F.E. Browder and P.HESS, "Nonlinear mappings of monotone type in Banach spaces", J. Fund.
Anal., 11, 251-294, 1972.
[3] D. Chan and l.S.PANG, "The generalized quasi-variational inequality problem", Math. Oper.
Res., 7, 211-222, 1982.
[4] P. Cubiotti, "An existence theorem for generalized quasi-variational inequalities", Set- Valued
Anal., 1, 81-87, 1993.
[5] M. Frasca and A.Villani, "A property of infinite-dimensional Hilbert spaces", J. Math. Anal.
Appl., 139, 352-361, 1989.
[6] P. Hartman and G. Stampacchia, "On some nonlinear elliptic differential equations" , Acta Math.,
115, 161-220, 1966.
[7] W.K. Kim, "Remark on a generalized quasi-variational inequality", Proc. Amer. Math. Soc.,
103, 667-668, 1988.
[8] B. Ricceri, "Un theoreme d'existence pour les inequations variationnelles", C.R. Acad. Sci.
Paris, Serie 1,301, 885-888, 1985.
[9] N.D. Yen, "On an existence theorem for generalized quasi-variational inequalities", Set- Valued
Anal., to appear.
SENSITIVITY ANALYSIS OF VARIATIONAL INEQUALITIES
BY NORMAL-MAP TECHNIQUES

Robinson Stephen M.

Department of Industrial Engineering

University of Wisconsin-Madison
1513 University Avenue
Madison, WI 53706-1572, USA
Email: smr@cs.wisc.edu

Abstract: This paper deals with the sensitivity analysis of solutions of variational
inequalities by an implicit-function approach that makes use of so-called normal maps.
Although some of the results can be extended to infinite-dimensional spaces, the treat-
ment here is restricted to variational inequalities over polyhedral convex sets in finite-
dimensional spaces, the case for which the strongest results can be established. Cov-
erage includes transformation of the variational inequality to the equivalent form of a
normal map, basic facts about normal maps, derivatives, conditions for nonsingularity,
the implicit-function theorem, and applications to sensitivity analysis. Some examples
are included to illustrate the techniques.

Key Words: Variational inequality, normal map, sensitivity analysis, implicit-function

theorem

1. INTRODUCTION

This is an expository paper about sensitivity analysis of variational inequalities: that

is, examining how solutions of such problems change when the data of the problem are
changed. The main tool for this analysis is an implicit-function approach that makes use
of so-called normal maps, which transform the variational inequality into a nonsmooth
equation. By extending the implicit-function theorem to this class of nonsmooth equa-
tions, one can then obtain information about the existence and behavior of solutions as
the problem data change.
In this paper we restrict ourselves to the case in which the function appearing in the
variational inequality has enough smoothness for what we want to do; it is enough for
the function to be strongly Fnkhet differentiable in the variables of interest. The main
reason for this is that in the differentiable case the theory has settled into a fairly stable
state. However, there is much current interest in pursuing these questions for cases in

257
258

which the functions are not differentiable, and this is an active research area. For a
sample of the kinds of results that can be obtained in that case we refer to [6].
The rest of this paper is organized in three main sections. The next section introduces
variational inequalities and normal maps, and shows how either one may be transformed
into the other. Then we examine the normal manifold associated with the projector
onto a polyhedral convex set C. This normal manifold plays an important part in the
analysis of normal maps, as well as in computations involving them. In that section we
also look at affine maps and ask when the normal maps generated from them will have
certain convenient nonsingularity prope!ties. In the following section, we exploit this
nonsingularity to obtain the basic implicit-function theorem, and we show how to apply
this theorem to do sensitivity analysis. Following these main sections we provide a brief
conclusion, acknowledgments of support, and references.
Throughout the paper we use without comment terms and constructs of convex anal-
ysis. An excellent reference for these is the fundamental book of Rockafellar [15].

2. VARIATIONAL INEQUALITIES AND NORMAL MAPS

In this section we introduce variational inequalities and normal maps, and we show that
these are essentially equivalent devices, in that either one may be transformed into the
other. Variational inequalities are more familiar from applications, whereas the normal
map formalism will be more useful for the analysis in this paper.
We begin with the parametric variational inequality

Find x(y) E Ire such that for each x E C, (x - x(y),J(x(y),y)) ~ O. (2.1)

Here C is a nonempty polyhedral convex set in IR" and f is a function from a product
(C n fl) X e to IRn;e is an open subset of a normed linear space P and fl is an open
subset of IRn with C n fl f. 0. For many applications P = IRk, but in general it is
convenient to let it be a normed linear space because the arguments go through and
because one sometimes needs the extra generality.
Of course, y might be fixed, in which case one has an ordinary variational inequality.
However, the setup in (2.1) permits us to change the form of f and then to observe
the effect of such changes on the solution point. Sensitivity analysis of this kind is the
focus of the present paper. We point out that in general there is no reason for C to be
polyhedral, but that we make that assumption here because it covers the cases in which
we are interested, and it makes possible much stronger results than would otherwise be
available.
The problem (2.1) includes a great many special cases, including linear and nonlinear
programming (the latter through the device of the Kojima function; see the discussion in
[13]), and various kinds of equilibrium problems including those posed over networks. It is
appropriate to note that the work of Stampacchia was decisive in the initial development
of this subject: the first two existence theorems for variational inequalities seem to have
been proved in [18] and [5] respectively, and an exposition of the state of the theory at
an early point in its development is in [19]. We do not go into detail here regarding the
various problems to which the formalism (2.1) can be applied, but we refer instead to
[4], in which numerous references are given.
 259

As mentioned above, another formulation of the problem will be more useful for
analysis. Suppose that we denote by TIc the Euclidean projector on C, and for the
moment we consider a function f : C n n --+ R n , where n is as above (that is, we
temporarily suppress the variable y). Define a new function fe from TIc1(n) to lR,n by
Jc(z) = f(IIc(z» + (z - IIe(z». (2.2)
Thus, to evaluate fe at a point z we first project z onto C to obtain a point x, then
evaluate f at x, and finally add to f(x) the remainder z - x. Given the set C, we
call this function fe the normal map induced by f. The definition implies that for a
function f and a point a E R n one has (f + a)c = fe + a, and that for functions f
and g and a scalar A E [0,1] one has [(1 - A)f + Ag]c = (1 - A)fe + Age.
To understand why this function fe is connected to the Variational Inequality
problem, suppose first that z and x are points of R n such that
fe(z) = 0, x = TIe(z). (2.3)
If we denote by Ne(s) the normal cone of Cat s, defined to be empty if s 1. C and
to be the set of all points c· with (c·, c - s) :S 0 for each c E C otherwise, then as
o = fe(z) we have - f(x) = z - x. The right-hand side is an element of the normal
cone Ne(x), so we see that
z =x - f(x), and 'Vc E C, (f(x),c - x) 2': 0, (2.4)
and therefore x = TIe(z) solves the Variational Inequality problem defined by f and
C.
Conversely, if z and x satisfy (2.4), we obtain z - x = -f(x) E Ne(x), or x =
(I + Ne)-l(z).But (I + Ne)-l is the projector TIc, so we have x = TIe(z). This
implies
0= f(x) + (z - x) = f(TIe(z» + (z - TIe(z» = fe(z),
and therefore z and x satisfy (2.3). Accordingly, we can move back and forth between
(2.4) and (2.3) because they are completely equivalent.

°
Returning to (2.1), we see that to analyze its solutions x(y) we can solve instead
the normal-map equation f(', y)e(z) = to obtain a point z(y), then project z(y) onto
C with TIc to obtain x(y) satisfying (2.1). That is what we shall do in the remainder
of this paper.
An example might help to clarify these ideas. Suppose that we want to minimize
the Frechet differentiable function 1> over the polyhedral convex set C. The first-order
necessary optimality condition is then just (2.4), with f = d1>, the derivative of 1>.
Accordingly, the corresponding normal-map equation is
d1>e(z) = 0,
and if z is a solution of this equation then the corresponding first-order point is x =
TIe(z).
In this section we introduced the Variational Inequality (2.1) and the formalism
of normal maps. We further showed that any Variational Inequality is completely
equivalent to the corresponding normal map, so that we can use the more convenient
single-valued normal map for analysis. In the next section we discuss a geometric
construction associated with normal maps, whose properties are important for the
analysis we shall do later.
 260

3. THE NORMAL MANIFOLD

In this section we study a subdivision of the space R n that is closely associated with
the normal map construction. This subdivision has numerous interesting geometric
properties, but we shall only note a few of these because most are not required for our
purposes. Additional detail, and proofs, can be found in [13].
To begin, recall that the set C is polyhedral, so it has a finite number of faces.
Further, it is easy to show that on the relative interior of each such face, say F, the
normal cone Ne(x) takes a constant value, which we denote by NF. Let us write
(J'F = F + NFj this set (J'F is a polyhedral convex subset of RR having dimension n.
Note that on (J'F the projector lIe coincides with the affine map aF that projects points
onto the affine hull of F, so that on (J'F fe coincides with (f 0 aF) + I - aF, a map
that is smooth if f is smooth, and even affine if f is affine. Therefore fe is piecewise
smooth or piecewise affine respectively, and the (J'F are in general the maximal sets
with respect to which the smoothness or the affine property will hold for fe.
It is easy to show that the union of the (J'F over all faces F of C covers Rnj in fact,
if we consider the slightly smaller sets u' = (riF) + N F , these actually partition the
space Rn. We shall work with the (J'F because they are more convenient to deal with
(being closed)j further, the collection

.Ne = {(J'F I F a face of C}

forms a .nibdivided piecewi&e linear (PLJ manifold in the sense of [2]j this is the normal
manifold of C. We call the sets (J'F the n-cells of .Ne . Note the difference between the
normal manifold.Ne and the normal-cone operator Ne.
If we wish to apply the ordinary implicit-function theorem to a sufficiently differ-
entiable function, an important hypothesis is that an appropriate partial derivative be
nonsingular: this is equivalent to saying that it is a local homeomorphism. It will turn
out later that a precisely analogous condition holds for our piecewise smooth function
fej this condition requires that a certain normal map Ae induced by an affine trans-
formation A be a local homeomorphism. Therefore it is of great importance to have a
suitable condition to ensure this property.
It turns out that the required condition is conceptually simple and easy to state
(these adjectives do not, however, also apply to its proof!). In order to state it, we need
some terminology. First, we say that a piecewise affine map a from an n-dimensional
PL manifold M c R n to R n is coherently oriented if the determinants of the affine
transformations with which a coincides in the n-cells of M all have the same nonzero
sign. Second, for any point z E Rn we define the critical cone of C with respect to z
to be the polyhedral convex cone defined by

Kc(z) = Te(IIe(z» n (z - IIe(z».L,

where Te(x) is the tangent cone of C at x (the polar of the normal cone), and (x·).L
is the orthogonal complement of the linear space spanned by x'.
To see why coherent orientation plays a part in determining whether a piecewise
affine map a is a homeomorphism, note first that if a is to be a homeomorphism it is
clearly necessary that the affine transformations ai with which a coincides in the various
n-cells of M be nonsingular. Now consider two adjacent cells Tl and T2, and let H be
 261

the common (n - I)-face separating them. Let al and a2 be the affine representatives
of a in these two cells. Choose a point Xo in the relative interior of H; by making
translations if necessary we can always assume that Xo = 0, so that H coincides near
the origin with a subspace L, and that al(O) = 0 = a2(0), so that al and a2 are
linear. Let v be any point in T2 but not in L, and let h2 , ••• , h n be a basis for L. By
nonsingularity ale v), al (h 2), . .. ,al(hn ) are linearly independent. Take the following
bases for R n , regarded as, respectively, the domain space and the image space of the
al:

With respect to these bases the matrix of al is the identity. If we define 0', f32, . .. ,f3n
by
n

a2(v) = O'al(v) + "Lf3ial(hi), (3.1)

i=2
and recall that al and a2 must coincide on L, then we see that the matrix of a2 is

Accordingly, the determinants of the matrices representing al and a2 will fail to have
the same sign exactly when 0' < O. In that case, the point Vi = O'V + Li=2 f3ihi belongs
to rl. But then (3.1) shows that

so that a cannot be even locally injective at the origin.

This argument shows why coherent orientation must be a necessary condition for
any piecewise affine map to be a homeomorphism. What is somewhat surprising is
that, in the case of piecewise affine normal maps, coherent orientation is also a sufficient
condition. This is not true for general piecewise affine maps; see e.g. the example in
Section 3 of [7]. This sufficiency is the content of part (a) of the following theorem;
part (b) gives the corresponding condition for a local homeomorphism property.

Theorem 1 Let A be an affine transformation from R n to R n and let C be a nonempty

polyhedral convex subset of R n and z be a point of Rn.
a. Ac is a homeomorphism if and only if it is coherently oriented,
b. Ac is a local homeomorphism at z if and only if AK(z) is coherently oriented.

Proof. Given a subdivided PL manifold M whose cells cover R n (a chamber .'lY.'ltem

in the terminology of [8]), we define the branching number of M to be the smallest
integer b (possibly 00) such that each face having codimension 2 of an n-cell of M is
contained in at most b n-cells. Theorem 5.3 of [8] says that if b ::; 4 then a piecewise
affine map on M is bijective if and only if it is coherently oriented and proper; a proper
map 9 is one for which limllxll_+oo IIg(x)11 = +00. In our case the properness condition
262

is irrelevant, because with finitely many cells in Nc any coherently oriented piecewise
affine map must be proper. Further, Ralph (10) has shown that for any polyhedral
set C, the branching number of Nc does not exceed 4. Therefore our map Ac will be
bijective if and only if it is coherently oriented. But any piecewise affine map on a PL
manifold with finitely many cells is Lipschitzian, so in this case bijectivity is equivalent
to the homeomorphism property that we are after. This proves our first assertion. A
different proof is given in (13). The second assertion is Theorem 5.2 of (13). 0
Although this theorem gives very precise conditions for the homeomorphism prop-
erties to hold, it is not necessarily the case that those conditions are easy to check. In
general, one has to evaluate the affine representatives in all of the cells of the normal
manifold and to check the signs of their determinants, which can be a lot of work. Of
course, sometimes it is possible to use structural information to make this job easier,
but that depends on the particular case at hand.
However, if the matrix of the transformation in question has a certain symmetry
property then the conditions of the theorem can be made much simpler. We now
describe this situation, giving results only for the global homeomorphism property
since by part (b) of Theorem 1 the local results for the original normal map are the
same as global results for a map based on the critical cone.
To introduce the symmetry property we need to make a certain reduction. We
describe this in the following proposition, which for simplicity we phrase in terms
of linear transformations; as noted above, if a( x) = Ax + a then we always have
ac = Ac + a, so there is no loss of generality in considering only the linear case.
The proposition uses the concept of Schur complement, defined as follows: if Mis
a matrix partitioned as

with Q nonsingular, then the Schur complement of Q in Mis (M/Q) =T - SQ-1R.

For more information about Schur complements see [1).

