Calculus of Variations PDF

CAMBRIDGE STUDIES IN
ADVANCED MATHEMATICS 64
EDITORIAL BOARD
D.J.H. GARLING, W. FULTON, K. RIBET, T. TOM DIECK,
P. WALTERS
CALCULUS OF VARIATIONS
Already published
1 W . M . L . Holcombe Algebraic automata theory
2 K . Petersen Ergodic theory
3 P.T. Johnstone Stone spaces
4 W,H. Schikhof Ultrametric calculus
5 J.-P. Kahane Some random series of functions, 2nd edition
6 H. Cohn Introduction to the construction of class fields
7 J . Lambek & P.J. Scott Introduction to higher-order categorical logic
8 H. Matsumura Commutative ring theory
9 C . B . Thomas Characteristic classes and the cohomology of finite groups
10 M. Aschbaeher Finite group theory
11 J . L . Alperin Local representation theory
12 P. Koosis The logarithmic integral I
13 A. Pietsch Eigenvalues and S-numbers
14 S.J. Patterson An introduction to the theory of the Riemann
zeta-function
15 H.J. Baues Algebraic homotopy
16 V.S. Varadarajan Introduction to harmonic analysis on semisimple
Lie groups
17 W. Dicks & M. Dunwoody Groups acting on graphs
18 L . J . Corwin & F.P. Greenleaf Representations of nilpotent Lie groups
and their applications
19 R. Pritsch & R. Piccinini Cellular structures in topology
20 H Klingen Introductory lectures on Siegel modular forms
21 P. Koosis The logarithmic integral II
22 M.J. Collins Representations and characters of finite groups
24 H. Kunita Stochastic flows and stochastic differential equations
25 P. Wojtaszczyk Banach spaces for analysts
26 J . E . Gilbert & M.A.M. Murray Clifford algebras and Dirac operators
in harmonic analysis
27 A. Prohlich & M.J. Taylor Algebraic number theory
28 K . Goebel & W . A . Kirk Topics in metric fixed point theory
29 J . F . Humphreys Reflection groups and Coxeter groups
30 D.J. Benson Representations and cohomology I
31 D.J. Benson Representations and cohomology II
32 C . Allday & V . Puppe Cohomological methods in transformation groups
33 C . Soule et al Lectures on Arakelov geometry
34 A. Ambrosetti & G . Prodi A primer of nonlinear analysis
35 J . Palis & F . Takens Hyperbolicity and sensitive chaotic dynamics at
homoclinic bifurcations
36 M. Auslander, I. Reiten & S. Smalo Representation theory of Artin algebras
37 Y . Meyer Wavelets and operators
38 C . Weibel An introduction to homological algebra
39 W. Bruns & J . Herzog Cohen-Macaulay rings
40 V . Snaith Explicit Brauer induction
41 G . Laumon Cohomology of Drinfeld modular varieties I
42 E . B . Davies Spectral theory and differential operators
43 J . Diestel, H. Jarchow & A. Tonge Absolutely summing operators
44 P. Mattila Geometry of sets and measures in Euclidean spaces
45 R. Pinsky Positive harmonic functions and diffusion
46 G . Tenenbaum Introduction to analytic and probabilistic number theory
47 C . Peskine An algebraic introduction to complex projective geometry I
48 Y . Meyer & R. Coifman Wavelets and operators II
49 R. Stanley Enumerative combinatories
50 I. Porteous Clifford algebras and the classical groups
51 M. Audin Spinning tops
52 V . Jurdjevic Geometric control theory
53 H. Voelklein Groups as Galois groups
54 J . Le Potier Lectures on vector bundles
55 D. Bump Automorphic forms
56 G. Laumon Cohomology of Drinfeld modular varieties II
59 P. Taylor Practical foundations of mathematics
60 M. Brodmann & R. Sharp Local cohomology
64 J . Jost & X . Li-Jost Calculus of variations
Calculus of Variations
Jiirgen Jost and Xianqing Li-Jost

Max-Planck-Institute for Mathematics in the Sciences,
Leipzig
CAMBRIDGE
U N I V E R S I T Y PRESS
PUBLISHED BY T H E PRESS SYNDICATE OF T H E UNIVERSITY OF CAMBRIDGE
The Pitt Building, Trumpington Street, Cambridge C B 2 1RP, United Kingdom
CAMBRIDGE UNIVERSITY PRESS

The Edinburgh Building, Cambridge C B 2 2 R U , U K http://www.cup.ac.uk
40 West 20th Street, New York, N Y 10011-4211, U S A http://www.cup.org
10 Stamford Road, Oakleigh, Melbourne 3166, Australia
Cambridge University Press 1998
T h i s book is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.
First published 1998
Typeset in Computer Modern by the authors using I A l ^ X 2e
A catalogue record of this book is available from the British Library
Library of Congress Cataloguing in Publication data
Jost, Jurgen, 1956-

Calculus of variations / Jurgen Jost and Xianqing Li-Jost.
p. cm.
Includes index.
I S B N 0 521 64203 5 (he.)
1. Calculus of variations. I . Li-Jost, Xianqing, 1956-
I I . Title.
QA315.J67 1999
515'.64-dc21 98-38618 C I P
I S B N 0 521 64203 5 hardback
Transferred to digital printing 2003

Dedicated to Stefan Hildebrandt
Contents
Preface and summary page x

Remarks on notation xv
P a r t one: One-dimensional variational problems 1
1 T h e classical theory 3
1.1 The Euler-Lagrange equations. Examples 3
1.2 The idea of the direct methods and some regularity
results 10
1.3 The second variation. Jacobi fields 18
1.4 Free boundary conditions 24
1.5 Symmetries and the theorem of E. Noether 26
2 A geometric example: geodesic curves 32

2.1 The length and energy of curves 32
2.2 Fields of geodesic curves 43
2.3 The existence of geodesies 51
3 Saddle point constructions 62

3.1 A finite dimensional example 62
3.2 The construction of Lyusternik-Schnirelman 67
4 T h e theory of H a m i l t o n and J a c o b i 79
4.1 The canonical equations 79
4.2 The Hamilton-Jacobi equation 81
4.3 Geodesies 87
4.4 Fields of extremals 89
4.5 Hilbert's invariant integral and Jacobi's theorem 92
4.6 Canonical transformations 95
vii
viii Contents
5 D y n a m i c optimization 104
5.1 Discrete control problems 104
5.2 Continuous control problems 106
5.3 The Pontryagin maximum principle 109
P a r t two: Multiple integrals in the calculus of

variations 115
1 Lebesgue measure and integration theory 117

1.1 The Lebesgue measure and the Lebesgue integral 117
1.2 Convergence theorems 122
2 B a n a c h spaces 125
2.1 Definition and basic properties of Banach and Hilbert
spaces 125
2.2 Dual spaces and weak convergence 132
2.3 Linear operators between Banach spaces 144
2.4 Calculus i n Banach spaces 150
p
3 L and Sobolev spaces 159
p
3.1 L spaces 159
3.2 Approximation of LP functions by smooth functions
(mollification) 166
3.3 Sobolev spaces 171
3.4 Rellich's theorem and the Poincare and Sobolev
inequalities 175
4 T h e direct methods in the calculus of variations 183

4.1 Description of the problem and its solution 183
4.2 Lower semicontinuity 184
4.3 The existence of minimizers for convex variational
problems 187
4.4 Convex functional on Hilbert spaces and Moreau-
Yosida approximation 190
4.5 The Euler-Lagrange equations and regularity questions 195
5 Nonconvex functionals. Relaxation 205

5.1 Nonlower semicontinuous functionals and relaxation 205
5.2 Representation of relaxed functionals via convex
envelopes 213
6 T-convergence 225
6.1 The definition of T-convergence 225
Contents ix
6.2 Homogenization 231

6.3 T h i n insulating layers 235
7 BV-functionals and T-convergence: the example of

M o d i c a and M o r t o l a 241
7.1 The space BV{Q) 241
7.2 The example of Modica-Mortola 248
Appendix A The coarea formula 257

Appendix B The distance function from smooth hypersurfaces 262
8 Bifurcation theory 266