Proposition 1 Let C be a nonempty polyhedral convex subset of R n and let A be a

linear transformation /rom R n to Rn.
a. If C has a lineality space L of dimension j > 0 let D = C n L.1. Consider a
basis for R n consisting of j elements of Land n - j elements of L.1. Let the matrix
of A in this basis be

(1~~ 1~:)·
Then Ac is a homeomorphism of R n if and only if (i) ALL is nonsinguiar, and (ii)
(A/ALdD is a homeomorphism of Rn-j.
b. Suppose H is the affine hull of C, and define an affine transformation T :
H --+ H by T(h) = IIHA(h) for h E H. Then Ac is a homeomorphism of R n if and
only if Tc is a homeomorphism of H.

Proof. Part (a) is Proposition 4.1 of (13) and part (b) is a specialization of Propo-
sition 2.3 of [14]. 0
We use Proposition 1 by starting with an arbitrary polyhedral convex set C, then
using part (a) to reduce the homeomorphism question to one for a normal map over the
 263

set D = C n LJ., regarded as a subset of RR-i, where L is the j-dimensionallineality

space of C. Then, if D has dimension k < n - j we further reduce the problem to one
in the space Rk as shown in part (b). After these reductions we are left with a problem
in which the underlying set has full dimension and contains no lines. The next theorem
states the homeomorphism condition for such problems when A is symmetric.

Theorem 2 Let C be a polyhedral convex set of dimension n in RR, containing no

lines, and let A be a symmetric linear transformation from RR to RR. Then Ac is a
homeomorphism if and only if A is positive definite.

Proof. This is Theorem 3.1 of [14]. 0

Note that to be able to apply Theorem 2 we do not necessarily have to know that
A is symmetric, only that the matrix appearing in the reduced problem is symmet-
ric. As we shall see in the next section, we are typically faced with the problem of
determining that a PL normal map of the form A K , where K is some critical cone, is
a homeomorphism. The critical cone K will often be much simpler in structure than
the underlying set C, and in consequence when we apply the reductions described in
Proposition 1 we may end up with a matrix that is symmetric even when A is not.
On the other hand, if the original matrix A is symmetric then the reductions in
Proposition 1 do not disturb the symmetry. To see this, note that the Schur com-
plement defined above will be symmetric if the original matrix was symmetric, so
the reduction in part (a) preserves symmetry. Further, the function T(h) = IIHA(h)
appearing in part (b) is of the form P A( h) + p, where P is the orthogonal linear pro-
jector on the subspace, say M, parallel to H and p is the orthogonal projection of
the origin on H. Then P must be symmetric, so T(h) = (PAP*)(h) + p and we have
Tc(z) = (PAP*)c(z) +p. Therefore the reduction of part (b) also preserves symmetry
if it is present.
In applying part (b) when C has dimension k < n, it is generally best to take Rk
as a model for H. An n x k matrix Q with orthonormal columns forming a basis for M
provides a convenient representation. Then points y E R k correspond to points h E H
by h = Qy+p and y = Q*(h - p); note that Q*Q = I, so we have Q*[(Qy+ p) - p] = y.
Then C corresponds to the (polyhedral convex) subset G = Q*(C - p) of Rk. Using
this representation we reduce the matrix in part (b) to Q* AQ, which may be very much
smaller than A. The normal-map equation in the smaller space Rk corresponding to
Ac(z) = a is
(Q* AQ)G(y) = Q*(a - Ap),
and a solution y of this equation corresponds to a solution z of the original equation
through the formula

z = (Qy + p) - (I - QQ*)[A(QYG + p) - a],

in which YG is the projection of yon G.

We finish this section with an example illustrating application of some of these
results. Suppose that we wish to find stationary points of the function

¢>(X) = -(1/2)x~ + 2XIX2 + (1/2)x~ + Xl - fX2

264

over the nonnegative orthant R!. Here I: can be any real parameter. We can express
this problem by writing the first-order necessary condition in the form

( -1 2)
where

A = 2 1 ' C=R!.

It is not hard to see that Zo = (-1,0). The critical cone of Cat Zo is then J{ = {O} xR+,
and to check whether Ac is a local homeomorphism at Zo we should see if AK is a
homeomorphism. The set K contains no lines, but it is not offull dimension. Therefore
we reduce the problem using part (b) of Proposition 1, using R 1 as a representation for
the affine hull {O} x R of K, and obtaining the trivial normal map (l)R+' The matrix
here is certainly positive definite, so the original Ac was a local homeomorphism.
In fact, for I: near zero the solution is

_ { Zo + (-21:, 1:) if I: ~ 0,
Z, - Zo + (0, 1:) otherwise,

and the stationary points are the projections of these z, onto C, namely

_ { (0, 1:) if I: ~ 0,
x, - (0, 0) otherwise,

so that both z, and x, are single-valued, continuous (but nonsmooth) functions of 1:.
The theory of this section showed us how to determine if solutions like the z, and
x, of the last example existed, were unique, and were continuous functions of f. In
the next section we see that this theory also helps us to establish a general implicit-
function theorem that will predict the existence, uniqueness, and (Lipschitz) continuity
of solutions for normal maps induced by differentiable nonlinear functions, rather than
just the affine functions we have been considering here. In addition, we shall see that
the affine theory provides a powerful computational tool for locally approximating the
solutions of the nonlinear problems.

4. IMPLICIT FUNCTIONS AND SENSITIVITY ANALYSIS

In this section we return to the Variational Inequality (2.1), and using the tools
we have already developed we rewrite it in the form

Find z(y) ERn such that f(·,y)c(z(y)) = o. (4.1)

Because of the equivalence of (2.3) and (2.4), a solution z(y) of (4.1) yields a solution
x(y) of (2.1) via the projection operator: x(y) = IIc(z(y)). Conversely, any solution
x(y) of (2.1) is derived via projection from the solution z(y) = x(y) - f(x(y), y) of
(4.1). So we need only be concerned with solving (4.1). We shall give a version of the
implicit-function theorem that applies to that equation, then a reformulated corollary
that applies directly to (2.1).
 265

In stating the first result we need the concept of B-derivativej a function 9 from
an open subset r of R m to Rq is said to be B-differentiable (Bouligand differentiable)
at Xo E r if there is a positively homogeneous function dg(xo) : R m -+ Rq (the B-
derivative of 9 at xo) having the property that

g(xo + h) = g(xo) + dg(xo)(h) + o(h).

H dg(xo) happens to be linear then it is an ordinary Frechet derivative (F-derivative),
but it need not be linear. We also need a certain quantitative measure of the local
injectivity of a function: for the 9 just considered we define

6(g,r) = inf{lIg(x) - g(y)llIlIx - ylll x '" y, X,y E r}.

This 6 amounts to the reciprocal of a Lipschitz constant for g-l. Clearly 9 is injective
on r if 6(g, r) > 0, but the converse need not be true (think of x 3 near the origin).
Let (zo, Yo) E II- 1(fl) X e with fch Yo)(zo) = 0, and write Xo = IIc(zo). Suppose
that the function f(x,y) has a strong partial F-derivative in x at (xo,Yo), which we
write d",f(xo,Yo)j let L(s) be the linearization operator defined by

L(s) = d",f(xo,yo)(s). (4.2)

We then write K for the critical cone K(zo), and observe that the normal map LK is
given by

One expects implicit-function theorems to involve derivatives, and this map is going
to play the part of a derivative for our purposes.

Theorem 3 Let C, fl, e, and f be as in Section 2. Let CIi = IIc1 (fl) and let Zo E CIi
and Yo E e with fh yo)c(zo) = 0. Suppose that for some 8 and each x E fl, f(x,.) is
Lipschitzian on e with modulus 8. Write K = K(zo) and Xo = IIc(zo), and define L
by (./.2). Assume the following hypotheses:
a. The partial derivative dxf(xo, Yo) is strong.
b. LK is a homeomorphism.
Then 6(LK' RR) > 0, and for each ,\ > 6(LK' RRt 18 there exist neighborhoods Z
of Zo, X of Xo, and Y of Yo, and a function z : Y -+ RR, such that:
a. z(Yo)= zoo
b. z is Lipschitzian with modulus ,\.
c. For each y E Y, z(y) is the unique point in Z satisfying (4.1), and IIc(z(y))
is the unique point in en X satisfying (2.1).

Proof. This theorem is a specialization of the general implicit-function result given

in Theorem 3.2 of [12]. Therefore we need to verify the hypotheses of that theorem.
Once we do so, the conclusions will follow from the general result, except for the claim
about x(y), which follows from our earlier discussion of the relationship between z(y)
and x(y).
Most of the hypotheses are direct translations of those in [12]j however, there are
two that we need to demonstrate. First, we have to show that LK(Z - zo) strongly
approximates fC(',y)(z) in z at (Zo, Yo), in the sense of [12]; then we can use this in
266

place of the function f appearing in Theorem 3.2 of (12) (not to be confused with
our present I). Second, we must prove the claim that 6(LK' Rn) > o. In fact, the
second point is easy, because we have assumed that LK is a homeomorphism, so that
its inverse is well defined on all of RR. As the inverse, like L K , is affine in each of the
finitely many cells of JVK, it must be Lipschitzian. Then the reciprocal of the Lipschitz
constant provides a positive lower bound for 6(LK,RR).
To prove the first assertion, we note that Proposition 4.1 of (12) shows that a strong
approximation of fO(·,y)(z) in z at (zo, Yo) is given by
d",f(xo, Yo)[I1c(z) - xo] + [(z - zo) - (I1c(z) - xo)]. (4.3)
All we need to do is to show that in our case (4.3) is identical to LK(Z - zo) for z
near zoo We first note that as C is polyhedral, by Lemma 5(i) of [9] we have for all
sufficiently small hERR the equation
I1c{zo + h) = I1c(zo) + I1K{h). (4.4)
Therefore we can rewrite (4.3) as
d,.f(xo, Yo)[IIK(z - zo») + [(z - zo) - IIK(z - zo»),
which is exactly LK(Z - zo). This completes the proof. 0
At this point, we know that under the hypotheses of Theorem 3 there is a locally
unique, Lipschitzian function giving a solution of (4.1), or equivalently of (2.1), for
each y near Yo. That is, we have identified conditions under which slightly perturbed
problems will have solutions that behave reasonably nicely as functions of the per-
turbations. Our final result will give a method of approximating those solutions by
solving linearized Variational Inequalities or normal maps that are, in general, much
simpler to deal with than are (2.1) or (4.1) respectively. The idea is simply to solve an
approximate problem to obtain an approximate solution, and the approximate problem
is, just as in the classical implicit-function theorem, constructed by using derivatives
of the nonlinear function of the original problem.
Theorem 4 Assume the notation and hypotheses of Theorem 9, and suppose in ad-
dition that f has a partial B-derivative dyf(xo, yo) with respect to y at (xo, yo). Then
the functions z(y) and xCV) of Theorem 9 are B-differentiable at Yo with
(4.5)
and
= IlK 0 dz{yo).
dx{yo) (4.6)
Proof. We know from Theorem 3 that xCV) = I1c(z{y». Therefore if we prove (4.5)
then (4.6) will follow from the chain rule for B-derivatives (e.g., Corollary A.4 of [11]),
since (4.4) implies in particular that dIlc(zo) = I1K(zo) = IlK.
To prove (4.5) we can use Part (a) of Theorem 3.3 in [12], provided that we show
that d,J(xo,yo)(·) is the partial B-derivative in y of fc(·,Y)(z) at the point (zo,Yo) (in
fact, it is then an F-derivative because it is linear). To do this we just write
fc(-,yo + k)(zo) = f(xo, Yo + k) + (zo - xo)
= f(xo, Yo) + dyf(xo, yo)(k) + o(k) + (zo - xo)
= fO(·, Yo){zo) + dyf{xo, Yo)(k) + o(k).
 267

The claim now follows from the uniqueness of the F-derivative. 0

Theorem 4 has an important computational meaning. If we note that (4.5) says

and that dX(Yo)(k) is the projection on K of dz(Yo)(k), then we see that w = dx(Yo)(k)
must solve the linear Variational Inequality problem of finding w E K such that

For each u E K, (u - w, d",f(xo, yo)(w) + dllf(xo, Yo)(k») ~ 0; (4.7)

further, under the assumptions of Theorems 3 and 4, (4.7) will have a unique solution
for each kERn. So we have reduced the question of sensitivity analysis of the nonlin-
ear Variational Inequality (2.1) to that of computing solutions of the linear problem
(4.7). At least in the finite-dimensional case, (4.7) is frequently easy to solve. Further,
the possibility of computing dz(yo) and dx(yo) leads applications other than sensitiv-
ity analysis; for example, Sellami [16] has recently used it in developing a homotopy
method of predictor-corrector type for solving nonlinear Variational Inequalities when
accurate starting points are not available.
Theorem 4 completes our mathematical coverage of sensitivity analysis for Varia-
tional Inequalities. We conclude this section with an example illustrating the use of
Theorems 3 and 4.
Consider finding stationary points for the problem of minimizing (in x)

if>(x, E) = -(1/4)fx~ + (1/2)(x~ + x~) - Xl - fX2 + X3,

over X in the polyhedral convex set

°
We can see that for Xo = and EO = 0, d",if>(xo, Eo) = (-1,0,1), so that the origin is
a stationary point (in fact it must be a minimizer, because for E = the objective °
function is convex). We write f for d",if> and take

= Xo - f(xo, Eo) = (1
zo -1), °
and then we have fc(zo, Eo) = o. The tangent cone of C at the origin is C itself, so the
critical cone K is the intersection of C with the hyperplane consisting of all X E R3
with Xl = X3. Therefore

We have
f(x, E) = (-1 X2 - E -EX~ + X3 + 1) ,
so that

d",f(xo, EO) = (01 °°°

0) ,
001
dd(xO,EO) = (0 -1 0).
We need to check that the homeomorphism condition of Theorem 3 holds. As d",f(xo, 1'0)
is symmetric and as K contains no lines, by Proposition 1 and Theorem 2 it suffices
268

to check that d.J(xo,f{J) is positive definite on the subspace M comprising the affine
hull of K. The two vectors

comprise an orthonormal basis for M, so as outlined in Section 3 we can examine the

matrix

U~~)o~nun~on.
which is certainly positive definite. Therefore we know that our theoretical results
apply, so for E near zero the locally unique stationary points comprise a Lipschitzian
function of E.
Moreover, we can obtain a first-order approximation to the curve of stationary
points by solving the linear problem (4.7) over our critical cone K. Here we have

d.J(xo, f{J)(w) + dd(xo, Eo)(k) = (0 W2 - k W3),

so we can minimize the quadratic function (1/2)( w~ + wD -

kW2 over the cone K.
By carrying out the computation we obtain w as the following function of the scalar
variable k:
(k /2) (1 1 1) ifk ~ 0,
{
w(k) = (k/2) (-1 1 -1) if k < o.
It is worth noting that w( k) is a nonsmooth function of k, so that this sensitivity
analysis could not have been carried out by the classical method using the smooth
implicit-function theorem (for which see, e.g., [3]).

5. CONCLUSION

We have presented a general method for performing sensitivity analysis on Varia-

tional Inequalities over polyhedral convex sets in RR. This method relies on techniques
of nonsmooth analysis: in particular it uses single-valued nonsmooth equations called
normal maps, which are equivalent to Variational Inequalities, and an implicit-function
theorem that predicts the existence and Lipschitz continuity of solutions to these nor-
mal maps. It also provides a technique for computing first-order approximations of
these solutions by solving a linearization of the original problem.

ACKNOWLEDGMENT. The research reported here was sponsored by the Na-

tional Science Foundation under Grant CCR-9109345 and by the Air Force Systems
Command, USAF, under Grant F49620-93-1-0068. The US Government has certain
rights in this material, and is authorized to reproduce and distribute reprints for Gov-
ernmental purposes notwithstanding any copyright notation thereon.

REFERENCES

[1] R. W. Cottle, "Manifestations of the Schur complement" , Linear Algebra and Its ApplicatioIls 8,
189-211, 1974.
 269

[2] B. C. Eaves, "A short course in solving equations with PL homotopies" ,in: "Nonlinear Program-
ming, SIAM-AMS Proceedings Vol 9," R. W. Cottle and C. E. Lemke, eds., American Mathemat-
ical Society, Providence, RI 1976.
[3] A. V. Fiacco, "Sensitivity analysis for nonlinear programming using penalty methods", Mathe-
matical Programming 10,287-311, 1976.
[4] P. T. Harker and J. S. Pang, "Finite-dimensional Variational Inequality and nonlinear complemen-
tarity problems: A survey of theory, algorithms and applications", Mathematical Programming
48, 161-220, 1990.
[5] P. Hartman and G. Stampacchia, "On some nonlinear elliptic differential functional equations",
Acta Mathematica 115, 153-188, 1966.
[6] A. J. King and R. T. RockafelIar, "Sensitivity analysis for nonsmooth generalized equations",
Mathematical Programming 55, 193-212, 1992.
[7] M. Kojima and R. Saigal, "A study of PC I homeomorphisms on subdivided polyhedrons" , SIAM
Journal on Mathematical Analysis 10,1299-1312,1979.
[8] D. Kuhn and R. Liiwen, "Piecewise affine bijections of R n , and the equation Sx+ - Tx- = y",
Linear Algebra and Its Applications 96, 109-129,1987.
[9] J. S. Pang, "Newton's method for B-differentiable equations", Mathematics of Operations Research
15,311-341,1990.
[10] D. Ralph, "On branching numbers of normal manifolds", Nonlinear Analysis: Theory, Methods
and Applications 22, 1041-1050,1994.
[11] S. M. Robinson, "Local structure offeasible sets in nonlinear programming, Part III: Stability and
sensitivity", Mathematical Programming Study 30,45-66, 1987.
[12] S. M. Robinson, "An implicit-function theorem for a class of nonsmooth functions", Mathematics
of Operations Research 16, 292-309, 1991.
[13] S. M. Robinson, "Normal maps induced by linear transformations", Mathematics of Operations
Research 17691-714, 1992.
[14] S. M. Robinson, "Nonsingularity and symmetry for linear normal maps", Mathematical Program-
ming 62,415-425, 1993.
[15] R. T. RockafelIar, "Convex Analysis," Princeton University Press, Princeton, NJ, 1970.
[16] H. Sellami, "A Continuation Method for Normal Maps," Ph.D. Dissertation, Departments of
Mathematics and Industrial Engineering, University of Wisconsin-Madison, Madison, WI, 1994.
[17] G. Stampacchia, "Variational inequalities", in: "Theory and Applications of Monotone Operators,
Proceedings of the NATO Advanced Study Institute, Venice, Italy," Edizioni Oderisi, Gubbio,
Italy, 1968.
MONOTONE RELATIONS AND NETWORK EQUILIBRIUM

Rockafellar R. Tyrrell

Departments of Mathematics and Applied Mathematics

University of Washington GN-50
Seattle, WA 98195, USA
Email: rtr@math.washington.edu

Abstract: Conditions for network equilibrium are developed in terms of vector-

valued flows and potentials and generalized resistance relations. The extent to which
the equilibrium can be expressed by a Variational Inequality or characterized by opti-
mization is analyzed. Emphasis is placed on maximal monotone relations, especially
subgradient relations associated with convex optimization.

Key words: Variational Inequalities, network equilibrium, vector flows, vector

potentials, maximal monotone relations, dual problems of optimization.

1. INTRODUCTION

Among the many important applications of Variational Inequalities is the expres-

sion of equilibrium conditions for flows in networks, in particular equilibrium involving
different kinds of traffic. Variational inequalities are a relative newcomer to the the-
ory of networks, however. Other approaches to equilibrium have grown out of the
classical study of electrical networks and their generalization to hydraulic networks
as well as the framework of transportation problems in operations research. A very
useful idea has been the duality between flows and potentials as expressed by systems
of possibly nonlinear or even multi valued relations imposed in the different elements
of a network.
The aim of this article is to illuminate the connections between Variational In-
equalities and these other approaches, with special attention paid to the extent to
which equilibrium may correspond to some sort of optimization. It is hoped that
the range of modeling possibilities thereby revealed will aid further in the formula-
tion of traffic problems and also in their solution by a wider class of computational
techniques.
Variational inequalities generalize conditions for optimality such as may be asso-
ciated with a variational principle, so we begin by reviewing how this comes about.
The need and desirability of working with multivalued mappings receives motivation
in this way, and a remarkable degree of flexibility in the application of numerical
methods is achieved. For simplicity the context here will be finite-dimensional.

271
 272

Next we develop a general formulation of network equilibrium for vector-valued

flows and potentials, paralleling the well known one for scalar-valued flows. We es-
tablish the circumstances in which this kind of equilibrium can be expressed by a
Variational Inequality. We show that, even when this is not the case, an expression is
available in which the many numerical approaches to calculating a zero of a possibly
set-valued mapping can be applied. In addition we study situations where network
equilibrium corresponds to solving a primal optimization problem for flows, a dual op-
timization problem for potentials, or a saddle point problems for flows and potentials
together.

2. VARIATIONAL INEQUALITIES AND OPTIMIZATION

The Variational Inequality problem with respect to a nonempty, closed, convex

set Z C RN and a continuous, single-valued mapping F : Z -+ RN is usually stated
in the form:

(VI) find zEZ such that (F(z), z - z) ~ 0 for all z E Z.

An alternative form, which is preferable for many reasons and will be especially fruitful
in what follows, is obtained by utilizing the notion of the normal cone Nz(z) to Z at
z, which in convex analysis consists of all vectors w such that (w, z - z) :::; 0 for all
z E Z (see [1]). The problem is then:

(VI') find zEZ such that - F(z) E Nz(z).

In either form the inspiration comes from the case where F is the gradient mapping
V' f associated with a continuously differentiable function f defined on a neighborhood
of Z. The Variational Inequality then expresses the first-order necessary condition for
optimality in the minimization of f over Z, this being not just necessary but sufficient
when f happens to be convex.
Variational inequalities in which F = V' f are usually called $ymmetric, whereas
all others are a$ymmetric. Really, this terminology is appropriate only under the
additional assumption that F is continuously differentiable, since that allows the
existence of a function f with F = V' f to be identified with the property of F that
the Jacobian matrix V' F( z) is symmetric everywhere. It is possible, of course, to have
F = V' f without F being differentiable at all, so that the Jacobian does not even exist.
On the other hand, in circumstances where F is differentiable but not continuously
differentiable the Jacobian might exist but not be symmetric, as indicated by classical
examples of twice differentiable functions f for which the matrix of second partial
derivatives is not symmetric.
The distinction between the symmetric and asymmetric cases is sometimes inter-
preted as marking the division between the Variational Inequality problems directly
reducible to optimization and the ones not so reducible. But this view is inaccurate
and potentially misleading. Variational inequalities can correspond to optimization
despite asymmetry, and whenever that is true they can be solved by optimization
techniques just as well as if they were symmetric, and without resorting to the intro-
duction of an artificial "gap" function.
For example, the Kuhn-Tucker conditions for a minimization problem with func-
tional constraints express first-order optimality in terms of an asymmetric Variational
Inequality in the primal and dual variables jointly. Suppose the problem consists of
minimizing fo(x) over all x E X satisfying (!I(x), ... ,fm(x)) E K, where K is a
 273

closed, convex cone in JR m , X is a nonempty, closed, convex set in JR n , and the func-
tions Ii are continuously differentiable. Let lex, y) stand for the Lagrangian expression
Io(x) + L~l Ydi(X). The generalized Kuhn-Tucker conditions that apply to this set-
ting, as established in [2] (Thms. 4.2, 10.6) under a basic constraint qualification, say
that for x E X to give a local minimum there must be a vector y E Y, where Y is the
cone polar to K, such that

- V'xl(x,y) E Nx(x), (2.1)

The search for a pair (x, y) satisfying this double relation is the Variational Inequality
problem (VI') in the case of

z = (x,y), Z=XxY, F(z) = (V'xl(x,y),-V'yl(x,y)). (2.2)

The Variational Inequality is asymmetric because F is not actually the gradient map-
ping of any function. Indeed, when F is continuously differentiable (through the
functions J; being twice continuously differentiable) its Jacobian matrix has an obvi-
ous lack of symmetry:

V'F(z) = [V'~xl(x,y) V'iyl(x,y)].

-V'YX1(x,y) -V'y)(x,y)
Beyond Kuhn-Tucker conditions, a similar pattern of asymmetry holds in the
characterization of optimality for broader problem models of composite type (see [2],
Sect. 10). These likewise concern dual vectors y along with primal vectors x as in
(2.1) and (2.2), but Y not necessarily a cone, and l(x,y) not necessarily the classical
Lagrangian. Also in this class are general problems of finding a saddle point (x, y)
of a differentiable function I( x, y) relative to closed, convex sets X and Y, since (2.1)
is necessary for lex, y) to have its minimum over x E X at x while lex, y) has its
maximum over y E Y at y.
In all these examples an asymmetric Variational Inequality is seen to be reducible
directly to optimization and therefore open to solution by methods of numerical op-
timization. Such methods, whether they are posed in a primal, dual, or primal-dual
context, inevitably aim at producing not only a primal vector x but an associated
dual vector y such that (2.1) holds.
The concept of monotonicity plays the key role in the theory of Variational In-
equalities that convexity plays in optimization. A Variational Inequality is monotone
if its mapping F : Z - t JRN is monotone in the sense that

(F(z') - F(z), z' - z) 2:: 0 for all z, z' E Z. (2.3)

(We denote by (.,.) the canonical inner product.) When F is continuously differ-
entiable this property is equivalent to the positive semidefiniteness of the (possibly
asymmetric) Jacobian matrix V' F( z) at every point z E Z. In the symmetric case
with F = V' f it corresponds to f being convex. The Variational Inequality then
describes the solution(s) to a problem of minimizing a convex function over a convex
set.
An important asymmetric example of monotonicity is encountered when F has the
form (2.2) relative to a product of convex sets X and Y, and the function I(x, y) is
convex in x E X and concave in y E Y. Such a Variational Inequality corresponds to
convex optimization as well. It characterizes solutions x to a certain primal problem of
minimization by means of a saddle point (x, jj), where jj solves a certain dual problem
of maximization.
274

As valuable as the notion of a Variational Inequality has turned out to be, it has
definite limitations which need to be appreciated if connections with optimization are
fully to be understood. One limitation, which fortunately is easy to get around, is the
single-valuedness of the mapping F. On the surface, this excludes applications to areas
like nonsmooth optimization. A more serious limitation, however, is the requirement
that the set Z be convex. When a Variational Inequality problem is stated in the
form (VI), the convexity of Z is essential for it to make good sense, but in form (VI ')
the way is open to assigning to the normal cone Nz(z) a definition appropriate not
only for convex sets Z but nonconvex sets as well. For instance, N z( z) can be taken
to be the Clarke normal cone or the smaller cone that has received special emphasis
in the work of Mordukhovich (see [2], Sect. 10).
The point is that although Variational Inequalities in which Z is convex do cover
some problems of nonconvex optimization through extended Kuhn-Tucker conditions,
as already discussed, there is something rather strained about the formulation. The
case where Z is convex and F is monotone is natural in providing a platform for a
theory of Variational Inequalities that mirrors convex optimization. The case where
Z is potentially not convex and F not monotone is well motivated too, if interpreted
in the manner just described. But the hybrid case where Z is convex, yet F is not
monotone, draws boundaries rather artificially.
Still another way of stating the basic Variational Inequality problem, which will
serve as a guide in our discussion of network equilibrium, is:

(VI") find z E Z such that 0 E T(i),

h T( ) _ {F(z)+Nz(z) if z E Z,
were z - 0 ifz~Z.

At first this format may seem unappealing because it requires working with a set-
valued mapping T. As support for an alternative point of view, however, it is rich in
theoretical implications.
A general mapping T that assigns to each z E R,N a subset T(z) C JRN can be
regarded as an ordinary single-valued mapping from JRN to the space 2RN. For most
purposes, though, there is much more to be gained by identifying T with the set

gphT:= ((z,w) E R,N X R,N 1 w E T(z)} (2.4)

as its "graph" and thinking of it thus as expressing a relation between vectors z and
w. The "effective domain" dom T and "effective range" rge T of T are defined then
by
domT:= {z 1 T(z) I- 0}, rgeT:= {w 13 z, wE T(z)}. (2.5)
In this framework, which we signal this framework by writing T : JRN =t R,N in place
of T : nN -+ n N, T is regarded as single-valued, empty-valued or multivalued at z
according to whether T(z) is a singleton, the empty set, or a set with more than one
element. The "inverse" of T is the mapping T- 1 : JRN =t JRN defined by

T-1(w) = {z 1wE T(z)}. (2.6)

Clearly domT- 1 = rgeT and rgeT- 1 = domT.

The monotonicity property introduced in (2.3) for a mapping F : Z -+ JRN has
the following generalization. A mapping T : JRN =t JRN is called monotone if

{Z' - Z, Wi - w} ~ 0 whenever wE T(z), Wi E T(z). (2.7)

 275

It is maximal monotone if it is monotone but its graph cannot be enlarged without

losing monotonicity, i.e., if for every choice of vectors £ and w with w ~ T(£), there
z
exist and w with w E T(Z) such that (£ - z,
w - w) < o.
Theorem 1 (Rockafellar [3], Thm. 3)
For the Variational Inequality problem that corresponds to a nonempty, closed,
convex set Z C RN and a continuous mapping F : Z -+ RN as expressed in (VI"), if
F is monotone, then the associated mapping T : RN =t JRN is maximal monotone.

This result, which characterizes monotone Variational Inequality problems as

problems of solving 0 E T(:Z) for certain kinds of maximal monotone mappings T,
will enable us to identify different ways in which conditions for network equilibrium
can be cast in terms of a Variational Inequality as long as monotonicity is present.
In appealing to it we will essentially be limiting our attention to problems with over-
tones of convexity, as already explained. An extension beyond monotonicity would
no doubt be possible, but we will not undertake it in this article.