8.1 Bifurcation problems in the calculus of variations 266
8.2 The functional analytic approach to bifurcation theory 270
8.3 The existence of catenoids as an example of a bifurca
tion process 282
9 T h e PalaisSmale condition and unstable critical

points of variational problems 291
9.1 The Palais-Smale condition 291
9.2 The mountain pass theorem 301
9.3 Topological indices and critical points 306
Index 319
Preface and summary
The calculus of variations is concerned with the construction of optimal

shapes, states, or processes where the optimality criterion is given in
the form of an integral involving an unknown function. The task of the
calculus of variations then is to demonstrate the existence and to deduce
the properties of some function that realizes the optimal value for this
integral. Such variational problems occur in many-fold applications, in
particular in physics, engineering, and economics, and the variational
integral may represent some action, energy, or cost functional. The cal
culus of variations also has deep and important connections with other
fields of mathematics. For instance, in geometrically defined classes of
objects, a variational principle often permits the selection of a unique
optimal representative, and the properties of this representative can fre
quently be used to much advantage to deduce additional information
about its class. For these reasons, the calculus of variations is a rich and
ample mathematical subject, and a good impression of this diversity
can be obtained by reading the beautiful book by S. Hildebrandt and
A. Tromba, The Parsimonious Universe, Springer, 1996.
In this textbook, we have attempted to present some of the many faces
of the calculus of variations, and a brief summary may be useful before
putting the contents into a broader perspective. A t the same time, we
shall also describe the logical connections between the various chapters,
in order to facilitate reading for readers with a specific aim. The book
is divided into two parts. The first part treats variational problems for
functions of one independent variable; the second, problems for functions
of several variables. The distinction between these two parts, however, is
also that the first treats the more elementary and more classical aspects
of the subject, while the second is concerned w i t h some more difficult
topics and uses somewhat more abstract reasoning. I n this second part,
x
Preface and summary xi
also some examples are presented in detail that occurred in recent ap

plications of the calculus of variations. This second part leads the reader
to some topics and questions of current research in the calculus of vari
ations.
The first chapter of Part I is of a somewhat introductory nature and
attempts to develop some intuition for the properties of solutions of vari
ational problems. I n the basic Section 1.1, we derive the Euler-Lagrange
equations that any smooth solution of a variational problem has to sat
isfy. The topics of the other sections of that chapter contain some reg
ularity questions and an outline of the so-called direct methods of the
calculus of variations (a subject that will be taken up in much more de
tail in Chapter 4 of Part I I ) , Jacobi's theory of the second variation and
stability of solutions, and Noether's theorem that deduces conservation
laws from invariance properties of variational integrals. A l l those results
will not be directly applied in subsequent chapters, but should rather
serve as a motivation. I n any case, basically all the chapters of Part I can
be read independently, after the reader has gone through Section 1.1.
In Chapter 2, we treat one of the most important variational prob
lems, namely that of geodesies, i.e. of finding (locally) shortest curves
under smooth geometric constraints. Geodesies are of fundamental im
portance in Riemannian geometry and several physical applications. We
shall make use of the geometric nature of this problem and develop some
elementary geometric constructions, to deduce the existence not only of
length-minimizing curves, but also of curves that furnish unstable criti
cal points of the length functional. I n Chapter 3, we present some more
abstract aspects of such so-called saddle point constructions. A t this
point, however, we can only treat problems that allow the reduction to
a finite dimensional situation. A deeper treatment needs additional tools
and therefore has to wait until Chapter 9 of Part I I . Geodesies will only
occur once more in the remainder, namely as an example in Section 4.3.
Chapter 4 is concerned with one of the classical highlights of the cal
culus of variations, the theory of Hamilton and Jacobi. This theory is
of particular importance in mechanics. Presently, its global aspects are
resurging in connection w i t h symplectic geometry, one of the most active
fields of present mathematical research.
Chapter 5 is a brief introduction to dynamic optimization and control
theory The canonical equations of Hamilton and Jacobi of Section 4.1
briefly reoccur as an example of the Pontryagin maximum principle at
the end of Section 5.3.
As mentioned, Part I I is of a less elementary nature. We therefore need
xii Preface and summary
to develop some general theory first. I n Chapter 1 of that part, Lebesgue

integration theory is summarized (without proofs) for the convenience
of the reader. While in Part I , the Riemann integral entirely suffices
(with the exception of some places in Section 1.2), the function spaces
that are basic for Part I I , namely the LP and Sobolev spaces, are es
sentially based on Lebesgue's notion of the integral. I n Chapter 2, we
develop some results from functional analysis about Banach and Hilbert
spaces that will be applied in Chapter 3 for deriving the fundamen
p
tal properties of the L and Sobolev spaces. (In fact, as the tools from
functional analysis needed in subsequent chapters are of a quite varied
nature, Chapter 2 can also serve as a brief introduction into the field of
functional analysis itself.) These chapters serve the purpose of making
the book self-contained, and for most readers the best strategy might
be to start with Chapter 4, or at most with Chapter 3, and look up the
results of the previous chapters only when they are applied. Chapter 4
is fundamental. I t is concerned with the existence of minimizers of vari
ational integrals under appropriate convexity and lower semicontinuity
assumptions. We treat both the standard method based on weak com
pactness and a more abstract method for minimizing convex functionals
that does not need the concept of weak convergence. Chapters 5-7 essen
tially discuss situations where those assumptions are no longer satisfied.
Chapter 5 deals with the method of relaxation, while Chapters 6 and
7 present the important concept of T-convergence for minimizing func
tionals that can be represented only in an indirect manner as limits of
other functionals. Such problems occur in many applications, including
homogenization and phase transitions, and several such examples are
treated in detail. Chapter 8 discusses bifurcation theory. We first dis
cuss the variational aspects (Jacobi fields), taking up the constructions
of Sections 1.1 and 1.3 of Part I , then develop a general functional an
alytic framework for analyzing bifurcation phenomena and then treat
the example of minimal surfaces of revolution (catenoids) in the light
of that framework. Chapter 8 is independent of Chapters 4-7, and of a
more elementary nature than those. The key tool is the implicit function
theorem in Banach spaces, proved in Section 2.4. The last Chapter 9 re
turns to the topic of the existence of non-miminizing, unstable critical
points of variational integrals. While such solutions usually cannot be
observed in physical applications because of their unstable nature, they
are of considerable mathematical interest, for example in the context of
Riemannian geometry. Chapter 9 is independent of Chapters 4-8.
Preface and summary xiii
The present book is self-contained, with very few exceptions. Prere

quisites are only the calculus of one and several variables.
Although, as indicated, there are important connections between the
calculus of variations and geometry, the present book is of an analytic
nature and does not explore those connections. One such connection con
cerns the global aspects of the space of solutions of one-dimensional vari
ational problems and their trajectories that started w i t h the qualitative
investigations of Poincare and is for example represented in V . I . Arnold,
Mathematical Methods of Classical Mechanics, G T M 60, Springer, New
York, 2nd edition, 1987. Here, geometric methods are used to study
variational problems. I n the opposite direction, variational methods can
often be used to solve geometric problems. This is the topic of geometric
analysis; we refer the interested reader to J. Jost, Riemannian Geome
try and Geometric Analysis, Springer, Berlin, 2nd edition, 1998, and the
references contained therein.
There is one important omission in this textbook. Namely, the reg
ularity theory for solutions of variational problems is not treated, w i t h
the exception of the one-dimensional case in Section 1.2 of Part I , and
the simplest example of the multi-dimensional theory, namely harmonic
functions (plus an easy generalization) in Section 4.5 of Part I I . There
fore, the solutions of the variational problems that are discussed usually
only are obtained in some Sobolev space. We think that a detailed treat
ment of regularity theory more properly belongs to the realm of partial
differential equations, and therefore we have to refer the reader to text
books and monographs on partial differential equations, for example
D. Gilbarg and N . Trudinger, Elliptic Partial Differential Equations of
Second Order, Springer, Berlin, 2nd edition, 1983, or J. Jost, Partielle
Differentialgleichungen, Springer, Berlin, 1998.
In any case, the present textbook cannot cover all the many diverse
aspects of the calculus of variations. For readers who are interested in a
more extensive treatment, we strongly recommend M . Giaquinta and St.
Hildebrandt, Calculus of Variations, several volumes, Springer, Berlin,
1996 ff., as well as E. Zeidler, Nonlinear Functional Analysis and its
Applications, Vols. I l l and I V , Springer, New York, 1984 ff. (a second
edition of Vol. I V appeared in 1995). Additional references are given
in the course of the text. Since the present book, however, is neither a
research monograph nor an account of the historical development of the
calculus of variations, references to individual contributions are usually
not given. We just list our sources, and refer the interested readers as
well as the contributing mathematicans to those for references to the
original contributions.
xiv Preface and summary
The authors thank Felicia Bernatzki, Ralf Muno, Xiao-Wei Peng, Mar-
ianna Rolf, and Wilderich Tuschmann for their help in proofreading and
checking the contents and various corrections, and Michael Knebel and
Micaela Krieger for their competent typing.
The present authors owe much of their education in the calculus of
variations to their teacher, Stefan Hildebrandt. In particular, the pre
sentation of the material of Chapters 1 and 4 in Part I is influenced
by his lectures that the authors attended as students. For example, the
regularity arguments in Section 1.2 are taken directly from his lectures.
For these reasons, and for his generous support of the authors over many
years, and for his profound contributions to the subject, in particular to
geometric variational problems, the authors dedicate this book to him.
Remarks on notation
d
A dot always denotes the Euclidean scalar product in R , i.e. if
d d d
x = (x\...,x ) ,y = (y\...,y )eR ,
then
d
x% % x% %
x -y y y (Einstein summation convention) ,
2=1
and
2
|x| = X X.
For a function u(t), we write
In Part I , the independent variable is usually called t, because in many

physical applications, it is interpreted as the time parameter. Here, the
dependent variables are mostly called u(t) or x(t). I n Part I I , the inde
1 d
pendent variables are denoted by x = ( x , . . . , x ) , conforming to estab
lished conventions.
We use the standard notation
k
c (n)
for the space of A;-times continuously differentiable functions on some
d
open set Q C M , for k = 0 (continuous functions), 1,2,..., oo (infinitely
often differentiable functions). For vector valued functions, w i t h values
d
in M , we write
k d
C (fl,R )
xv
xvi Remarks on notation
for the corresponding spaces.
Co(ft)
denotes the space of functions of class C on ft that vanish identically
outside some compact subset K C ft (where K may depend on the
function, of course). Occasionally, we also use the notation
fc
<? (fi)
0
k
for C functions on ft that again vanish outside some compact subset
K Cft.
Finally, we use the notation
to indicate that the expression on the left of this symbol is defined by

the expression on the right of i t .
P a r t o n e
One-dimensional variational problems