3. NETWORK EQUILIBRIUM

For purposes here, a network consists of a finite set of nodes indexed by i E I =

{I, ... , m} and a finite set of arcs indexed by j E J = {I, ... , n}. Each arc j has an
initial node and a terminal node, which are different. The information about these
nodes is embodied in the m x n incidence matrix E = (eij) for the network, where

I, if node i is the initial node of arc j,

eij = { -1, if node i is the terminal node of arc j, (3.1)
0, otherwise.

Classical network theory is concerned with scalar-valued flows, but here we will
be occupied with vector-valued flows. A d-dimensional flow x is a "supervector"
(Xl, ... , xn) where each component Xj is a vector (Xjl, ... , Xjd) E JRd. In applications,
X jk will represent the amount of scalar flow of type k in arc j. Constraints on the
magnitude and direction of such flow amounts may be imposed later, but for now we
note merely that a quantity Xjk > 0 is to be interpreted as flowing from the initial
node of arc j to its terminal node, whereas a quantity Xjk < 0 refers to physical flow
in the opposite direction.
The divergence of the flow X at node i is the vector Yi = (Yil, .. . ,Yid) E JRd in
which Yik gives the net amount of flow type k that originates at node i. This is
expressed by

Yi = L eijXj for each i E I, or in summary, Y = Ex, (3.2)

jeJ

where y is the supervector (Yl, ... ,Ym). Node i is a source for flow type k under x if
Yik > 0 and a sink if Yik < o. Flow type k is conserved at node i if Yki = o.
Dual to the concept of flow is that of potential. Ad-dimensional potent'jul U
is a supervector (Ub ... , urn), each component of which designates a vector Ui =
(Uil, ... , Uid). The quantity Uik refers to the potential of type k at node i, and abstract
quantity which in economic applications may have a price interpretation. Relative to
such a vector-valued potential u, the tension Vj in arc j is the difference Ui' - Ui, where
 276

i is the initial node of arc j and i' is the terminal node. In terms of the incidence
matrix E this comes out as

Vj =- L Uieij for each j E J, or in summary, v = _ET u, (3.3)

iel

where v = (VI, ... ,vn ). Each tension vector Vj = (Vjl,' .. ,Vjd) E IRd has components
Vjk giving the difference in potential type k in arc j.
Equilibrium problems in this context can usefully be set up on several levels.
To begin with, we consider the case of fixed supplies and demands. By a supply
b = (bI, ... ,bm ) in the network we will mean the assignment to each node i of a vector
bi = (bi1 , ... , bid), where bik designates the supply of flow type k at node i, this being
the value that the divergence Yki will be required to have. Negative supply values bik
correspond of course to demand. A value bik = 0 indicates that flow type k is required
to be conserved at node i.
By a flow-tension relation in arc j we will mean a subset of IRd x IRd specifying
the flow-tension pairs (xj,Vj) permitted to coexist in arc j. We interpret this subset
as the graph gphRj of a mapping R j : IRd ~ IRdj thus, Xj and Vj are related in the
required manner if and only if Vj E Rj(xj), or equivalently Xj E R j l(Vj).
The classical analogy for scalar-valued flows in an electrical network lies with
resistance and conductance. In such a network each arc j represents an electrical
component with a certain "characteristic curve" which describes how the flow (elec-
trical current) through j corresponds to the tension (voltage difference) across j. This
characteristic curve is the graph of R j , and the R j is "resistance mapping" for the arc
j j the inverse Rjl is the "conductance mapping" for the arc.
If arc j represents an ideal resistor, behaving in accordance with Ohm's Law with
resistance value r j > 0, its characteristic curve is a line in IR x IR with slope r j.
Then both R j and Rt are single-valued and linear. Nonlinear resistors correspond
to more complicated curves in IR x IR. Sometimes R j or Rjl, or both, can fail to be
single-valued in such a context. For instance, in the case of an ideal diode, the graph
of R j is the subset of IR x IR formed by the union of the nonnegative xj-axis and the
nonpositive vraxis.

Equilibrium Problem 1 Given for each arc j a mapping R j : IRd ~ IRd and
for each node i a supply vector bi E IR d, find a flow x and a potential U for which the
corresponding divergence Y and tension v satisfy

{ Vj E Rj(xj) for all~ E J,

Yi = bi for all z E I.

A more general formulation of equilibrium dispenses with fixed supplies and de-
mands and instead allows a divergence-potential relation to be assigned to each node.
Again, we think of such a relation in terms of a subset of IRd x IRd viewed as the
graph G(Si) of a mapping Si : IRd ~ IRd. The divergence Yi and potential Ui at node
i are related in the required manner when Ui E Si(Yi), or equivalently, Yi E Si- 1( Ui).
In Equilibrium Problem 1, the graph of Si is the set {bi} X IRd for every node ij in
other words, we have

Si- 1 : Ui 1--+ bi (constant mapping). (3.4)

But instead now, Si- 1 might for instance be a nonconstant, single-valued mapping.
An economic interpretation in some models where the kinds of flow represent different
 277

kinds of commodities is that Ui is a vector of prices at i for these commodities, and

S;l{ Ui) gives the amounts supplied (produced), or with negative signs, demanded
(consumed) at i in response to these prices.

Equilibrium Problem 2 Given for each arc j a mapping Rj : lR d =t lR d and

for each node i a mapping Si : F =t R d , find a flow x and a potential U for which
the corresponding divergence y and tension v satisfy

{ Vj E Rj(xj) for all ~ E J,

Ui E Si(Yi) for aliI E I.

By the maximal monotone version of Equilibrium Problem 1, we will mean the

version where every mapping R j is maximal monotone. Likewise, by the maximal
monotone version of Equilibrium Problem 2, we will mean the version where every
mapping Rj and every mapping Si is maximal monotone. (Note that Si is maximal
monotone in particular when S;l is a constant mapping, as seen when Equilibrium
Problem 1 is imbedded within Equilibrium Problem 2.) For scalar-valued flows and
potentials (d = 1), the role of maximal monotonicity was first explored by Minty [4J,
who concentrated on Equilibrium Problem 1 with bi = 0 for all i. The theory of this
case is fully presented in the book [5J.
One of the many nice features of maximal monotonicity with d = 1 is that the
graphs of the relations are indeed "curves," i.e., sets nicely parameterized by a real
variable. For d > 1 there is the following generalization.

Theorem 2 (Minty [6])

The graph gph T of any maximal monotone mapping T : Rd =t lRd is homeomor-
phic to Rd. Moreover, the homeomorphism can be set up to be Lipschitz continuous
in both directions.

In applications to traffic equilibrium, the following model is basic. For each node i
consider a supply vector bi = (bib ... , bid). For each arc j let Xj = [0, eilJ x··· X [0, eNJ,
where eik is the upper bound for flow of type k in the arc in question, ejk ~ O. (When
ejk = 0, flow of type k is forbidden in this arc.) For Xj E X j let

for a continuous, nondecreasing function Cj : [0,00) -+ [0,00) and fixed weights Wjk ~
o. The conditions for traffic equilibrium are taken then to be those of Equilibrium
Problem 1 with

Xi E Xj, v·J - F-(x·)

J J E N xJ(x·)
J.

In models of this kind it is common to have only one source and one sink for each
type of traffic. Then for each k there is exactly one node i with bik > 0 and exactly
one other node i' with bilk < O. Often the models are set up in terms of flows along
particular paths instead of just flow amounts in each arc. Such models are much more
complicated to work with, yet they seem not to offer any serious advantages, because
the flow of traffic of type k can readily be represented, at any stage of computation
or analysis where desired, as a sum of flows along paths from source to sink. See
Rockafellar [5], Secs. 4A and 4B, for the elementary algorithm that is involved.
Generalized traffic models in the format of Equilibrium Problem 2 instead of Equi-
librium Problem 1 might arise from situations in which the supply and demand for the
278

different kinds of flow could be affected by the state of congestion. Models attempting
to treat the difficulties of passing through various nodes would not necessarily require
passage to Equilibrium Problem 2. Instead one might use the device of introducing
"internal arcs" in such nodes; the needed equilibrium conditions could then be cen-
tered on such arcs, see [5], Sec. 3L. In either approach, dynamical networks could be
formulated in a space-time framework to handle traffic equilibrium in the sense of a
day-to-day cycle; see [5], Secs. 1H and 3L.

4. EQUILIBRIUM AS A VARIATIONAL INEQUALITY

To what extent are the equilibrium conditions in Equilibrium Problems 1 and 2

expressible in terms of a Variational Inequality? In general they go beyond the limited
format served by Variational Inequalities, but there are important cases where they
fit with it. It is instructive to see that this can occur in several different ways.
Let us say that a flow-tension relation for arc j is of primal VI-type if its mapping
Rj : F =¥ JRd has the form

R .( .) _ {Fj(Xj) + Nx,(xj), if Xj E Xj, (4.1)

1 X, - 0, 'f Xj dy; X j,
1

where Xj is a nonempty, closed, convex subset of JRd and F j : X j -+ JRd is continuous

(then Xj = dom R j ). On the other hand, let us say that this relation is of d'aul
VI-type if the inverse mapping Rjl has such form:

(4.2)

where Vi is a nonempty, closed, convex subset of F and c)j : Vi -+ JRd is continuous

(then Vi = rge Rj). In a similar vein, let us say that a divergence-potential relation
for node i is of primal VI-type if its mapping Si : JRd -+ JRd has the form

(4.3)

where Y; is a nonempty, closed, convex subset of JRd and Gi : Y; -+ JRd is continuous

(then Y; = domSi), while it is of dual VI-type if the inverse mapping Sil has the
form
Si-I(Ui) = {IliAuj) + Nu,(uj}, ~f Ui E Ui , (4.4)
0, If Ui rt Ui ,
where Ui is a nonempty, closed, convex subset of JRd and llI i : Ui -+ JRd is continuous
(then Ui = rgeSi)'
The divergence-potential relations in the special case (3.4) used in imbedding
Equilibrium Problem 1 within Equilibrium Problem 2 are obviously of dual VI-type
with Ui = JRd and IlI(Ui) == bi, but they are also of primal VI-type with Y; = {bi} and
Gi(Yi) = o.
Our results will utilize the theory of relative interiors. Recall that the relative
interior ri C of a convex set C is the interior of C relative to its affine hull (see [1],
Sec. 6). An affine set is its own relative interior; in particular, if C = {a} (a singleton
set) then ri C = {a}. Recall further the C is polyhedral when it is representable as
the intersection of a finite collection of closed half-spaces, or equivalently as the set
of solutions to a system of finitely many (weak) linear inequalities.
 279

Theorem 3 (Rockafellar [1), Cor. 23.8.1)

Suppose that C = Ct n ... n CT> where each Cl is a convex subset of JRN, and
z z
suppose there exists E C such that actually E ri C 1 for each I such that C 1 is not
polyhedral. Then at all points z E C one has Nc(z) = Nc.(z) + ... + Ncr(z).

Because Equilibrium Problem 1 is covered by Equilibrium Problem 2, we develop

results in terms of Equilibrium Problem 2 and then specialize.

Theorem 4 (Variational Inequalities from Equilibrium Problem 2).

(a) (primal case). In Equilibrium Problem 2, suppose all the flow-tension relations
and divergence-potential relations are of primal VI-type: (4.1) and (4-3) hold. Ass'ume
there is at least one flow X which, with its divergence ii, satisfies Xj E riXj for all
j E J and iii E ri Yi for all i E I; in this assumption, uri" can be omitted for any
Xj or Yi that is polyhedral. The problem is equivalent then to solving the Variational
Inequality for

Z ={z=(Xt, ... ,xn,Yt, ... ,Ym)lxjEXj , Y;EYi, y=Ex},

F(z) = (Ft(xd, ... , Fn(xn), Gt(Yt), ... , Gm(Ym)).
This Variational Inequality is monotone when every F j and G; is monotone; then one
has a maximal monotone version of Equilibrium Problem 2.
(b) (dual case). In Equilibrium Problem 2, suppose all the flow-tension relations
and divergence-potential relations are of dual VI-type: (4.2) and (4.4) hold. Assume
there is at least one potential ii which, with its tension v, satisfies Vj E ri Vj for all
j E J and iii E ri Ui for all i E I; in this assumption, uri" can be omitted for any
Vj or Ui that is polyhedral. The problem is equivalent then to solving the Variational
Inequality for

Z ={z=(Vt, ... ,vn,ut, ... ,urn)lvjEVj, uiEUi, v=-ETu},

F(z) = (1)t(Vt), ... , 1>n(vn), Wt(Ut), ... , Wrn(u m)).
This Variational Inequality is monotone when every 1>j and Wi is monotone; then one
has a maximal monotone version of Equilibrium Problem 2.
(c) (primal-dual case). In Equilibrium Problem 2, suppose all the flow-tension
relations are of primal VI-type and all the divergence-potential relations are of dual VI-
type: (4.1) and (4.4) hold. The problem is equivalent then to solving the Variational
Inequality for

Z = {z = (Xt, ... ,xn,Ut, ... ,urn) I Xj E Xj, Ui E Ui},

F(z) = (Ft(xt}, ... , Fn(xn), Wt(Ut), ... , Wm(u m)) + (ET U, -Ex).

This Variational Inequality is monotone when every Fj and Wi is monotone; then one
has a maximal monotone version of Equilibrium Problem 2.