1
The classical theory
1.1 T h e E u l e r - L a g r a n g e equations. E x a m p l e s
The classical calculus of variations consists in minimizing expressions of
the form
d d
where F : [a, 6] x R x R > E is given. One seeks a function u : [a, 6]
d
R minimizing J. More generally, one is also interested in other critical
points of J. Usually, u has to satisfy some constraints, the most common
one being a Dirichlet boundary condition
u(a) = u\
u(b) = 1*2-
Also, one needs to specify a class of admissible functions among which

one seeks a minimizing u. For example, one might want to take the
class of continuously differentiable or piecewise continuously differen-
tiable functions. Let us consider some examples of such variational prob
lems:
(1) We want to minimize the arc-length of the graph of a function u :

2
[a, 6] K, i.e. the length of the curve (t,u(t)) C K among all
graphs with prescribed boundary values u(a),u(b). This leads to
the variational problem
mm.
Of course, one knows and easily proves that the solution is the
straight line between u(a) and u(b), i.e. satisfies u(t) = 0.
3
4 The classical theory
(2) Historically, the calculus of variations started with the so-called

brachystochrone problem that was posed by Johann Bernoulli.
2
Here, one wants to connect two points (to,yo) and (t\,y\) in R
by such a curve that a particle obeying Newton's law of gravi
tation and moving without friction travels the distance between
those points in the fastest possible way. After falling the height y,
the particle has speed {2gy)^ where g is the gravitational accel
eration. The time the particle needs to traverse the path y = u(t)
then is
I[u)=
L i-^w d t
(3) A generalization of (1) and (2) is

b 2
f Jl 4- u(t) ,
I(u)
V ;
= / ~rdt,
Ja 7 (t,(t))
where 7 : [a, 6] x R R is a given positive function. This vari

ational problem also arises from Fermat's principle..That princi
ple says that a light ray chooses the path that needs the shortest
time to be traversed among all possible paths. I f the speed of
light in a given medium is y(t,u(t)), we obtain the preceding
variational problem.
If one seeks a minimum of a smooth function

d
/ : fi R ( f i open in R ) ,
one knows that at a minimizing point Zo fi, one necessarily has
Df(z )
0 - 0,
where Df is the derivative of / . The first variation of / actually has

to vanish at any stationary point, not only at minimizers. I n order to
distinguish a minimizer from other critical points, one has the additional
2
necessary condition that the Hessian D f(z ) is positive semidefinite and
0
(at least for a local minimizer) the sufficient condition that i t is positive
definite.
In the present case, however, we do not have a function / of finitely
many independent real variables, but a functional Z o n a class of func
tions. Nevertheless, we expect that a first derivative of J something
still to be defined needs to vanish at a minimizer, and moreover that
a suitably defined second derivative is positive (semi)definite.
In order to investigate this more closely, we assume that F is of class

1
C and that we have a minimizer or, more generally, a critical point of /
1
that also is C . We also assume prescribed Dirichlet boundary conditions
u(a) = u i , u(b) = U2. I n other words, we assume that u minimizes / in
1
the class of all functions of class C satisfying the prescribed boundary
d
condition. We then have for any 77 G CQ ([a, 6], M ) f and any s G R
I(u + sri) > I(u).
Now
I(u + sri)= I F(t,u(t) + sri(t),u(t) + sf}(t))dt.

Ja
1
Since F , u, and 77 are assumed to be of class C , we may differentiate
the preceding expression w.r.t. s and obtain at s = 0
I(u + sr,)Uo (1.1.1)
= J {F (t,u{t),u{t))-r)(t)
u + F (t,u{t),u{t))-T](t)}dt,
p
Ja
where F is the vector of partial derivatives of F w.r.t. the components
u
of u, and F the one w.r.t. the components of u(t).

p
We now keep 77 fixed and let s vary. We are thus just in the situation of
a real valued f(s), s G R, (f(s) = I(u + srj)), and the condition / ' ( 0 ) = 0
translates into
0= / (1.1.2)
/ a
and this actually then has to hold for all rj CQ. We now assume that F
2
and u are even of class C . Equation (1.1.2) may then be integrated by
parts. Noting that we do not get a boundary term since 77(a) = 0 = 77(6),
we thus obtain
b
0 = ^ |(F(t, (t),u(<))-|(F (i u p ) U ( < ) ^ ) ) ) ) !,() J d
) (1.1.3)
d
for all 7] Co ([a, 6 ] , R ) . I n order to proceed, we need the so-called
fundamental lemma of the calculus of variations:
f This means that r) is continuously differentiable as a function on [a, b) with values

d
in R and that there exist a < a\ < b\ < b with rj(x) = 0 if x is not contained in
[aiM]-
d
L e m m a 1.1.1. If h e C ( ( a , 6 ) , R ) satisfies
b
L h(t)(p(t)dt = 0 for all <p G CQ ((a, 6), R ) , D
then h = 0 on (a, 6).

Proo/. Otherwise, there exists some t G (a, 6) with 0
M*o) ^ 0.
l
Thus, h (to) i=- 0 for some index io { 1 , . . . , d}. Since / i is continuous,
there exists some <5 > 0 w i t h
a<t -6<t
0 0 + 6 ^ |/i (t )| 0 whenever | t - t\ < 6.
0
d
We then choose 6
0
io
0 if |t -t|<(5
0
<^(t)=0 fort^to, t{l,...,d}.

For this choice of </?, however
/ h(t)(p(t)dt= / h(t)(p(t)dt ^ 0,
./a J to 6
r
contradicting our assumption. Thus, necessarily /i(o) = 0 f all
(a, 6).
g.e.d.
2
Lemma 1.1.1 and (1.1.3) imply that a minimizer of I of class C has
to satisfy the so-called Euler-Lagrange equations, namely:
2 d d 2 D
T h e o r e m 1.1.1. Let F G C ( [ a , 6 ] x R x R , R ) , and letu G C ( [ a , 6 ] , R )
be a minimizer of
I(u) = J F(t,u{t),u(t))dt
Ja
among all functions with prescribed boundary values u(a) andu(b). Then
u is a solution of the following system of second order ordinary differ
ential equations, the Euler-Lagrange equations
(F (t,u(t),u(t)))
p - F (t,u(t),u(t))
u = 0. (1.1.4)
Written out, the Euler-Lagrange equations are
F (t,
pp u(t),u(t))u(t) + F (t, pu u(t), u(t))u(t)
+ F (t,u(t),u(t))
pt - F (t,u(t),u(t))
u = 0, (1.1.5)
i. e. a system of d ordinary differential equations of second order that are

linear in the second derivatives of the unknown function u.
Let us compute the Euler-Lagrange equations for our preceding three

examples:
(1) Here F = 0, F =
u p / t ^ a > and we get
i/i+u(t)
d u(
u(t) u(t) 2
u(t) il(t)
0 =
3 '
i.e.
u(t) = 0
meaning that u has to be a straight line, a fact that we know of

course.
(3) For the general example (3), we obtain as Euler-Lagrange equa
tions
o= | + 2 ^ + ^
2 8
d*7(t,(*))>/l + ( * ) T
2
u{t) ii{t) u(t) 7 t
hence
2 2
0 = u(t) - ^ u ( t ) (1 + 7i(t) ) + ^ (1 + < i ( t ) ) . (1.1.6)
(2) We just need to insert 7 = yj2gu(t) into (1.1.6) to obtain
0 = fi(t) + (l+(*))
The classical theory
Actually, (2) is an example of an integrand F(t,u,ii) that does not

depend explicitly on t, i.e. F = 0. I n this case t
j (Ft - uF ) = u(F
p u - jF ) p = 0 by (1.1.4),
and hence every solution of the Euler-Lagrange equation (1.1.4) satisfies
F(t,u{t),u{t)) -u(t)F (t,u(t),u(t))

p = constant. (1.1.7)
Conversely, every solution of (1.1.7), w i t h the exception of ii = 0, i.e.