Proof. The analysis of the normal cone N z( z) at points z E Z is crucial in each

case. In (a) and (b) the main tool for this purpose will be Theorem 3.
In (a) we have Z = L n Zo for the subspace L = {z = (x,y) I Y = Ex} and
the product set Zo = X t x ... X Xn X ••• x Yi x ... X Ym. We can also express Zo
as the intersection X~ n ... n X~ n Y; n ... n Y':' by taking Xj to be the subset of
(JRd)n+m having the same formula as Zo but with all factors except X j replaced by
JRd, and likewise for li'. Then ri Xj and ri li' have this form as well, with ri Xj and
ri Yi replacing the factors X j and Yi. Also, Xi and li' are polyhedral when X j and
280

Yo are polyhedral. Since ri L = L we see that the assumption in (a) about a certain
How x corresponds to the hypothesis of Theorem 3 when applied to the intersection
Z = L n X~ n ... n X~ n Y{ n ... n Y':'. We deduce thereby that
Nz(z) = NL(Z) + NXj(z) + ... + Nx:.(z) + Ny;(z) + ... + Ny,:. (z).
Here NL{Z) = LJ. = {(v,u) I v = -ETu}, whereas
NXj{z) + ... + Nx:.{z) + Ny;(z) + ... + Ny,:.{z)
= Nx,(xd X ••• x NXn(xn) X Ny,(Yl) x··· X NYn(Yn) = Nzo(z).
Thus, Nz(z) = LJ. + Nzo(z).
The Variational Inequality for Z and F, expressed in form (V!'), refers therefore
to the existence of z E L n Zo such that there exists 'Iii E LJ. with 'Iii - F(z) E Nzo{z).
To say that z E L n Zo is to say that z = (x,y) with y = Ex, Xj E Xj and y E Yo. To
say that 'Iii E LJ. with 'Iii - F{z) E Nzo(z) is to say that w = (ii, u) with ii = -ETu,
Vj - Fj(xj) E NxJ(xj) and Ui - Gi(Yi) E Ny,(Yi) for all arcs j and nodes i. From (4.1)
and (4.3) we conclude that the Variational Inequality comes down to the equilibrium
conditions in Equilibrium Problem 2.
When all the mappings F j and Gi in (a) are monotone, F is obviously monotone
as well. Then too, every relation Rj and Si is maximal monotone by Theorem 1, so
we have a maximal monotone version of Equilibrium Problem 2.
In case (b) the argument is closely parallel. We have Z = L n Zo for the subspace
L = {{v,u) I v = -ETu} and set Zo = 'Vi x ... x Vn X Ul X ••• X Urn. Again
through Theorem 3, the assumption about a potential u guarantees that Nz(z) =
NL(z) + Nzo{z) with NL(z) = LJ. = {(x,y) I Y = Ex} and Nzo(z) = NV,{Vl) x ... x
Nvn{vn ) x Nu,{ut} x··· X Num(u m ). The specified Variational Inequality in form (VI')
reduces then to the existence of z = (v,u) ELand w = (x,y) E LJ. such that iij E Vi
and Xj - CIi;(vj) E NvJ(vj) for all arcs j, while Ui E Ui and Yi -Wi(iii) E Nu;(iii) for all
nodes i. Because of (4.2) and (4.4), these conditions are identical to Xj E Rjl(Vj) and
Yi E Si- l (iii)' which are just another way of writing the ones in Equilibrium Problem 2.
When all the mappings CIi j and Wi in (b) are monotone, F is monotone too, and
through Theorem 1 the relations Rjl and Si- 1 are maximal monotone. Then R j and
Si are maximal monotone and we have a maximal monotone version of Equilibrium
Problem 2.
Case (c) is simpler and does not require Theorem 3. Without having to invoke
any constraint qualification we know that
Nz(z) = Nx,(xt} X .•• X NXn(x n) x NU,(Ul) X ••• X Num(u m ).
Consider z = (x, u) and let y = Ex and v = -ETu. To say that z E Z is to say
that Xj E Xj for all j and Ui E Ui for all i. The condition -F(z) E Nz(z) takes the
form then that -Fj(xj)+Vj E NxJ(xj) and -Wi(Ui)+Yi E Nu;(iii) for all i andj. By
(4.1) and (4.4) these properties are equivalent to having iij E Rj(xj) and Yi E S-I{Ui),
which are the same as the equilibrium conditions in Equilibrium Problem 2.
When Fj and Wi are monotone, the mapping
(x,u) 1-+ (Fl(Xl),'" ,Fn(xn), Wl(Ul),"" wm(um »
in (c) is monotone. The linear mapping (x,u) 1-+ (ETu,-Ex) is always monotone
(because its matrix is antisymmetric). Then F, as the sum of two monotone mappings,
is itself monotone. In this case the Variational Inequality in (c) is monotone. At the
same time the mappings Rj in (4.1) and Si- l in (4.4) are maximal monotone by virtue
of Theorem 1, so we have a maximal monotone version of Equilibrium Problem 2. 0
 281

Theorem 5 (Variational Inequalities from Equilibrium Problem 1)

(a) (primal case). In Equilibrium Problem 1, suppose all the flow-tension relations
are of primal VI-type: (4.1) holds. Assume there is at least one flow x with Ex = b
such that Xj E ri Xj for all j; in this assumption, "ri" can be omitted for any Xj that
is polyhedral. The problem is equivalent then to solving the Variational Inequality for

Z ={Z=(Xll""Xn) I xjEXj, Ex=b},

F(z) = (Fl(Xl),"" Fn(xn)).
This Variational Inequality is monotone when every Fj is monotone; then one has a
maximal monotone version of Equilibrium Problem 1.
(b) (dual case). In Equilibrium Problem 1, suppose all the flow-tension relations
are of dual VI-type: (4.2) holds. Suppose there is at least one potential it whose
tension v satisfies Vj E ri Vi for all j; in this assumption, "ri" can be omitted for
any Vi that is polyhedral. The problem is equivalent then to solving the Variational
Inequality for

Z ={z=(v}, ... ,vn,u}, ... ,um)lvjEVi, uiEJRd, v=-ETu},

F(z) = (<fll(vd, ... , <fln(vn ), b}, ... , bm ).

This Variational Inequality is monotone when every <flj is monotone; then one has a
maximal monotone version of Equilibrium Problem 1.
(c) (primal-dual case). In Equilibrium Problem 1, suppose all the flow-tension
relations are of primal VI-type: (4.1) holds. The problem is equivalent then to sol'ving
the Variational Inequality for

Z ={z=(xl, ... ,xn,u}, ... ,um)lxjEXj, UiEJR d },

F(z) = (Fl(xd, ... , Fn(xn), bl , ... , bm) + (ET u, -Ex).

This Variational Inequality is monotone when every Fj is monotone; then one has a
maximal monotone version of Equilibrium Problem 1.

Proof. Here we take the divergence-potential relations in Theorem 4 all to have

form (3.4). As observed, these relations are simultaneously of primal VI-type and of
dual VI-type, and they are maximal monotone. In (a) the form is so special that the
Yi arguments trivialize. 0
The standard traffic equilibrium model described at the end of Section 3 fits case
(a) of Theorem 5. Because Cj is nondecreasing in this example, F j is monotone and
we have a monotone Variational Inequality corresponding to a maximal monotone
version of Equilibrium Problem 1.

5. OPTIMAL FLOWS AND POTENTIALS

The network equilibrium problems in Section 3 correspond to in some important

cases to problems of optimization of flows and potentials, and vice versa. This corre-
spondence is more general than that associated with the Variational Inequality format
in its multiple modes in Section 4. We concentrate here on optimization of convex
type and describe the connections in terms of subgradients of convex functions. The
results obtained generalize the ones presented for scalar-valued flows and potentials
in Rockafellar [5J, Sec. 8H.
Let JR denote the extended real line JRu {±oo }. Recall that a function j : JRN -+ JR
is called convex if its "epigraph" epij:= ((z,G) E JRN X JR I G ~ j(z)} is a convex
 282

set. The "effective domain" domf := {z I fez) < oo} of such a function is then a
convex set in particular.
A convex function f on R,N is proper if fez) > -00 for all z, and fez) < 00 for
at least one z. It is lower Ilemicontinuous (lsc) if the set epif is closed in JRN x JR.
The function j* conjugate to f is defined by

r(w) = sup ((w,z) - fez)}.

zERN

When f is convex, proper and lsc, the conjugate function j* likewise is convex, proper
and lsc, and the function j** conjugate to j* is in turn f:

fez) = sup ((w,z) - r(w)}.

wERN

These matters are developed in detail in [1].

For a convex, proper, lsc function f on JRN, the subgradients of f at a point z are
the vectors w (if any) such that

fez) ~ fez) + (ro, z - z) for all z.

The set of these is denoted by af(z). Thus af denotes a mapping JRN =f JRN in the
general sense adopted in Section 2. Furthermore, the inverse mapping (al)-1 is the
subgradient mapping associated with the conjugate function j*:

wE af(z) {::::::} z E aj*(w). (5.1)

Theorem 6 (Moreau [6], Rockafellar [1], Cor. 31.5.2)
When f : JRN -+ JR is convex, proper and lsc, the mapping T = af : JRN =f JRN
ill maximal monotone.
When N > 1, not every maximal monotone mapping T : JRN =f JRN is of the
form af. The ones that are have the property of maximal cyclic monotonicity; see
[1], Thm. 24.9. Maximal cyclic monotonicity is the same as maximal monotonicity in
the one-dimensional case, however.
As with network equilibrium problems, it will be useful to consider network opti-
mization problems on two levels.

Primal Problem 1 Given for each arc j a convex, proper, lsc function fi on JRd
and for each node i a supply vector bi E JRd, minimize 1:jeJ fi(Xj) over all flows x
with divergence Y satisfying Yi = bi for all i E I.

Primal Problem 2 Given for each arc j a convex, proper, lsc function fi on JRd
and for each node i a convex, proper, lsc function gi on JRd, minimize 1:jEJ fi(Xj) +
1:iEI gi(Yi) over all flows x, where Y is the divergence of x.
It should be kept in mind that these problems have implicit constraints represented
through 00. In Primal Problem 1, a flow x with divergence y = b is not regarded as
feasible unless I;(Xj) < 00 for all j, i.e., Xj belongs to the convex set dom I; C JRd for
all j. In Primal Problem 2, there is the further requirement that Yi should belong to
the convex set domgi for all i. Primal Problem 2 reduces to Primal Problem 1 when

gi (Yi) = { 000 ifIf. YiYi i-= bi,

bi.
(5.2)

With these primal problems we associate the following dual problems in terms of
the conjugate convex functions.
 283

Dual Problem 1 Given for each arc j a convex, proper, lsc function Ii on JRd
and for each node i a supply vector bi E JRd, maximize - LjEJ f;(vj) - LiEI -(bi, Ui)
over all potentials U and their tensions v.

Dual Problem 2 Given for each arc j a convex, proper, lse function fj on JRd
and for each node i a convex, proper, lsc function gi on JRd, maximize - LjEJ f;( Vj)-
LiEf g:( Ui) over all potentials U and their tensions v.

In Dual Problem 1, the constraint Vj E domf; is implicit. In Dual Problem 2,

one also needs Ui E dom gi, since otherwise the expression being maximized has the
value -00. Note that Dual Problem 2 reduces to Dual Problem 1 under the choice
(5.2) for gi, because the conjugate convex function is then

gi( Ui) == (bi, Ui)' (5.3)

By the optimal values in these primal and dual problems we will mean the values
(in JR) giving the infimum or supremum in each case. The optimal solutions are the
elements (flows or potentials) for which these values are achieved, if any.
Finally, we formulate saddle point problems of Lagrangian type corresponding to
the optimization problems on both levels.

Saddle Problem 1 Given for each arc j a convex, proper, lsc function fj on JRd
and for each node i a supply vector bi E JRd, find a saddle point of the Lagrangian
function
L(x, u) = 'Lfi(Xj) - 'L(bi, Ui) + 'L
eij(ui, Xj)
jEJ iEI jEJ, iEI
with respect to minimizing over flows x but maximizing over potentials u.

Saddle Problem 2 Given for each arc j a convex, proper, lsc function Ii on JRd
and for each node i a convex, proper, lsc function gi on JRd, find a saddle point of the
Lagrangian function

L(x,u) = 'Lfi(Xj)- 'Lgi(Ui) + 'L eij(Uj,Xj)

jeJ iEI jEJ, iEI

with respect to minimizing over flows x but maximizing over potentials u.

In Saddle Problem 2, the expression L(x,u) is interpreted as -00 unless 'Ui E

domg; for every i E I. A pair (x,it) furnishes a saddle point relative to all flows and
potentials if and only if it furnishes a saddle point relative to X xU, where X is the
product of the sets dom Ii, and U is the product of the sets domgi; see [lJ, Sec. 36.

Theorem 7 In Equilibrium Problem 2, suppose the flow-tension and divergence-

potential relations have the form R j = ali for all j E J and Si = agi for all i E I,
where Ii and gi are convex, proper, lsc functions on JRd. Then one has a maximal
monotone version of the problem in which the following conditions on a flow x with
divergence jj and a potential it with tension v are equivalent to each other:
(a) i and it solve Equilibrium Problem 2.
(b) i and it solve Saddle Problem 2.
(c) i is an optimal solution to Primal Problem 2, it is an optimal solution to Dual
Problem 2, and the optimal values in these two problems are equal.
284

Proof. The condition for (x,u) to be a saddle point in (b) is that the expression
L(x, ii) should achieve its minimum over all flows x at x, whereas L(x, u) should
achieve its maximum over all potentials u at ii. The minimization part is equivalent
to having f;(xj) - (Vj,Xj) achieve its minimum over Xj E lR,d at Xj for each j E J,
where Vj = - EiEleijiii. This means that Vj E 8fj(xj) for all j. The maximization
part is equivalent to having (Yi,Ui) - g;(Ui) achieve its maximum over Ui E lR,d at
Ui for each i E I, where iii = EjEJ eijXj. This means that iii E 8g;(iii) for all i, or
since 8g: = (8git\ that iii E 8gi(Yi). These subgradient conditions are the same as
Vj E Rj(xj) and Ui E Si(Yi) under our hypothesis. Thus, (a) and (b) are equivalent.
To establish the equivalence of (b) with (c), let r( x) = supu L( X, '11.) and s( '11.) =
inf",L(x,u). It is elementary and well known in general minimax theory (cf. [1],
Sec. 36) that (x,u) furnishes a saddle point of L if and only if x minimizes rex),
ii maximizes s(u), and the minimum value rex) agrees with the maximum value
s(u). We merely need to observe now that rex) = EjEJ I;(Xj) + EiEI gi(Yi) (with Yi
standing for EjEJ eijXj), whereas s(u) = - EjEJ fi(Vj) - EiEI g;(Ui) (with Vj standing
for - EiEI eijUi). 0

Theorem 8 In Equilibrium Problem 1, suppose that the flow-tension relations are

of the form R j = 81; for all j E J, where each fj is a convex, proper, lsc function on
R,d. Then one has a maximal monotone version of the problem in which the following
conditions on a flow x with divergence Y and a potential u with tension v are equivalent
to each other:
(a) x and ii solve Equilibrium Problem 1.
(b) x and ii solve Saddle Problem 1.
(c) x is an optimal solution to Primal Problem 1, u is an optimal solution to Dual
Problem 1, and the optimal values in these two problems are equal.

Proof. This specializes Theorem 7 to the case of (5.1)-(5.2). o

To give these results their full force, supplementary conditions need to be provided
under which the optimal values in the primal and dual problems are equal. Then the
pairs (x, ii) satisfying the equilibrium problem in question are precisely those such
that x solves the primal problem and ii solves the dual problem. In other words,
equilibrium is fully reducible to optimization.