u = constant, also satisfies (1.1.4).
In the case of example (2), we have F = and (1.1.7) becomes
2
= ^ ( 1 + u ), i f we denote the constant in (1.1.7) by A.
In all examples ( l ) - ( 3 ) , we actually had d = 1. I f one modifies e.g. (1)
d
and seeks a curve g(t) = ( # i ( ) , . . . , 9d(t)) C R connecting two given
points g(a) and ^(6), our variational problem becomes
The Euler-Lagrange equations in this case are

d d
2
, . 9i J2(9j) -9i E 9j9j
Q = d ^ ) _ _ ^ = _ =i
J i-i
d t
( d \ * ( d \
3 { ) 2
f jgftW ) Ls ^' )
for i = 1 , . . . , d.
d
We now recall that any smooth curve g(t) C R may be parameterized
by arc-length, i.e.
= 1. (1.1.9)
We also know that a reparameterization of a curve g(t) does not change

its arc-length 1(g). Consequently, we may assume (1.1.9) in (1.1.8). The
latter then becomes
0 for i= l d
d
so that we see again that a length minimizing curve in E is a straight
line.
1.1 The Euler-Lagrange equations. Examples
Often, one also meets the task of minimizing
I(u) = / F(t,u{t),u(t))dt
Ja
subject to some constraint, for example
S(u)= G(t,u(t),u(t))dt = c 0 (a given constant). (1.1.10)

Ja
As i n the case of finite dimensional minimization problems, one then

finds a Lagrange multiplier A with
0 = A (I( u + srj) + \S{u + sr])) | o 5 = (1.1.11)

as
d
for all rj G C g ( [ a , 6 ] , E ) . This leads to the Euler-Lagrange equations
j t (F (t,u(t),ii(t))
p + \G (t,u(t),ii(t)))
p
- (F (t,u(t),u(t))
u + \G (t,u(t),u(t)))
u = 0. (1.1.12)
Example. We wish to miminize
2
J( ) = / <i(*) d*
M
./a
under the constraint
2
5(ti)= / u(t) dt=l, (1.1.13)
with u(a) = 0 = u(b). (1.1.12) becomes

u(t)-Xu{t) = 0. (1.1.14)
2 2
Thus, A is an eigenvalue for the differential operator d /dt under the
Dirichlet boundary conditions u(a) = 0 = u(b). Of course, this example
can easily be generalized.
Summary. We seek solutions of the variational problem
I(u) > min,
with
I(u) = J F(t,u(t),ii(t))dt
Ja
d
for given F and unknown u : [a, 6] R . I f F and u are differentiate,
one may consider some kind of partial derivative, namely
61{u,rj) : = I{u + srj)y
d
for rj E Co([a,6],R ). For a minimizer u then
61 (u, rj) = 0 for all such rj.

2
If F and u are of class C , this leads to the Euler-Lagrange equations
-F (t,u(t),ii(t))
p - F*(t,u(t),u(t)) = 0.
The classical strategy for solving the problem
I(u) > min
consists in solving the Euler-Lagrange equations and then investigating

whether a solution of the equations is a minimum of / or not.
1.2 T h e idea of the direct methods and some regularity

results
So far, our formulation of the variational problem
I(u) > min
has been rather vague, because we did not specify in which class of
functions u we are trying to minimize / . The only things we did prescribe
were boundary conditions of Dirichlet type, i.e. we prescribed the values
d
u(a) and u(b) for our functions u : [a, 6] * R .
Because of our derivation of the Euler-Lagrange equations in the pre
2
ceding section, i t would be desirable to have a solution u of class C .
So one might want to specify in advance that one minimizes / only
2
among functions of class C . This, however, directly leads to the ques
2
tion whether / achieves its infimum among functions of class C (with
prescribed Dirichlet boundary conditions, as always) or not, and if i t
1
does, whether the infimum of / in some larger class of functions, say C ,
2
could be strictly smaller than the one in C . I n the light of this question,
it might be preferable to minimize / in the class of all functions u for
which
1.2 Direct methods, regularity results 11
is meaningful. Here, we assume that F(t,u,p) is continuous in u and p

and measurable in t. For this purpose, one needs the class of functions
for which the derivative u(t) exists almost everywhere and is finite. This
is the class
AC([a,b})
of absolutely continuous functions. A function u G AC([a,b]) satisfies for

ti,t e2 [a,6]
u{t )-u(h)
2 = / ii(t)dt.
Jti
Note that F(t, u(t),u(t)) is a measurable function of t for u AC by our
assumptions on F and the fact that the composition of a measurable and
a continuous function is measurablef. The idea of the direct methods in
the calculus of variations, as opposed to the classical methods described
in the preceding section then consists i n minimizing / in a class of func
tions like AC([a,b]) and then trying to show that a solution u because
of its minimizing character actually enjoys better regularity properties,
2
for example to be of class C , provided F satisfies suitable assumptions.
This minimizing procedure will be treated later J, since we want to
return to the classical theory for a while. Nevertheless, even for the
classical theory, one occasionally needs certain regularity results, and
therefore we now briefly address the regularity theory. To simplify our
notation, we put / : = [a, 6]. A class of functions intermediate between
1
C and AC is
1 d d
D ( / , E ) : = {u : / M , u continuous and piecewise
continuously differentiable, i.e. there exist
a = to < t\ < ... < t = b with u G
m
d
C H M j + i ] , M ) for j = 0 , . . . , m - l } .
1
u G D then has left and right derivatives u~(tj) and u+(tj) even at the
points where the derivative is discontinuous, and
f Lebesgue integration theory is summarized in Chapter 1 of Part I I . T h e required

composition property is stated there as Theorem 1.1.2. Here, this point will not
be pursued or used any further.
t See Chapter 4 of Part I I .
We shall use the same letter J to denote the functional to be minimized and the
domain of definition of the functions, inserted into this functional. T h i s conforms
to standard notations. T h e reader should be aware of this and not be confused.
Examples
Example 1.2.1. [a, 6] = [ - 1 , 1 ] , d = 1
2 2
I(u) = ( l - ( f k ( t ) ) ) di,
t i ( - l ) = 1 = u(l).
A minimizer is
1
ti(0 = | t | D ( / , R )
1
which is not of class C . The minimizer of / is not unique (exercise:
1
determine all minimizers), but none of them is of class C .
Example 1.2.2. [a,6] = [ - 1 , 1 ] , d = 1
2 2
I(u) = J (l-u(t)) u(t) dt
u(-l) = 0 , ti(l) = l .
Here, the unique minimizer is
, . x f0 for - 1 < t < 0

u { t )
= \t forO<*<!
1 1
which again is of class D , but not C .
Example 1.2.3. [a, 6] = [ - 1 , 1 ] , d = 1
2 2
J(u) = ^ (2t~u(t)) u(t) dt,
u(-l) = 0 , ti(l) = l .
The unique minimizer is
, . f0 for - 1 < t < 0

=
forO<*<!
1 2
which is of class C , but not of class C .
1
T h e o r e m 1.2.1. Let F(t,u,p) be of class C w.r.t. u and p and con
d d d
tinuous w.r.t t ( F : J x R x R -+ R ) , and let u G AC(I,R ) be a
solution of
6I(u,rj) = J {F (t,u,u)-rj
u + F (t,u,u)-f)}dt
p =0 (1.2.1)
Ja
d d
for all rj G AC (I,R )0 (i.e. rj G AC(I,R )) and we require that if I =
[a, 6], there exist a < a\ <b\ < b with rj(x) 0 if x is not contained in
d
[ai,&i], as in the definition of C$([a, 6], R )). Then for almost all points
in I
j F (t,
t p u(t), u(t)) = F (t, u(t), u(t)) u (1.2.2)
(note, however, that the derivative on the left hand side cannot be com
l d
puted by the chain rule). If u G C (I,R ), (1.2.2) holds for all t G J,
1 d
and if u G J D ( / , E ) , at those points tj where u(tj) is discontinuous
~ (F (t ,u(t ),u(t ))_

p j j j = Fuitjiuit&ii-itj)),
and analogously for the right derivative.

d
Remark. I t actually suffices to assume (1.2.1) for all rj G Co(I,R ),
because functions in ACQ may be approximated by CQ functions. I f
1 1
u G C or D , the proof anyway only requires (1.2.1) for 77 G CQ or Dp,
respectively (where Dg is defined analogously to C Q ) .
Proof. We have, omitting the obvious arguments of F,F , U etc.,
T]dt = Fudy
jf F rjdt
u = j[ ^ (j[ F d
^ y) ~J a (/ ) V dt
(1.2.1) then implies
0=J (Fp-J F dy^f]dt.

u
We now make use of:

X
L e m m a 1.2.1. Let h G L (I,R) satisfy
b
h(t)ip(t)dt = 0 for all <p G AC (I, 0 E). (1.2.3)
Then there exists a constant cGE
/i() = c / o r almost all t G / .
Remark. I t actually suffices to assume (1.2.3) for all <p G C Q ( / , R ) . I f

1
h G C , one directly sees from the proof that cp C$ suffices.
Proof. We put
1
c := / h(t)dt
b a
~ Ja
and
<p(t):= J (h(y)~c)dy.
Ja
Then
<p(a) = 0, and (p(b) = f (p(t)dt = 0. (1.2.4)

Ja
Equation (1.2.3) implies
0 = f \h{y)-c)h{y)dy = f \h(y) - cf dy
Ja Ja
because of (1.2.4). This implies the claim.