Theorem 9 Either one of the following assumptions suffices to guarantee that the
optimal values in Primal Problem 2 and Dual Problem 2 are equal:
(a) There is a flow x with divergence y such that xjEri ( dom 1;) for all arcs
j E J, and Yi E riC domgi) for all nodes i E f.
(b) There is a potential u with tension v such that Vj E ri ( dom fn for all arcs
j E J, and Ui E riC domgt) for all nodes i E I.
Proof. This follows from the Fenchel duality theorem in convex optimization,
specifically as a case of [1], Cor. 31.2.1. 0

Theorem 10 Either one of the following assumptions suffices to guarantee that

the optimal values in Primal Problem 1 and Dual Problem 1 are equal:
(a) There is a flow x with divergence Y such that xjEri ( dom 1;) for all arcs
j E J, and Yi = bi for all nodes i E f.
(b) There is a potential U with tension v such that Vj E ri ( dom fn for all arcs
j E J.
 285

Proof. Again we merely choose gi as in (5.1), so that g; is given by (5.2). 0

The constraint qualifications in Theorem 9 could be refined along the lines of the
one in Theorem 3. Recall that a convex function is polyhedral when its epigraph is a
polyhedral set; then the conjugate function is polyhedral as well; see [1], Sees. 19 and
20. Theorem 10 remains valid if in (a) the condition Xj E ri (domJ;) is weakened to
ifj E dom J; when J; is polyhedral, while the condition iii E ri ( dom gi) is weakened to
iii E domgi when gi is polyhedral; similarly in (b) and Theorem 10. This refinement
can be verified on the basis of the polyhedral results in [1], Theorem 31.1, by splitting
up the polyhedral and nonpolyhedral parts of the problem in a suitable way.
For scalar-valued flows and potentials, "ri" can be dropped entirely from Theo-
rems 9 and 10-see [5], Sec. 8H.
An example of particular importance is the one where the functions J; have the
form
r(X') _ {<pj(Xj), if Xj E Xj, (5.4)
J J - 00
, ifx·dX·
J y:. J'

with X j a nonempty, closed, convex subset of JRd and <Pj a differentiable convex
function defined on a open set containing Xj. Then

ofj(xj) 0,<p;(Xj) + Nxj(xj),

= {V if Xj E Xi>
if Xj ~ Xj.
(5.5)

In other words, the relation Rj = of; is of primal VI-type (4.1) with Fj = V<pj (this
mapping being continuous, because any differentiable convex function is continuously
differentiable). Then Primal Problem 1 corresponds to minimizing LjEJ <PiC x j) subject
to Xj E Xj and Yi = bi.
Other cases, where Rj is of dual VI-type, or where Si is of primal or dual VI-type,
can be identified in like manner. This reveals the extent to which the Variational
Inequalities developed for Equilibrium Problems 1 and 2 in Theorems 4 and 5 can
be viewed as coming from optimization with convexity. It should be noted that the
primal-dual Variational Inequalities correspond to the saddle point problems.

6. EQUILIBRIUM AS FINDING A ZERO OF A MAPPING

It was seen in formulation (VI") of the standard Variational Inequality problem

that such a problem in R,N can be stated as the "zero problem"
(ZP) find z satisfying 0 E T(z)
for a certain mapping T : R,N .=f R,n. If the Variational Inequality is monotone,
this mapping T is maximal monotone by Theorem 1. When Equilibrium Problems
1 and 2 reduce to Variational Inequalities in the manner of Theorems 4 and 5, they
fit this pattern in particular. We will demonstrate now, though, that these problems
always have an expression as (ZP), even when they lie beyond the scope of Variational
Inequalities, and that in their maximal monotone versions the associated mappings
T are maximal monotone-under mild assumptions.
Our motivation for this effort, beyond the nicety of tying several ideas together, is
that many computational methods have been developed for solving (ZP), even with T
not single-valued. The results open the way to applying those methods to Equilibrium
Problems 1 and 2.
As background we will need the following fact about the effective domain and
range of a maximal monotone mapping.
286

Theorem 11 (Minty [7])

The effective domain dom T and effective range rge T of any maximal monotone
mapping T : R,n =¥ R,N are almo~t convex, in the ~en~e that the sets C = cl ( dom T)
and D = cl ( rge T) are convex and such that dom T J ri C and rge T J ri D.

Because of this property of dom T and rge T we can speak of the relative interiors
ri ( dom T) and ri ( rge T), these being the same as the relative interiors of the convex
sets C and D in Theorem 11. When T = af for a convex, proper, lsc function
f : R,N -+ R, (cf. Theorem 6), one has ri ( dom T) = ri ( dom f) and ri ( rge T) =
riC domj*), see [1], Thm. 23.4.

Theorem 12 (Rockafellar (3], Thm. 2)

If Tl : R,N =¥ JRN and T2 : R,N =¥ JRN are maximal monotone and such that
ri (dom T1 ) U ri (dom T2) i- 0, then the mapping T = Tl + T2 likewise is maximal
monotone.

This result holds not only for the sum of two mappings, but any number. A proof
by induction is immediate from the fact that for convex sets CI, ... ,CT one has

ri (C1 n ... n C = ri C1 n ... n ri C

T ) T when ri C1 n ... n ri C i- 0,
T

cf. [1], Thm. 6.5.

Theorem 13 (Equilibrium Problem 2 as a Zero Problem)

(a) (primal case). Equilibrium Problem 2 can be identified with (ZP) in the case
of z = (x,y) and the mapping

where L = {(x,y) I y = Ex}, ~o that Ll. = {(v,u) I v = -ETu}. In the maximal

monotone version of Equilibrium Problem 2, T is maximal monotone as long as there
exi~ts a flow x with divergence fj ~uch that Xj E ri (dom Rj ) for all j E J and Yi E
ri ( dom Si) for all i E I.
(b) (dual case). Equilibrium Problem 2 can be identified with (ZP) in the case of
z = (v, u) and the mapping

if Z E L,
if Z rJ. L,

where L = {(v,u) I v = -ETu}, so that Ll. = {(x,y) I y = Ex}. In the maximal

monotone version of Equilibrium Problem 2, T is maximal monotone as long as there
exists a potential u with ten~ion v such that Vj E ri ( rge R j ) for all j E J and Ui E
ri (rge Si) for all i E I.
ec) (primal-dual case). Equilibrium Problem 2 can be identified with (ZP) in the
case of z = (x,u) and the mapping

In the maximal monotone ver~ion of Equilibrium Problem 2, T is maximal monotone.

 287

Proof. In all cases the equivalence of (ZP) with Equilibrium Problem 2 is elemen-
tary. What we have to verify are the maximal monotonicity assertions. For these we
will rely on Theorem 12. In (a), let

Clearly T = Tl + T2 • The maximal monotonicity of Tl follows at once from that

of every Rj and Sj. That of T2 is seen from Theorem 1 with F(z) == 0 through
the fact that T2(Z) = NL(z). Alternatively, T2 is maximal monotone by Theorem
6 because T2 = 88L for the indicator function 8L (which has the value 0 on L but
00 everywhere else). We have domT2 = L, a set which is its own relative interior,

whereas dom Tl is the product of the sets dom R j and dom Sj. The assumption in (a)
about x is equivalent therefore to the assumption that ri (dom T1 ) n ri (dom T2 ) "# 0,
and Theorem 12 then gives us the maximal monotonicity of T.
Similarly in (b), let

T1(z) = (Ri1(xd, ... , R;;-l(xn)' Sll(Yl)' ... ' S;;;l(Ym»'

T2 (z) = {L1., ifzEL,
0, if z ~ L.
Again, T = Tl +T 2. The maximal monotonicity of Tl is implied by that of every Rj and
Sj. The maximal monotonicity of T2 is justified by the same arguments used in case
(a). We have domT2 = L = ri domT2, while domTl is the product ofthe sets rgeRj
and rge Sj. The assumption in (b) about u means that ri ( dom T1 ) n ri ( dom T2 ) "# 0,
and it thus guarantees through Theorem 12 the maximal monotonicity of T.
In case (c) we get T = Tl + T2 by taking

T1(z) = (R1(xt), ... ,Rn(xn),Sll(Ut), ... ,S;;;l(U m »,

T2 (z) = (ETu,-Ex).
Here Tl is maximal monotone when R j and Sj are maximal monotone for all j and
i. On the other hand, T2 is maximal monotone by Theorem 1 as invoked for the
linear mapping F(z) = (ET u, -Ex) with Z taken to be the whole space (so that
Nz(z) = {OJ for all z). Then domT2 is the whole space, so that the condition
ri ( dom T1 ) n ri ( dom T2) "# 0 is satisfied trivially. Hence T is maximal monotone by
Theorem 12. 0

Theorem 14 (Equilibrium Problem 1 as a Zero Problem)

(a) (primal case). Equilibrium Problem 1 can be identified with (ZP) in the case
of z = x and the mapping

if Ex = b,
if Ex"# b.
In the maximal monotone version of Equilibrium Problem 1, T is maximal monotone
as long as there exists a flow x with Ex = b and x jEri ( dom Rj) for all j E J.
(b) (dual case). Equilibrium Problem 1 can be identified with (ZP) in the case of
z = (v, u) and the mapping

if z E L,
if z ~ L,
288

where L = {(v,u) I v = -ETu}, so that L.l. = {(x,y) I y = Ex}. In the maximal

monotone version of Equilibrium Problem 1, T is maximal monotone as long as there
exists a potential it with tension v such that Vj E ri (rge R j ) for all j E J.
(c) (primal-dual case). Equilibrium Problem 1 can be identified with (ZP) in the
case of z = (x,u) and the mapping

In the maximal monotone version of Equilibrium Problem 1, T is maximal monotone.

Proof. We specialize Theorem 13 here to (3.4). o

REFERENCES

[1] R. T. Rockafellar, "Convex Analysis", Princeton University Press, Princeton, NJ, 1970.
[2] R. T. Rockafellar, "Lagrange multipliers and optimality," SIAM Review 35, 183-238, 1993.
[3] R. T. Rockafellar, "On the maximality of sums of nonlinear monotone operators," Transactions
Amer. Math. Soc. 149, 75-88, 1970.
[4] G. J. Minty, "Monotone networks," Proc. Roy. Soc. London A 257,194-212,1960.
[5] R. T. Rockafellar, "Network Flows and Monotropic Optimization", Wiley, New York, 1984.
[6] J. J. Moreau, "Proximite et dualite dans un espace hilbertien," Bull. Soc. Math. France 93,
273-299, 1965.
[7] G. J. Minty, "On the maximal domain of a 'monotone' function," Michigan Math. J. 8,135-137,
1961.
GENERALIZED MONOTONICITY - CONCEPTS AND USES

Schaible Siegfried

Graduate School of Management

University of California
Riverside, CA 92521, U.S.A.
e-mail: schaible@ucracl.ucr.edu

Abstract. The role of monotonicity of the operator in Complementarity Problems

and Variational Inequality Problems corresponds to the role of convexity of the objec-
tive function in Mathematical Programming Problems. Various concepts of generalized
monotonicity are presented which are related to generalized convexity concepts. Spe-
cial characterizations in case of differentiable operators can be obtained. Using some
of these concepts, two sets of existence results are given for generalized monotone Vari-
ational Inequality Problems in Banach spaces. In the paper with Hadjisavvas recent
existence results by Cottle and Yao for pseudomonotone Complementarity Problems in
Hilbert spaces are extended to the general case of quasimonotone Variational Inequality
Problems in Banach spaces. In the work with Yao the equivalence of strictly mono-
tone Complementarity Problems, least element problems and other related problems in
Banach lattices, studied by Riddel, is extended to the strictly pseudomonotone case.

Key words. Quasimonotone map, (strictly) pseudomonotone map, Variational Inequal-

ity Problem, least element problem.

1. INTRODUCTION

Monotone operators play an important role in Complementarity Problems and Vari-

ational Inequality problems, much like convex functions in mathematical programming
problems. Whereas a theory of generalized convex functions has already been quite
well established (see e.g. [1]), various concepts of generalized monotone operators have
only very recently emerged, starting with the paper by Karamardian and Schaible [27]
in 1990. Even more recently is the study of their uses in Variational Inequalities. To
see the rapid progress in the area of generalized monotonicity, the reader is invited to
compare this survey with an earlier one in [46].
In the first part of this paper we present nine kinds of (generalized) monotone maps

289
290

as well as characterizations in the differentiable case. This is followed by two sets of

existence results for generalized monotone Variational Inequality Problems in the second
part, summarizing major results by Hadjisavvas and Schaible [19] and Schaible and Yao
[48].

2. CONCEPTS OF GENERALIZED MONOTONICITY

We begin with a brief description of three models in order to provide the necessary
context for generalized monotone maps. These models have been extensively analyzed
and applied in the second part of this century, e.g. [6], [7], [8], [15], [20], [38], [46].

2.1 Background: three models

The Mathematical Programming Problem is given by

MP min{f(x)lx E C} (2.1)
where f : C -+ R for C ~ Rn. Numerous applications are found in management, eco-
nomics, applied mathematics, statistics, natural sciences and engineering, for example.

The Complementarity Problem can be stated as follows:

CP find x E C such that F(x) E C* , x T F(x) =0 (2.2)
where F : C -+ R n for a cone C ~ Rnj C* denotes the dual cone. In addition to the
Karush-Kuhn-Tucker optimality conditions for differentiable MP a variety of equilib-
rium problems in economics, game theory and mechanics give rise to CPo

The Variational Inequality Problem can be stated as follows:

VIP find x E C such that (x - x)T F(x) ~ 0 for all x E C (2.3)
where F : C -+ R for a set C ~ R which is not necessarily a cone as in CPo Many
n n

applications in the natural sciences, often in infinite dimensional spaces, and economic
equilibrium problems as well as problems in a management science give rise to VIP.

The following relationship exists between these models in case of differentiable MP

(2.4)

Under certain regularity assumptions theoretical as well as algorithmic results for these
models can be established. Classical assumptions are, in addition to closedness and
convexity of C:
for MP: f is convex
for CP/VIP: F is monotone.

Under these assumptions the models above have many nice properties, and algorithms
converge to a solution. Indeed, in many applications these convexity /monotonicity con-
ditions hold.
 291

However there are also many applications were these classical assumptions fail to be true.
We realize that they are only sufficient conditions which are not necessary in the theory
of MP, CP and VIP. In case of MP, much progress has been made to develop various
kinds of generalized convex /unctions and demonstrate their usefulness in application,
see [1), [3), [33), [49), [50). Recently various kinds of generalized monotone maps have
been introduced and first uses in CP /VIP have been shown. Most of the work so far has
been conceptual [2), [4), [5), [10), [11), [12), [16), [17), [18), [22), [23), [24), [25), [26), [27),
[28), [29), [30), [31), [32), [33), [35), [36), [39), [40), [46], [47], [54].
However first results demonstrating the usefulness of generalized monotonicity in CP /
VIP have already appeared [2), [9), [19), [20), [23), [24), [26), [48), [52), [53). Many of
these results are related to the existence and uniqueness of solutions.

2.2 Nine kinds of (generalized) monotone maps

In [18), [27) the following types of (generalized) monotone maps have been intro-
duced:
quasimonotone
i
monotone --+ pseudomonotone --+ semistrictly quasimonotone
i i i
strictly monotone --+ strictly pseudomonotone --+ strictly quasimonotone
i i
strongly monotone --+ strongly pseudomonotone
(2.5)
In the special case of a gradientmap, these properties correspond to the following (gen-
eralized) convexity properties of the underlying function [1) as shown in [18), [27):

quasi convex
i
convex --+ pseudoconvex --+ semistrictly quasi convex
i i i
strictly convex --+ strictly pseudo convex --+ strictly quasi convex
i i
strongly convex --+ strongly pseudoconvex
(2.6)
We remark that equivalence between strong pseudoconvexity and strong pseu-
domonotonicity of the gradient only holds if the new definition in [18) is used. For
the more restrictive definition in [27) strong pseudomonotonicity of the gradient implies
strong pseudoconvexity of the underlying function, but the reverse is not true, as shown
in [18).

We now give the definitions of (generalized) monotone maps [18), [26), [27).
For the definition of (generalized) convex functions and their characteristic geometrical
properties see [1). We assume F: C -+ R n where C ~ Rn.

F is monotone (m) on C if for all x, y E C

(y - xf(F(y) - F(x» ~ o. (2.7)

292

F is strictly monotone (s.m) on C iffor all x, y E C , x =f. y

(y - xf(F(y) - F(x» > O. (2.8)

F is strongly monotone (str.m) on C if there exists (3 > 0 such that for all x, y E C

(2.9)

F is pseu.domonotone (pm) on C iffor all x, y E C

(y-xfF(x)~O implies (y-x)TF(y)~O (2.10)

which is equivalent to

(y - xfF(x) > 0 implies (y - x)TF(y) > O.

F is strictly pseudomonotone (s.pm) on C if for all x, y E C , x =f. y

(y - xfF(x) ~ 0 implies (y - x)TF(y) > O. (2.11 )

F is quasimonotone (qm) on C iffor all x,y E C

(y - x)TF(x) > 0 implies (y - x)TF(y) ~ O. (2.12)

F is strongly pseudomonotone (str.pm) on C (open) if for all x E X and vERn such

that IIvll = 1 , vTF(x) = 0 there exists (3 > 0, f> 0 such that

(2.13)

This definition in [18] replaces the more restrictive one in [27].

F is strictly quasimonotone (s.qm) on C if F is quasimonotone and for all x, y E C, x =f. y

there exists z = tx + (1 - t)y, t E (0,1) such that

(2.14)

F is semistrictly quasimonotone (ss.qm) on C if F is quasimonotone and for all x, y E

C, x =f. y
(y-xfF(x»O implies (y-x)TF(z»O (2.15)

for some z = tx + (1- t)y, t E (0, t).

If F is continuous, quasimonotonicity of F does not have to be required explicitly for
s.qm and ss.qm maps since it is implied by (2.14), (2.15), respectively.

As mentioned before, in case of a gradient map F = 'V f where f is a differentiable

function on an open convex set C, a particular kind of (generalized) monotonicity of
F holds if and only if the underlying function f has the corresponding (generalized)
convexity property; e.g. F is pseudomonotone if and only if f is pseudoconvex, etc.
 293

Extensions of these characterizations of differentiable functions through their gra-

dient to the nondifferentiable case have been obtained in [2), [22], [25], [30), [31), [32],
[35), [36), [39).

2.3 Characterizations of generalized monotone maps

In the special case of a differentiable map F on an open convex set C ~ IR n first-

order characterizations in terms of the Jacobian JF(X) have been established [28], [29].
Consider the following conditions, given x E C, vERn:

A: vTF(x) = 0 (2.16)

B : v T F(x) = v T h(x)v = 0 implies there exists t > 0 such that

i<o, vTF(x+iv»o x+tvEC, vTF(x+tv)~O (2.17)
for all 0::; t ::; t
implies there exists t > 0 such that
x+tvEC, vTF(x+tv)~O (2.18)
for all 0::; t ::; t.
It can be shown [29]:

Theorem 1. Let F be differentiable on an open convex set C ~ Rn. F is qm on C if

and only if A and B hold. F is pm on C if and only if A and C hold.

Additional characterizations of various types of generalized monotone maps are

obtained in [28], [29) as well. See also [4] for a related, but different approach.

The characterization in Theorem 1 have very recently been strenghtened in [10] by

replacing the assumptions Band C by weaker ones. In [37] the results in [29) and in [10]
have been extended to nonsmooth, locally Lipschitz functions. For this the Jacobian has
been replaced by the generalized Jacobian in the sense of Clarke.

To further specialize, affine maps F(x) = Mx + q can be studied where M is an

n x n matrix and q ERn. F is a gradientmap V f if and only if M is symmetric
in which case f is a quadratic. Hence generalized monotonicity results for affine maps
extend classical characterizations of generalized convex quadratic functions [1], [14], [43],
[44J, [45J. A number of characterizations of generalized monotone affine maps have been
derived in [28], [29]. For instance, condition A above is both necessary and sufficient for
pseudomonotonicity and quasimonotonicity. For additional results see [40].
 294

3. GENERALIZED MONOTONE VARIATIONAL INEQUALITY

PROBLEMS

We now turn to two applications of generalized monotone maps in the context of

Variational Inequality Problems and related models in 2.1. These results have been
obtained very recently in [19] and [48]. They are derived in Banach spaces. Obviously,
generalized monotonicity concepts in Section 2 can be extended to Banach spaces. The
results in [19] and [48] deal with the existence and uniqueness of solutions of Variational
Inequality problems.

3.1 Quasimonotone Variational Inequality Problems [19].

Let B be a real Banach space, B* its dual and (x,u) the primal-dual pairing
for x E B, u E B*. Furthermore, let K ~ B be a nonempty closed convex set and
F : K -+ B*. We will study the Variational Inequality Problem:

VIP find x E K such that (x - x, F(x)) ~ 0 for all x E K. (3.1)

According to the classical result by Hartman and Stampacchia [21] in 1966 there exists a
solution of the VIP if B = R,n , K is compact and F is continuous. This existence result
has been extended in many ways, often assuming monotonicity of F. For monotone VIP
the following properties hold:
- the solution set is convex (it may be empty)
- a solution is unique if F is strictly monotone
- there exists a unique solution if F is strongly monotone.
Also, most algorithms for VIP are derived under the monotonicity assumption.

Turning now to nonmonotone VIP, Karamardian [26] proved the following existence
result for Complementarity Problems:
- the VIP (3.1) has a solution if B = R,n , K ~ R,n is a pointed solid closed convex
cone, there exists x E K such that F( x) E int K*, where K* is the dual of K, and
F is continuous and pseudomonotone.
This result, overlooked for some time, has recently been extended in several ways,
always assuming pseudomonotonicity of F. According to [20], K does not have to be
a cone, i.e. the existence result holds not only for Complementarity Problems, but for
Variational Inequality problems. Cottle and Yao [9] extend Karamardian's result to
Complementarity Problems in Hilbert space, at the same time relaxing the continuity
assumption to continuity on finite dimensional subspaces. Yao [53] extends these results
further to reflexive Banach spaces. In [52] F is allowed to be multivalued, i.e. existence
of solutions is established for generalized VIP.

In [19] Hadjisavvas and Schaible extend Karamardian's result [26] in three direc-
tions: a) B is a reflexive Banach space, b) F is quasimonotone, but not necessarily
pseudomonotone, and c) K has an "inner" point, but not necessarily an algebraic inte-
rior point. The proofs of the existence results in [19] differ significantly from those in
 295

previous extensions of Karamardian's existence result. The main reason for this is that
quasimonotone maps form a much larger class than pseudomonotone maps so that new
proof techniques are needed.

In [19) the following four types of generalized monotone maps are used:
pseudomonotone ---+ quasimonotone
1 1
strictly pseudomonotone ---+ striclty quasimonotone.
In the result in (19) the existence of a socalled inner point is assumed, replacing the
more restrictive assumption of the existence of a relative algebraic interior point. A
point Xo E B is called inner point of a nonempty set D ~ B iffor all u E B* \ {O}

(x,u):::; (xo,u) for all xED implies (x,u) = (xo,u) for all xED. (3.2)

A relative algebraic interior point is an inner point, but the converse is not true. It can
be shown that all nonempty separable closed convex sets have inner points. They may
however not have algebraic interior points; e.g. important subsets in CP(1 :::; p < 00).

The main result in (19) is:

Theorem 2. Let B a real, reflexive Banach space. Consider a nonempty closed convex
set K ~ B which has an inner point and which is either bounded or there exists p > 0
such that for all x E K with IIxll ~ p there exists Y E K satisfying IIYII < p and
(x - Y , F( x» ~ O. Let F : K -+ B* be hemicontinuous and quasimonotone. Then the
VIP (3.1) has a solution.
This extends not only Karamardian's result in (26), but also the extensions of it in (9), (20)
and [53). The Hilbert space is replaced by a reflexive Banach space, pseudomonotonicity
by quasimonotonicity, and the continuity and interior point assumptions are weakened.

In (19) additional existence results for quasimonotone VIP are obtained. Further-
more, properties of the set of solutions of (3.1) and sufficient conditions for the unique-
ness of a solution are established. An application to a problem in equilibrium analysis
is presented as well.

We now turn to a second application of generalized monotonicity.

3.2 Equivalence of strictly pseudomonotone Complementarity Problems and

least element problems [48].

As above, let B be a real Banach space. Consider a nonempty closed convex cone
K ~ B which induces the partial order

x :::; Y for x, Y E B if Y - x E K. (3.3)

Correspondingly, a partial order is introduced in B* by the dual cone K*. Points x E

K, x#-O and u E K*, u #- 0 are called positive. A point u E K* is strictly positive if
(x,u»OforallxEK, x#-O.
296
The space B is called a vector lattice with respect to ::; if each x, y E B has a unique
infimum (meet) x A y such that

x A Y ::; x , x A Y ::; Y and [z::; x, z::; y :::} z ::; x A y].

Consider T: K -+ B* , f: K -+ R. The set :F = {x E B I x E K, T(x) E K*} is called

the feasible set of T. We consider the following five problems:

(I) nonlinear program (u E B* given):

find x E :F such that (x, u) = min{(x, u) Ix E :F} j
(II) least element problem: find x E :F such that x ::; x for all x E :Fj

(III) Complementarity Problem: find x E :F such that (x, T(x)) = OJ

(IV) Variational Inequality Problem: find x E K such that (x - x, T( x)) :?: 0

for all x E Kj
(V) unilateral minimization problem (given I):

find xEK such that f(x) = min{f(x) I x E K} .

For strictly monotone operators T the equivalence of the above five problems has
been established by ruddell in [42] under certain regularity assumptions. His work
extends earlier results by Stampacchia [51], Lions and Stampacchia [34] and Cryer and
Dempster [13]. In [48] Schaible and Yao extend ruddell's results further from strictly
monotone to strictly pseudomonotone operators.
The map T is called Z-map if

(z,T(x) - T(y))::; 0 whenever (x - y) A z = O. (3.4)

In the special case of B = R n, K = R+. and T linear this is equivalent to offdiagonal

anti tonicity [41].
The map T is called positive at infinity iffor all x E K there exists p( x) > 0 such that

(y - x, T(y)) > 0 for all y E K, lIyll :?: p(x).

The results in [48) make use the following types of generalized monotone maps:

pseudomonotone
T
strictly pseudomonotone
T
strongly pseudomonotone (as in [27]).
Using an existence result by Yao in [52] for certain Variational Inequalities, the following
extension of ruddell's result to strictly pseudomonotone operators is derived in [48]:

Theorem 3. Let B be a real, reflexive Banach space, K ~ B a nonempty closed convex

cone such that B is a vector lattice. Consider a hemicontinuous, strictly pseudomonotone
 297

Z-map T : K -+ B* which is positive at infinity. Let u E K* be strictly positive. Then

there exists x E F which is a solution of (I), (II), (III), (IV), and x is unique. If T is the
Gateaux-derivative of f : K -+ R, then x is also the unique solution of (V).

In [48] additional results on the equivalence of the above five problems under generalized
monotonicity of the operator are derived.

REFERENCES

[1) M. Avriel, W.E. Diewert , S. Schaible and I. Zang, "Generalized concavity", Plenum Publishing
Corporation, New York, 1988.
(2) M. Bianchi, "Pseudo P-monotone operators and Variational Inequalities", Working Paper No.
6, Istituto di Matematica Generale, Finanziaria ed Econometrica, Universitit Cattolica, Milano,
1993.

(3) A. Cambini, E. Castagnoli, L. Martein, P. Mazzoleni and S. Schaible (eds), "Generalized convexity
and fractional programming with economic applications", Springer-Verlag, Berlin-Heidelberg-
New York, 1990.

(4) E. Castagnoli and P. Mazzoleni, "Order-preserving functions and generalized convexity", Rivista
di Matematica per Ie Scienze Economiche e Sociali 14, 33-45, 1991.

[5) E. Castagnoli and P. Mazzoleni, "Orderings, generalized convexity and monotonicity", in: S.
Komlosi, T. Rapcsak and S. Schaible (eds.) "Generalized convexity", Proceedings, Pees. Hungary
1992, Springer-Verlag, Berlin-Heidelberg-New York, 250-262, 1994.
[6) R.W. Cottle, F. Giannessi and J.L. Lions (eds.) "Variational Inequalities and Complementarity
Problems: theory and applications" , J. Wiley, New York, 1980.

[7) R.W. Cottle, J. Kyparisis and J.S.Pang (eds.), "Variational inequality problems", Mathematical
Programming, Series B, 48, No.2, 1990.

[8] R.W. Cottle, J. Kyparisis and J.S.Pang (eds.), "Complementarity problems", Mathematical Pro-
gramming, Series B, 48, No.3, 1990.

[9) R.W. Cottle and J.S. Yao, "Pseudomonotone complementarity problems in Hilbert space", J. of
Optimization Theory and Applications 75, 281-295, 1992.

[10) J.P. Crouzeix and J .A. Ferland, "Criteria for differentiable generalized monotone maps", M athe-
matical Programming (to appear), 1993.

[11] J.P. Crouzeix and A. Hassouni, "Quasimonotonicity of separable operators and monotonicity in-
dices", Working Paper, Mathematiques Appliquees, Universite Blaise Pascal, Clermont-Ferrand,
1992.

[12) J.P. Crouzeix and A. Hassouni, " Generalized monotonicity of a separable product of opera-
tors: the multivalued case", Working Paper, Mathematiques Appliquees, Universite Blaise Pascal,
Clermont-Ferrand, 1993.

[13] C.W. Cryer and M.A.H. Dempster, "Equivalence of linear Complementarity Problems and linear
programs in vector lattice Hilbert spaces", SIAM J. Control Optimization 18, 76-90, 1980.