q.e.d.
We now may complete the proof of Theorem 1.2.1:

d
By Lemma 1.2.1 there exists c G R with
F {t,u(t),u(t))
p = f F {y,u{y),u(y))dy
u +c (1.2.5)
Ja
for almost all t e l . Therefore, F is of class AC, and differentiating

p
1 1
(1.2.5) gives (1.2.3). The claims for u e C or D are obvious from the
proof.
q.e.d.
d d 1
T h e o r e m 1.2.2. Let F : I x R x R be of class C , and let F be also p
l or
of class C , and let det (F i j (t, u ( t ) , ^
p p f all t E I
l d
and a solution u G C (I,R ) of
d
6I(u,T])=0 for allr)eCl(I,R ).
2
Then u is of class C .
Proof. We define
d d d
<f>: R x R x R x R R
via
<t)(t,u,p,q) := F (t,u,p) p - q.
Our assumption d e t F ^ 0 makes it possible to apply the implicit

p p
function theorem to conclude that
<t>(t,u,p,q) = 0
may be uniquely solved w.r.t. p near UQ u(t ), 0 po = u(t ),0 q0 =

F(to,u ,po)
0 for any to G I . Thus, there exists a neighbourhood U of
(to,uo,qo) such that for each (t,u,q) G /, <t> 0 has a unique solution
d 1
p = <p(t, u, q) and that (p : U E is of class C . Since we already know
a solution of <fi = 0, namely (, u(t), u(t), F (t, u(t),u(t))), the uniqueness p
of the solution cp implies
u(t) = </?(, ii(),F (,i(),u())) p for t near - 0
1 2
Since <p is of class C , so then is ii(t), hence t i G C . Since to e I was
2 d
arbitrary, i i G C ( J , E ) .
T h e o r e m 1.2.3. Let F satisfy the assumptions of Theorem. 1.2.2, and

d
in addition assume that F pp is (positive or negative) definite on ft x R
d + 1 d
where ft C E contains {(t,u(t)) : t G / } . Let u G AC(I,R ) satisfy
d
61{u, rj) = 0 for all rj G AC (I, 0 R)
(assume that F (t,u(t),u(t)) u and F (t,u(t),u(t))

p are integrable). Then
2 d
ueC {I,R ).
Proof. Since the uniqueness result of the implicit function theorem is

only local, i t cannot be applied anymore because u(t) might be discon
tinuous. We thus need a global argument. Thus, assume that for given
d d
(t,u,q) G ft x R , there are two solutions pi,p2 G R of (f)(t,u,p,q) = 0,
i.e.
q = F (t,u,pi) p and q= F (t,u,p ).
p 2
Thus
/ F (t,u,
pp Pl + s(p ~ ))ds 2 Pl (p -pi)=0.
2 (1.2.6)
Jo
By our assumption on F , (1.2.6) is invertible, hence p = P i , hence
pp 2
uniqueness.
Using this global uniqueness together w i t h the existence result of
the implicit function theorem, we now see that for any (t,u,q) in a
sufficiently small neighbourhood of ( ,^(b<Zo) (ô I , Ho = u(to), 0
qo = F (t ,^(bPo),
p 0 Po = ô(ô)), there is a unique solution (p(t,u,q)
of
F (t,u,p)
p -q =0
1
and <p is of class C . Thus, as in the proof of Theorem 1.2.2,
u(t) = <p(t,u(t),F (t,u(t),u(t)))

p
for almost all f i n a neighbourhood of to. Since u(t) and F (t,u(t),u(t)) p
are absolutely continuous w.r.t. t (the latter by Theorem 1.2.1), u(t)

coincides for almost all t near to with an absolutely continuous function
v(t). We put
to
1
w then is of class C . Since u is absolutely continuous, by a theorem of
Lebesgue
1
Since v = u almost everywhere, we conclude u = w, hence u G C near
2
to, which was arbitrary in / . Theorem 1.2.2 then gives u G C .
q.e.d.
Corollary 1.2.1. Under the assumptions of Theorem 1.2.3, any AC-

d
solution of 6I(u,rj) = 0 for all rj G ACo(I,R ) is a solution of the
Euler-Lagrange equations
TF (t,u(t),ii(t))
p - F (t,u(t),u(t))
u = 0 (1.2.7)
or equivalently of
F (t,pp u(t),u(t))ii(t) + F (t, pu u(t), u(t))u(t)

+F (t,u(t),ii(t))
pt - F (t,u(t),u(t))
u = 0. (1.2.8)
l
The same holds under the assumptions of Theorem 1.2.2 for a C - so
d
lution of6I(u,rj) = 0 for all rj G C<J(J,R ).
q.e.d.
d d k
T h e o r e m 1.2.4. Let F : I x R x R -+ R be of class C , and let F p
k 1
also be of class C , k G { 2 , 3 , . . . , o o } . Suppose u is of class C and a
d
solution of 6I(u,rj) = 0 for all rj G C o ( / , R ) , and suppose
det (F pipj (*, u{t), ii(t)) j= ... )

i li id ^ 0 for all t G / . (1.2.9)
k+l d 1
Then u G C (I,R ). (The same result holds if we assume that u G C
is a solution of the Euler-Lagrange equations (1.2.8).)
2
Proof. By Theorem 1.2.2, u is of class C , and by Corollary 1.2.1, i t
solves (1.2.8). Because of (1.2.9), F (t, pp u(t),u(t)) is an invertible matrix,

hence
1
il(t) = F (t,u(t),u(t))
pp
{-F (t,u(t),u(t))
pu - F (t,u(t),u(t))
pt + F (t,u(t),ii(t))}
u
(1.2.10)
3
Let now j < k, and suppose inductively u E C . The right hand side of
3 x 7 1
(1.2.10) then is of class C ~ . Therefore, u is of class C- " , hence u is
j+1
of class C .
q.e.d.
The preceding proof most clearly shows the importance of the as

sumption det(Fpt j(t, u(t),ii(t))) ^ 0 that already occurred in the proof
p
of Theorem 1.2.2. Namely, i t implies that the Euler-Lagrange equations

(1.2.8) can be solved for u in terms of u and u.
Corollary 1.2.2. If under the assumption of Theorem 1.2.3, F and F p
k
are of class C , then a solution u of 6I(u,rj) = 0 for all rj ACo is of
k + 1
class C .
q.e.d.
Summary. I f one wants to solve
I(u) > min
by a direct minimization procedure, i t is preferable to admit a class of

d
comparison functions u that is as large as possible. AC (I, E ) seems to
be a good choice, because this is the largest class for which
7(u)= J F(t,u(t),u(t))
is well defined, assuming F(t,u,p) to be continuous in u and p and
measurable in t. However, if one then finds a minimizer u, it might not
be a solution of the Euler-Lagrange equations, because it is not regular
enough. I f the invertibility condition d e t F ^ 0 is satisfied, however, p p
one may show that a minimizer u is as regular as F allows. Namely, if

fc k + 1
F and F are of class C , k G { 1 , 2 , . . . , oo}, then u is of class C
p .
Examples show that without such an invertibility condition, regularity
need not hold. This invertibility condition det F ^ 0 implies that the pp
Euler-Lagrange equations allow the expression of u(t) in terms of u(t)

and u(t).
1.3 T h e second variation. Jacobi fields

1 d
We assume that u G D ( / , E ) is a critical point of
I(u) = / F(t,u(t),u(t))dt,
Ja
i.e.
d
6I(u,T]) = 0 for all 77 G >o(/,M ). (1.3.1)
We recall that
:= ^ / ( u + s77) u = 0 ,
and 8I(u,rj) = 0 is equivalent to s = 0 being a critical point of the

function
f(s)=I(u + sri).
If we want to decide if a given solution u minimizes J instead of just

being a critical point, we immediately see that a necessary condition
would be
/"(0) > 0 (1.3.2)
d
for the above function / and all 77 G D o ( J , R ) . Namely, by Taylor's
theorem, since / ' ( 0 ) = 0
2 2
m-f(0) = \s f"(0)+o(s ) forsÔ.
More precisely, (1.3.2) is needed for u to minimize / when compared w i t h

u 4- srj for sufficiently small s. I n other words, we want u to minimize i"
1
in a D -neighbourhood of itself, i.e. among functions
with
u(a) = v(a), u(b) = v(b) and (1.3.3)
sup (\u(t) - v(t)\ + \ii-(t) - v-(t)\ 4- \ii+(t) - < e (1.3.4)
for some e > 0. (Note: I t is not clear that e may be chosen independently
of v.) We define the second variation of / at u in the direction rj e DQ
as
2
d
2
6 I(u,rj) := _ / ( w - f s 7 ? ) u = 0 .
In order that this variation exists, we require for the rest of the section
2
that F is of class C . We then compute
2
6 I(u, 77) = ^ J" F(t, u(t) + sr](t), ii(t) + 7(<))d*|.
rb
I
J+a 2F i (t,u(t),u(t))r) {t)r {t)
p UJ i )j
+ F^jûi^ûit^rji^rjjit)} dt. (1.3.5)
Here, and in the sequel, we employ the standard summation conventions,

e.g.
d
Fpipjrjirjj = ] T Fpipjfiifjj.
We abbreviate (1.3.5) as
fb
2
6 I{u,r])= {F r)r)
pp + 2F r)rj pu + F rjrj} uu dt. (1.3.6)
Ja
Our preceding considerations imply:

2 d d l d
T h e o r e m 1.3.1. SupposeF e < 7 ( J x R x R x R ) andletu G D (I,R )
satisfy I(u) < I(v) for all v with {1.3.3), (1.3.4). Then
2 d
6 I(u,rj)>0 forallrjeDl(I,R ). (1.3.7)
We now put, for given u,
<p(t, 77, TT) : = F pipj (t, u(t), u(t))'Ki'K + 2F i j p UJ (t, u(t), u^))-*^
(,u(),^))^%,
and we define the accessory variational problem for J(M) min as
d
Q(rj) : = / cf)(t,r)(t),r)(t))dt -> min among all 77 G Z ^ ( J , R ) .
(1.3.8)
If u satisfies the assumptions of Theorem 1.3.1, then
Q(rj) > 0 for all 77 G >J, (1.3.9)
and hence 77 = 0 is a trivial solution of (1.3.8). We are interested in the

question whether there are others. The Euler-Lagrange equations for
(1.3.8) are
= ^(*,r?W,i)W), (1.3.10)
i.e.
~ (F (t, pp u(t), u(t))f)(t) + F {t,

pu u(t), u(t))ri(t))
= F (t,
pu u(t), u(t))fj(t) + F (t, uu u(t), u(t))rj(t). (1.3.11)
Since u is considered as given, our first observation is that (1.3.11) is a

linear homogeneous system of second order equations for the unknown
77. These equations are called Jacobi equations.
2 d
Definition 1.3.1. A solution 77 G C (I,R ) of the Jacobi equations
(1.3.11) is called a Jacobi field along u(t).
3 d d
L e m m a 1.3.1. Let F G C (I x R x R , R ) , det F {t, u{t), u{t)) ^ 0 pp
2 d d
for all t e I , u e C ( J , R ) . Then any solution of rj e AC {I,R ), 0
d 2
6Q(rj,(p) = 0 for all V(t)) = Fpp(t, u{t), u(t)) for all t and 77
and so the assumption det F (t, u(t), u(t)) ^ 0, that is seemingly weaker
pp
than the one of Theorem 1.2.3, indeed suffices to apply that Theorem.
q.e.d.
We now derive the so-called necessary Legendre condition:
T h e o r e m 1.3.2. Under the assumption of Theorem 1.3.1, i.e. u G

1 d
D ( / , R ) minimizes I in the sense described there, we have that
F (t,u(t),u(t))
pp is positive semidefinite for all t G / ,
i.e.
d d
F pipj (t, u(t), u{t))?? > 0 for all = (\ ..., ) e R.
(At points where ii(t) is discontinuous, this holds for the left and right
derivatives.)
Proof. We may assume that t e I and ii is continuous at t . The result

0 0
at the points where u jumps then follows by taking appropriate limits,

and likewise at to a, 6. We then consider 0 < e < min(to a, b to)
d
and define 77 G Z ^ ( J , R ) by
{ 0
e
for a < t < 10 e and to 4- e < t < b
for t - to
linear for to e < t < 10 and for to < t < to + e
d
for given R . Then
{ 0

-
for a < t < t or t + e < t < 6
for t - e < t < t
0
for t < t < t + c.

0
0
0
0
0
We apply Theorem 1.3.1 to obtain
2 + C J 2
0 < 6 I(u, rj) = r F p V (t, ix(t), u(t))CZ dt + 0(e ) for c 0,
Jto-e
since all other terms contain a factor e, and we integrate over an interval
of length 2e. Hence
j j
F i (t ,u{t ),u(t ))Cl;
p pJ 0 0 o - lim - / F {t,u{t),u(t))C^ dt
pipj > 0.
0 Jt -e 0
q.e.d.
The Jacobi equations and the notion of Jacobi fields are meaningful
for arbitrary solutions of the Euler-Lagrange equations, not only for
minimizing ones. I n fact, Jacobi fields are solutions of the linearized
Euler-Lagrange equations. Namely:
3 d d
T h e o r e m 1.3.3. Let F e C {I x R x R , R ) , and let u (t) s be a family
2
of C -solutions of the Euler-Lagrange equations
j F (t,u (t),ii (t))

t p s s - F (t,u (t),u (t))
u s s = 0, (1.3.12)
with u s depending differentiably on a parameter s 6 (e,e). Then
rj(t) := - ^ ( % s = 0
ds
is a Jacobi field along u = uo.
Proof. We differentiate (1.3.12) w.r.t. s at s = 0 to obtain
~ (F (t, pp u(t), u(t))r)(t) + F (t, pu u(t), u(t))ri(t))
-F (t,u(t),u(t))r)(t)
pu - F {t,u(t),u(t))r](t)
uu = 0.
i.e. the Jacobi equation (1.3.11). q.e.d.

2
L e m m a 1.3.2. Let a < a\ < a <b, and let F and F be of class C 2 p
l d
in [ a i , a ] , and suppose r\ G C ([a\,a ],R )
2 is a Jacobi field on [ a i , a ] 2 2
with r](ai) = 0 = r)(a ). Then 2
/ <p(t,r)(t),r)(t))dt = 0. (1.3.13)
Proof. Since <f> is homogeneous of second order in (77,7r), we have
2<f>(t, 77, 7T) = (f) (t, 77, 7r)77 -h (f>n(t, 77, 7T)7T.
v
Therefore
2/ <f>(t,r),r,)dt = / {^,(,?, 1 7 ) + ))')}* (1-3-14)

/ai /ai
1
Comparing (1.3.10) and (1.3.11), we see that (f> is of class C as a n
function of t. We may hence integrate the last term in (1.3.14) by parts.

Since 77(01) = 0 = 77(02), we obtain
2 j <t>(t,T7,r))dt = j (j>ri{tiT7,rj) - j <t>A^V, f ^ = 0,
since 77 is a Jacobi field. q.e.d.
3 2
As before, let F be of class C , and let u(t) be a solution of class C
on [a, 6] of the Euler-Lagrange equations
j F {t,
t p u(t), ii(t)) - F (t, u(t), ii(t)) = 0.
u
Definition 1.3.2. Let a < a\ < a < b. We call the parameter value 2
a conjugate to a\ and the point (a ,u(a ))

2 conjugate to (a\,u(a\)) if 2 2
there exists a not identically vanishing Jacobi field 77 on [a\,a ] with 2
77(01) = 0 = 77(02).
We may derive the important result of Jacobi:
3 d d 2 d
T h e o r e m 1.3.4. LetF e < 7 ( J x R x R , R) and suppose u e C (I,R ).
Suppose that F (t,u(t),u(t))
pp is positive definite on I. If there exists a*
with a < a* 0, there exists v D (I, R ) with
v(a) = u(a), v(b) = u(b),
sup (\u{t) - v(t)\ 4 \u{t) - v (t)\) <e

tl
and
I(v) < I(u).
Proof. Let rj(t) be a nontrivial Jacobi field on [a, a*]. We put
rj(t) for a <t < a*

77 w 1
' " for a* < t < b.
d
Then 77* G ^ ( J , ] R ) , and by Lemma 1.3.2
Q(V*)= f* <P(t,v*,v*)dt = o.
Ja
If u were a local minimum, then by Theorem 1.3.1

2 d
0 < S I(u,fj) = Q(fj) for all 77 G Z ^ ( J , M ) .
Hence 77* would be a minimizer of Q, hence by Lemma 1.3.1 77* G

2 d
C (I,R ). Since ?)*(a*) = 0, then
7)*(a*)=0.
Since also 77*(a*) = 0, and since 77* solves the Jacobi equation, a (linear)
second order ordinary differential equation, the uniqueness theorem for
solutions of such equations implies
a contradiction, because by assumption 77 does not vanish identically.