[14] J .A. Ferland, "Quasi-convex and pseudo-convex functions on solid convex sets" , Technical Report
71-4, Department of Operations Research, Stanford University, 1971.
298

[15] F. Giannessi and F. Niccolucci, "Connections between nonlinear and integer programming prob-
lems". Symposia Mathematica, Vol. XIX, Academic Press, 160-175,1976.
[16] M.S. Gowda, "Pseudomonotone and copositive star matrices, Linear Algebra and Its Applications
113, 107-118, 1989.
[17] M.S. Gowda, "Affine pseudomonotone mappings and the linear Complementarity Problem", SIAM
J. of Matrix Analysis and Applications 11,373-380, 1990.
[18] N. Hadjisavvas and S. Schaible, "On strong pseudomonotonicity and (semi) strict quasimonotonic-
ity", J. of Optimization Theory and Applications 79, 139-155, 1993.
[19] N. Hadjisavvas and S. Schaible, "Quasimonotone Variational Inequalities in Banach spaces" , Work-
ing Paper 94-01, Graduate School of Management, University of California, Riverside, 1994.
[20] P.T. Harker and J.S. Pang, "Finite dimensional Variational Inequality and nonlinear Complemen-
tarity Problems: a survey of theory, algorithms and applications", in: R.W. Cottle, J. Kyparisis
and J.S. Pang (eds.), "Variational inequality problems", Mathematical Programming 48, Series E,
161-220, 1990.
[21] P. Hartman and G. Stampacchia, "On some non-linear elliptic differential-functional equations",
Acta Math. 115,271-310, 1966.
[22] A. Hassouni, "Sous-differentiels des fonctions quasi-convexes" , These de 3 erne Cycle, Mathemati-
ques Appliquees, Universite Paul Sabatier, Toulouse, 1983.
[23] A. Hassouni, "'Quasimonotone multifunctions, applications to optimality conditions in quasiconvex
programming", Numerical Functional Analysis and Optimization 13,267-275, 1992.
[24] A. Hassouni, "Operateurs quasimonotone, applications a certain problemes variationnels" , These,
Mathematiques Appliquees, Universite Paul Sabatier, Toulouse, 1993.
[25] A. Hassouni and R. Ellaia, "Characterizations of nonsmooth functions through their generalized
gradients", Optimization 22, 401-416, 1991.
[26] S. Karamardian, "Complementarity over cones with monotone and pseudomonotone maps", J. of
Optimization Theory and Applications 18,445-454, 1976.
[27] S. Karamardian and S. Schaible, "Seven kinds of monotone maps", J. of Optimization Theory and
Applications 66, 37-46, 1990.
[28] S. Karamardian and S. Schaible, "First-order characterizations of generalized monotone maps" ,
Working-Paper 90-5, Graduate School of Management, University of California, Riverside, 1989.
[29] S. Karamardian, Schaible S. and J.P. Crouzeix , "Characterizations of generalized monotone
maps", J. of Optimization Theory and Applications 76, 399-413, 1993.
[30] S. Komlosi, "On generalized upper quasidifferentiability", in: F. Giannessi (ed.), "Nonsmooth
optimization methods and applications". Gordon and Breach, Amsterdam 189-200, 1992.
[31] S. Komlosi, "Generalized monotonicity in non-smooth analysis", in: S. Komlosi, T. Rapcsak and
S. Schaible (eds.), "Generalized convexity", Proceedings, Pees. Hungary 1992, Springer-Verlag,
Berlin-Heidelberg-New York, 263-275, 1994.
[32] S. Komlosi, "Generalized monotonicity and generalized convexity", J. of Optimization Theory and
Applications (to appear), 1993.
[33J S. Komlosi, T. Rapcsak and S. Schaible (eds.), "Generalized convexity", Proceedings, Pees. Hun-
gary 1992, Springer-Verlag, Berlin-Heidelberg-New York 1994.
[34] J.L. Lions and G. Stampacchia, "Variational inequalities", Comm. Pure Appl. Math. 20,493-519,
1967.
 299

[35] D.T. Luc, Characterizations of quasiconvex functions", Bulletin of the Australian Mathematical
Society 48, 393-405, 1993.
[36] D.T. Luc, "On generalized convex nonsmooth functions", Bulletin of the Australian Mathematical
Society 49, 139-149, 1994.
[37] D.T. Luc and S. Schaible, "Generalized monotone nonsmooth maps", Working Paper, Graduate
School of Management, University of California, Riverside, 1994.
[38] A. Maugeri, "Convex programming, Variational Inequalities and applications to the traffic equi-
librium problem", Appl. Math. Optim. 16, 169-185, 1987.
[39] J.P. Penot and P.H. Quang, "On generalized convex functions and generalized monotonicity of
set-valued maps", J. of Optimization Theory and Applications, (to appear), 1993.
[40] R. Pini and S. Schaible, ''Invariance properties of generalized monotonicity", Optimization, 28,
211-222, 1994.
[41] W.C. Rheinboldt, "On M-functions and their application to nonlinear Gauss-Seidel iterations and
network flows", J. Math. Anal. Appl. 32, 274-307, 1970.
[42] R.C. Riddel, "Equivalence of nonlinear complementarity problems and least element problems in
Banach lattices", Mathematics of Operations Research 6, 462-474, 1981.
[43] S. Schaible, "Beitraege zur quasikonvexen Programmierung", Doctoral Dissertation, Universitaet
Koeln, 1971.
[44] Schaible S., "Quasiconvex, pseudoconvex and strictly pseudoconvex quadratic functions", J. of
Optimization Theory and Applications 35, 303-338, 1981.
[45] S. Schaible, "Generalized convexity of quadratic functions", in: S. Schaible and W.T. Ziemba
(eds.) , "Generalized concavity in optimization and economics". Academic Press, New York, 183-
197, 1981.
[46] S. Schaible, "Generalized monotone maps", in: F. Giannessi (ed.), "Nonsmooth optimization
methods and applications", Proceedings, Erice 1991, Gordon and Breach, Amsterdam, 392-408,
1992.
[47] S. Schaible, " Generalized monotonicity - a survey", in: Komlosi S., Rapcsak T. and S. Schaible
(eds.), "Generalized convexity", Proceedings, Pees. Hungary 1992, Springer-Verlag, Berlin-Heidel-
berg-New York, 229-249, 1994.
[48] S. Schaible and J .C. Yao, "On the equivalence of nonlinear Complementarity Problems and least
element problems", Mathematical Programming (to appear), 1993.
[49] S. Schaible and W.T. Ziemba (eds.), "Generalized concavity in optimization and economics",
Academic Press, New York, 1981.
[50] C. Singh and B.K. Dass (eds.), "Continuous-time, fractional and multiobjective programming",
Analytic Publishing Company, New Delhi, 1989.
[51] G. Stampacchia, "Le probleme de Dirichlet pour les equations elliptiques du second ordre a coef-
ficients discontinus", Ann. Inat. Fourier (Grenoble) 15, 189-258, 1965.
[52] J .C. Yao, "Multi-valued Variational Inequalities with K-pseudomonotone operators", J. of Opti-
mization Theory and Applications 83, 1994 (to appear).
[53] J .C. Yao, "Variational inequalities with generalized monotone operators", Mathematics of Opera-
tions Research (to appear), 1993.
[54] D.L. Zhu and P. Marcotte, "New classes of generalized monotonicity", J. of Optimization Theory
and Applications (to appear), 1994.
CONTRIBUTORS

Antoni Carla National Institute for Advanced Mathematics "F. Severi"

(INDAM),Citta Universitaria, Roma, Italy; and Naval Aca-
demy, Vi ale Italia, Livorno, Italy.
e-mail: ANTONI@BIANCHLDM.UNIPI.IT
Bonuccelli Maurizio Univ. of Roma "La Sapienza", Computer Science Depart-
ment, Via Salaria 113, 1-00198 Roma, Italy.
e-mail: ALGOR@VM.CNUCE.CNR.lT
Castellani Marco Dept. of Mathematics, Univ. Pisa, Via Buonarroti 2, 56127
Pisa, Italy.
e-mail: CASTELLA@GAUSS.DM.UNIPI.IT
Cubiotti Paolo Dept. of Mathematics, Univ. di Messina, 98166
Sant' Agata-Messina, Italy.
e-mail: PCUB@IMEUNIV.UNIME.lT
De Luca Marino Dept. of Mathematics, Univ. of Palermo, Via Archirafi, 34
- 90123 Palermo, Italy.
e-mail: DELUCA@CSIINS.CSII.UNIRC.lT
Elster Karl-Heinz European Laboratory for Intelligent Techniques Engineering
Aachen, Germany.
Elster Rosalind Overseas Publishers Association 10405 Berlin, Germany
Facchinei Francisco Dept. of Sistems and Computer Science, Univ. of Roma "La
Sapienza", Via Buonarroti 12, 00185 Roma, Italy
e-mail: SOLER@PEANO.DIS.UNIROMAl.IT
Ferrari Paolo Dept. of Road and Transport, Univ. Pisa Via Diotisalvi 2,
56126 Pisa, Italy. Fax: ++3950555057.
Fukushima Masao Nara Inst. of Science and Technology, Graduate School of
Information Science, Ikoma, Nara 630-01, Japan.
e-mail: FUKU@IS.AIST-NARA.AC.JP
Giannessi Franco Dept. of Mathematics, Univ. of Pisa, Via Buonarroti 2,
56127 PISA, Italy.
e-mail: GIANNESS@GAUSS.DM.UNIPLIT
Gwinner Joachim Institute of Mathematics, Depart. of Aerospace Engeene-
ring, Universitaet de Federal Army, Wener-Heisenberg-Wey,
85577 Neubiberg, Germany.
e-mail: L11BGWIN@RZ.UNIBW-MUENCHEN.DE
Ibaraki Toshihide Department of Applied Mathematics and Physics Faculty of
Engineering, Kyoto University, Kyoto 606, Japan.

301
302

Iwaoka Kouchiro I. Department of Applied Mathematics and Physics Faculty of

Engineering, Kyoto University, Kyoto 606, Japan.
Koksal Muhammet Engeenering Faculty Inonu University, 44100-Malatya, Tur-
kie. e-mail: INONU01@TREARN.BITNET
Larsson Torbjorn Division of Optimization Department of Mathematics, Lin-
koping Institute of Technology, S-581 83 Linkoping, Sweden.
e-mail: TL@MATH.LIU.SE
Marcotte Patrice Dept. of Informatique et Rech. Op., Univ. de Montreal, CP
6128, Succ. Centre-Ville, Montreal, H3C 3J7 Canada.
e-mail: PATRICEM@CRT.UMONTREAL.CA
Mastroeni Giandomenico Dept. of Mathematics, Univ. Pisa, Via Buonarroti 2, 56127
Pisa, Italy.
Maugeri Antonino Dept. of Mathematics, Univ. of Catania, Viale A. Doria 6,
95125 Catania, Italy.
e-mail: MAUGERI@DIPMAT.UNICT.lT
Murthy Venkatesha M.K. Dept. of Mathematics, Univ. of Pisa, Via Buonarroti 2,
56127 Pisa, Italy.
e-mail: MURTHY@GAUSS.DM.UNIPI.IT
Oettli Werner Facultat fur Mathematik und Informatik, 68131 Mannheim,
Germany.
Patriksson Michael Dept. of Mathematics, Linkoping Inst. of Technology, 58183
Linkoping, Sweden.
e-mail: MIPAT
Rapcsak Tamas Hungarian Academy of Sciences, Computer and Automation
Institute, Lagymanyosi utca 11, 1111 Budapest,
Hungary.
e-mail: H632RAP@ELLA.HU
Ricceri Biagio Dept. of Mathematics, Univ. of Catania, Vi ale A. Doria 6,
95125 Catania, Italy.
e-mail: RICCERI@DIPMAT.UNICT.lT
Robinson Sthephen M. Dept. of Industrial Engineering, Univ. of Wisconsin at
Madison, 1513 University Avenue, Madison, WI 53706-1572,
USA
e-mail: SMR@CS.WISC.EDU
Rockafellar Tyrrell R. Univ. of Washington, Dept. of Mathematics, Seattle, WA
98195, USA.
e-mail: RTR@MATH.WASHINGTON.EDU
Schaible Siegfried Graduate School of Management, Univ. of California at
Riverside, Riverside, CA 95521, USA.
e-mail: SCHAIBLE@UCRACl.UCR.EDU
Yen Nguyen Dong Vietnamese Academy of Sciences, Institute of Mathematics,
P.O. Box 631 BO HO, 10000 Hanoi, Vietnam.
INDEX

Asymptotic cone 127 Function( s)

Bilevel programming 187, 190 copositive 229
cost 86, 93
Cauchy-Dirichlet problem 218
demand 88 gap 1,4, 5, 6, 72, 101,
Conservation law 213
103, 180, 181, 198, 201, 272
Control systems 39
hemicontinuous 126
Capacity constraints 85, 86, 90, 92, implicit 264
93,95 Lagrangian 1, 3, 73
Clarke Jacobian 71 merit 69, 72, 73
Complementarity problems 6, 30, 69, monotone 71, 124
P-71
102, 225, 227, 235, 289, 295
penalty 95, 96, 188
linear 225, 231
q-regular 229
quasi- 104
semi coercive 126
nonlinear 235
strictly monotone 71
Conic extension 104
strong gap 108, 202
Contact problems 134 strongly monotone 71
Dual(ity) 5, 21, 27, 115, 283 weak gap 104, 202
Fenchel4, 26 Gap function(s) 1,4,5,6,72, 101, 103
Lagrange 4, 27
180, 181, 198, 201, 272
problem 4, 23, 25
strong 108, 202
Variational Inequality I, 4, 22, 24,
weak 104, 202
203
Geometric
Wolfe 6, 27, 118
optimization 55, 63
Equilibrium 81,85,86,89,93,113,184, vector inequality 61
197, 225, 227, 271, 275, 276, Hierarchical systems 15
277, 278, 285 Image 104, 201
Cournot-Nash 184 transformed 104
homogeneous 225, 227
Implicit function 264
market 131
Integrability property 25
solution 89
traffic 45, 81, 169 Kuhn--Tucker
conditions 28, 273
Existence theorems 251
point 94

303
304
Lagrangian 73
duality 27
function 1, 3, 30, 73
multipliers 86, 94, 99, 100
Least element problems 295
Map(ping)
co-coercive 180
monotone 180, 277, 285, 289
pseudo co-coercive 180
pseudomonotone 180, 289
quasimonotone 289
semicontinuous 226
strictly monotone 180
strongly monotone 180
Matrix
double stochastic 11
pseudo-regular 226
Ro- 226
switching 10, 13
Merit function 69, 72, 73
Multifunctions 34-36
Multipliers 86, 94, 99, 100, 144
Network
equilibrium 271, 275
flow 9, 113, 195
interconnection 9
permutation 10
switched 155
traffic 196
transport 85
Normal map 257, 258, 260
O1igopoly models 181
Paradoxis 206
Potentials 281
Projection methods 203
Proximal point 143
primal-dual algorithm 143
Quasi-Complementarity System 104
Quasi-Variational Inequalities 33, 45,
101, 102, 195, 197,251
Regularity 70
Saddle point 24, 25
Schur complement 70, 262
Sensitivity analysis 206, 257, 264
Separation 3, 103, 202
functions 23, 103
of sets 1, 101
strong 3, 202
weak 3,202
Signorini problem 135, 138
Stability 123, 127, 134, 138
Switched networks 155
Switching modes 15
Telecommunication systems 9
Tensor approximations 235
Theorem
Birkhoff-von Neumann 11
Traffic assignment 151, 175, 184
Traffic equilibrium 45, 81, 169
Transponders 13
Unilateral constraints 213
Variational Inequality 1, 21, 24, 55, 91,
92, 124, 132, 143, 179, 180,
195, 213, 251, 257,258,271,
272, 278, 289, 294
asymmetric 272
dual 24, 203
elliptic 134, 140
monotone 123, 124, 294
Quasi- 33,45, 101, 102, 195, 251
stability of 123
symmetric 272
vector 55, 58, 61, 63
 305

Vector optimization 55, 63

Complementarity 55 Variational Inequality 58
Fenchel conjugate 60 Wardrop principle 170, 172, 184
geometric inequality 61