Hence u cannot be a local minimizer.
q.e.d.
In words, Theorem 1.3.4 says that a solution of the Euler-Lagrange

equations cannot be minimizing beyond the first conjugate point. Turned
the other way round, Theorem 1.3.4 says that i f u is a local minimizer,
then there cannot be any parameter value a* w i t h a < a* < b that is
conjugate to a. I t may happen, however, that 6 is conjugate to a. A n
example will be given in the next chapter.
Summary. I n order to obtain necessary conditions for a solution of the

Euler-Lagrange equations
F (t,u(t),ii(t))
p = F (t,u(t),ii(t))
u
to minimize
one needs to study the second variation
Qfa) 2
: = 6 I(U,T ) ] = - ^ I ( U+ ST )
1 1 for r? Do-
If, for fixed u, we consider the variational problem Q(rj) * 0, we are led
to the Jacobi equations
~ (F (t, pp u(t), u(t))rj(t) + F (t, pu u(t),u(t))rj(t))
= F (t,
up u(t),u(t))r)(t) + F (t,
uu u(t),u(t))ri(t)
for 77.
Solutions rj with 77(a) = 77(6) = 0 are called Jacobi fields, a* G (a, 6) for
which there exists a nontrivial Jacobi field on [a, a*] is called conjugate
to a, and if there exists such a*, u cannot be locally minimizing on [a, 6].
In other words, a solution of the Euler-Lagrange equations cannot be
minimizing beyond the first conjugate point.
1.4 Free boundary conditions

We recall the definition of an n-dimensional embedded differentiable sub-
d
manifold M of R : For every p G M , there have to exist a neighbourhood
d n
V = V(p) C M , an open set U cR and an injective differentiable map
/ : U * V of everywhere maximal rank n (i.e. for every z U, the
n d
derivative Df(z), a linear map from E to E , has rank n) w i t h
M nv = f(U).
n
A n example is the sphere S described in detail in Section 2.1 (Exam
ple 2.1.1). The tangent space T M of M at p then is the vector space
P
n d
D / ( z ) ( E ) . I t can be considered as a subspace of the vector space T E , p
d
the tangent space of E at p.
As in 1.1, we now consider the variational problem
I(u)= / F(t,u(t),u(t))dt > min

Ja
2
w i t h F of class C . This time, however, we do not impose the Dirichlet
boundary condition that the values of u(a) and u(b) were prescribed,
but the more general condition that for given submanifolds M i , M 2
d
(differentiable, embedded) of E , we require that
u(a) G M i , i i ( 6 ) G M . 2
(Dirichlet boundary conditions constitute the special case where M\ and

M are points.)
2
In this section, we do not consider regularity questions. As an exercise,

1.4 Free boundary conditions 25
the reader should supply the necessary regularity assumptions on F , w,

etc. at each step.
Let u be a solution. Then, as before, u has to satisfy the Euler-
Lagrange equations, because i f u(a) G M i , 77(a) = 0, then also u(a) +
577(a) G M i for any s, and likewise at 6, and so we may again consider
variations of the form 72 + 577, 77 G DQ. This time, however, also more
general variations are admissible. Namely, let u (t) be a family of maps s
d
from / into M. depending differentiably on s G (e, c), w i t h u(t) = Uo(t)
and
u (a)
s GMi , u (b) G M
s 2 for all s.
Let
Then again
0= ^ / K ) | . _ 0 = F(t,u(t),u(t))dt^ 0
= f {F {t,u(t),u(t))-f,{t)
p + F {t,u(t),u{t))-T}(t)}dt
u
Ja
= f a \ - j F P + ^ } - v + F P ( * . ( * ) , ( * ) ) m l z i
i
= F (t,u(t)M*))-V(t)\ Za>
p
since u solves the Euler-Lagrange equations.

We now observe that 77(a) G T ( ) M i (and likewise at 6), since we may
u a
find a 'local chart' / as above w i t h MiDV(u(a)) = f(U) for a neighbour

n i
hood V of u(a) and some open set U C M ( n i = dim M i ) . By choosing
e smaller i f necessary, we may assume u (a) G M i f l V = f(U) for 5 G s
(~, e). Since / is injective, there then has to exist a curve 7(5) C U w i t h
1 ,
u (a) = fo>y(s) for all s. Hence 77(a) = u (a)
s = D/(/- 7i(a)) (0) s u=0 7
is indeed tangent to M i at u(a). Moreover, any tangent vector to M i at

u(a) can be realized in this manner. Therefore, since we may choose the
values of 77 at a and 6 independently of each other, we conclude
F (a,u(a),u(a))
p V = 0 for all V G T M
u{a) u
and likewise
F (6, u(6), u(b)) -W = 0

p for all W G r u ( 6 ) M . 2
We have thus shown:
T h e o r e m 1.4.1. Let u be a critical point of I among curves withu(a) G

Mi, u(b) G M {Mi, M2 given differentiable embedded submanifolds
2
d
of R ), i.e. ^ ^ ( ^ s ) | = 0 for all variations u (t) differentiable in
s = 0 s
s with u (a) G M u (b) G M for all s G ( - e , e ) (e > 0 ) . Then

s b s 2
u is a solution of the Euler-Lagrange equations for I , and in addi

tion, F (a,u(a),u(a))
p and F (b,u(b),u(b)) are orthogonal to
p T ^Mi u
d
and T ( 5 ) M 2 , respectively. In particular, if for example Mi = R , then
U
F (a,u(a),u(a))
p = 0.
Summary. I f instead of a Dirichlet boundary condition, we more gen

erally impose a free boundary condition that u(a) and u(b) are only
required to be contained in given differentiable submanifolds Mi and
d
M 2 , respectively, of E , then F (a,u(a),u(a)) and F (b,u(b),u(b)) are p p
orthogonal to these submanifolds for a critical point of / under those

boundary conditions.
1.5 Symmetries and the theorem of E . Noether

In the variational problems of classical mechanics, one often encounters
conserved quantities, like energy, momentum, or angular momentum. I t
was realized by E. Noether that all those conservation laws result from
a general theorem stating that invariance properties of the variational
integral / lead to corresponding conserved quantities. We first treat a
special case.
T h e o r e m 1.5.1. We consider the variational integral
I(u) = / F(t,u(t),u(t))dt,
Ja
2 d d
with F G C ([a,6] x l x E , E ) . We suppose that there exists a smooth
one-parameter family of differentiable maps
d d
h : R
s -> R
(the precise smoothness requirement is that
h(s,z) := h (z)s
2 d
is of class C ( ( - e , e ) x E , E ) for some e > 0),
0 0 0
with
d
h (z) 0 = z for all zeR
and satisfying
j\(t,h (u(t)),
s f h (u(t)^J
t s dt = j\(t,u(t), j u(t^J
t dt (1.5.1)
2 d
for all s G (~e,e) and all u G C ( [ a , 6], R ).
Then, for any solution u(t) of the Euler-Lagrange equations (1.1.4)
fori,
F (t, u(t),ii(t))
p ~h (u(t))\ s s=0 (1.5.2)
is constant in t G [a, 6].

Definition 1.5.1. A quantity C(t,u(t),u(t)) that is constant in t for
each solution of the Euler-Lagrange equations of a variational integral
I(u) is called a (first) integral of motion.
Proof of Theorem 1.5.1: Equation (1.5.1) yields for any t G [a, 6], using 0
h (z) = z,
0
= F t h s k s { u { t ) ) d t s = 0
^ s J a { ' ^ ^ Jt ) ^
= jT {F u (t,u(t),ii(t)) ^hs(u(t)) (1.5.3)
h
+F P (t, u(t),u(t)) J f sW))}dt\s=o-
t s
We recall the Euler-Lagrange equations (1.1.4) for u:
0 = jF t p (t, u(t), ii(t)) - F (t, u(t),u(t)). u (1.5.4)
Using (1.5.4) in (1.5.3) to replace F , we obtain u
0 = f [jF {t,u(t),u(t)) p fh {u{t))

a
h u
+FP (t,u{t),u(t)) J -ff s( (t))}dt\ =o
t s s (1-5.5)
= f j t (F (t,u(t),u(t))^- h (u(t))\ )
p s 3 s=0 dt.
Therefore
F (t ,u(to),u(t ))~h (u(t ))\

p 0 0 s 0 s=0 = F (a,u(a),u(a))~h (u(a))\
p s s=0
(1.5.6)
for any to G [a, 6]. This means that (1.5.2) is constant on [a, 6].
q.e.d.
Examples
3 n
Example 1.5.1. We consider for u : E > E , u = (ui,...,u )
n with
2
F(t,u(t),u(t)) = pmj-^f-- ÎI^H = E ^ i j,
3
i.e. a mechanical system in E w i t h point masses m*, and a potential
V(u) that is independent of the third coordinates of the Ui. Then
h (z)
a = z + se , 3
3
where e is the t h i r d unit vector i n M , leaves F invariant i n the sense
3
of Theorem 1.5.1. Since

d
^-h \ o s s= = e ,
3
we conclude that
1=1
i.e. the third component of the momentum vector of the system is con
served.
Example 1.5.2. Similarly, i f a system as in Example 1.5.1 is invariant

under rotations about the e -axis, and i f h now denotes such rotations,
3 8
then (up to a constant factor)

d
L ,
n \s=oUi = es A Ui.
s
as
Hence, the conserved quantity is the angular momentum w.r.t. the e - 3
axis,
n
m u e
^ F ez A Ui = ] P ( i i)
v ' ( 3 A Ui) = ] P (ui A ra^) e . 3
i=l i i
We now come to the general form of E. Noether's theorem
T h e o r e m 1.5.2 ( T h e o r e m of E . N o e t h e r ) . We consider the varia

tional integral
rb
I(u) = I F(t,u(t),u(t))dt
Ja
2 d d
with F C ([a,6] x R x E , E ) . We suppose that there exists a smooth
one-parameter family of differentiate maps
d d
h = (h ,h )
s s s : [a,b] x E > E x E
(s G ( - e o , e ) as before) with
0
d
h (t, z) = (t, z)
0 for all (t, z) G [a, b] x E
and satisfying
b
rh (b)
3 ( d \ r
/ F(t ,h (u(t )),h (u(t )))dt =
s s s s s s / F(t,u(t),u(t))dt
(1.5.7)
2 d
fort = h (t), all s G ( - e , e ) and a// it G C ( [ a , 6 ] , E ) . Then, for any
s s 0 0
solution u(t) of the Euler-Lagrange equations (1.1.4) for I ,
F {t,u(t),u(t))
p h {u{t))\
s s=0
'ds'
f c
+ ( F ( t , t i ( t ) , t i ( t ) ) - F (*,ti(*),A(*))wW) ^ 2 W I - = o p (1-5.8)
is constant in t G [a, 6].
Proof. We reduce the statement to the one of Theorem 1.5.1 by artifi

cially considering t as a dependent variable on the same footing with u.
Thus, we consider the integrand
F(t(T),u(t(T)),^,û(t(r))
:=F t M t ) , * ^ ) Z (1.5-9)
dr
\ dr
Then
I(t,u) := j H F ( i ( r ) , u ( i ( r ) ) , | : ( * ( r ) ) ) dr
= J F(t,u(t),u{t))dt, i f i ( r ) = a, f ( n ) = 6
0 (1.5.10)
= /()
By our assumption, F remains invariant under replacing (t,u) by

h (t,u). Consequently, Theorem 1.5.1 applied to I yields that
s
F {t,u(t)Mt))~h (u(t))\ ^
p s s=Q
with p standing for the place of the argument ^ of F (while p stands

as before for the arguments ii), is invariant. Since, by (1.5.9),
Fp
Fpi
Fo
p = F - Fu
p
at s = 0 (note ^ = 1 for s = 0 since h^t) = t), this implies the

invariance of (1.5.8).
q.e.d.
Example 1.5.3. Suppose F = F(u,u), i.e. F does not depend explicitly

on t. Then
h (t,z)
s = (t + s,z)
leaves / invariant as required in Theorem 1.5.2. Therefore, the 'energy'
F(t, u(t),u{t)) - F (t, u(t),

p u(t))u{t)
is conserved. We shall see another proof of this fact in Section 4.1.
Summary. The theorem of E. Noether identifies a quantity that is pre

served along any solution u(t) of the Euler-Lagrange equations of a
variational integral, a so-called first integral of motion, with any differ
entiable symmetry of the integrand. For example, in classical mechan
ics, conservation of momentum and angular momentum correspond to
translational and rotational invariance of the integral, respectively, while
time invariance leads to the conservation of energy.
Exercises
d
1.1 For mappings u : [a, 6] > E , consider
2
E(u) : = i f\u{t)\ dt
d l d
(| | is the Euclidean norm of E , i.e. for z (z ,..., z ),
1 2
| | 2 _ J2i i(z ) ).
z = Compute the Euler-Lagrange equations and
the second variation. Also, let
L(u) := I \ii{t)\dt.
Ja
Show that
L(u) < yj2(b-a)E(u),
Exercises 31
w i t h equality if \u(t)\ = constant almost everywhere. (What

is an appropriate regularity class for the mappings u that are
considered here?)
1.2 Determine all minimizers of the variational integral
w i t h u(~l) = 0 = u{l).
1.3 Develop a theory of Jacobi fields for variational problems with
free boundary conditions. I n particular, you should obtain an
analogue of Jacobi's theorem.
d
1.4 For mappings u : [a, 6] E , consider
Compute the first and second variation of / and the Jacobi

equation. Can you find Jacobi fields?
2
A geometric example: geodesic curves
2.1 T h e length and energy of curves

d
We let M be an n-dirnensional embedded submanifold of R . I n this
3
section, we assume that / is of class C , i.e. that all local charts are
thrice differentiable. We let c G AC([0,T],M) be a curve on M . This
means that c is an absolutely continuous map from the interval [0, T] into
d
R with the property that c(t) G M for every t G [0,T]. The derivative
of c w.r.t. t will be denoted by a dot ',
c(t) := Jt{t).
The length of c is given by
a y
L(c):=\c(t)\dt = ^ ( n j dt, (2.1.1)
1 d
where ( c , . . . , c ) are the coordinates of c = c(t). We also define the
energy of c as
2 2
E(c) := ^ \c(t)\ d t = \ Y . (<H (2.1-2)
We let now
f:U-~>V , f(U) = Mf)V
be a local chart for M as defined in Section 1.4. We assume for a moment

that c([0, T]) is contained in /(J7). Since / maps U bijectively onto f(U),
there exists a curve
7 ( t ) C C/
32
with
c(t) = f( (t)). 1 (2.1.3)
Since the derivative Df(z) has maximal rank everywhere (by definition
of a chart, cf. 1.4), 7 is absolutely continuous, since c is, and we have
the chain rule
c(t) = (Df) ( ( ) ) o 7 ( t ) , 7
or
where the index i is summed from 1 to n. Thus
< , 2
L(c) = (^(7W)7 ()^(7(*))7 '(*)) dt
and
T a a
1 f df df
In these formulae, and in sequel, the index is summed from 1 to d. For

zeU,
we put
a a
9f df
W i t h this notation, the preceding formulae become
j h
L{c)= I * (9iMt))fm (t)) dt (2.1.5)
Jo
and E{c) = J 9iMmW(t)dt (2.1.6)
Definition 2.1.1.
QfOt QfOt
is co//ed! /ie metric tensor of M w.r.t the chart f U >V.
We note that (gij(z)) i J = 1 n is symmetric, i.e.
z f o r a 1 1
9ij(*) =9ji( ) hi
and positive definite, i.e.
1 n
9ij{z)rfrf > 0 whenever 77 = ( 7 7 , . . . , rf) ^ 0 G E .
34 Geodesic curves
Remark 2.1.1. The use of local charts for M seerns to have the obvious
disadvantage that the expressions for length and energy of curves be
come more complicated. The advantage of this approach, namely not to
d
consider curves on M as curves in R satisfying a constraint, is that this
constraint now is automatically fulfilled. A l l curves represented in local
charts lie on M . This more than compensates for the complication in
the formulae for L and E.
Our aim will be to find curves of shortest length or of smallest energy

on M , i.e. to minimize the functionals L and E among curves on M . For
this purpose i t will be useful to observe certain invariance properties of
d d
L and E. First of all, whenever i : H& > H& is a Euclidean isometry, i.e.
d
i(y) = Ay + b with A G 0 ( d ) , the orthogonal group, and b G E , then
L(i(c)) = L(c) (2.1.7)
E(i(c)) = E(c) (2.1.8)
d
for any curve c : [0, T] -> R .
Secondly, L is parameterization invariant in the sense that whenever
T:[0,S]->[0,r]
l
is a diffeornorphisrn (i.e. r is bijective, and both r and its inverse r~
are everywhere differentiable), then
L(c) = L ( c o r ) ,
Namely
L(cor)
E, however, is not parameterization invariant. By the Schwarz inequality,

we have instead
2
L(c) = J l<\c(t)\dt/E(cj,
with strict inequality, unless
\c(t)\ = constant almost everywhere.
If
\c(t)\ = constant almost everywhere ,
we say that the curve c is parameterized proportionally to arc-length,

and if
\c(t)\ = 1,
we say that i t is parameterized by arc-length. We recall that a Jordan

d
curve, i.e. an injective curve c : [0, T] R , is rectifiable if i t is absolutely
continuous (which we always assume), and this implies that i t may be
parameterized by arc-length, i.e. there exists a diffeornorphisrn
r:[0,L(c)]^[0,T]
with
^ ( c o r)(s) = 1 for almost all s,
i.e. the reparameterized curve

C CO T
is parameterized by arc-length. From Lemma 2.1.1, we obtain:

d
Corollary 2.1.1. Let c : [0, L(c)] R be a curve parameterized on
[0,L(c)]. Among all reparameterizations
T:[0,L(C)}^[0,L(C)}
36 Geodesic curves
(i.e. we keep the interval of definition fixed, namely [0, L(c)]), the par
ameterization by arc-length leads to the smallest energy. Namely, if
d
c : [0, L(c)] E is parameterized by arc-length
L(c) = 2JE7(c), (2.1.12)
whereas for any other parameterization of c on the same interval,
L(c) < 2E(c). (2.1.13)
We now return to those curves c that are confined to lie on M , in order

to discover a third invariance. Namely,we compare the two expressions
(2.1.1) and (2.1.5) for the length of c, and similarly (2.1.2) and (2.1.6)
for its energy. (2.1.1) is obviously independent of the chart / : U V
and its metric tensor, and therefore (2.1.5) has to be independent of
them, too. I n order to study this more closely, let
f:U-+V
be another chart with
C ([0,T])c/(/).
Then there exists a curve 7 in U w i t h c(t) = /(7()) for all t. Putting

QfOt QfOt
we then also have
L{c) = (s(7(*))7*7 (*)) * dt. (2.1.14)
In order to study this invariance property more closely, we define
f == r 1
of-.r1
(/([/) n / ( # ) ) ^ r 1
(/([/) n /"(#))
(see Figure 2.1).

(p is called a coordinate transformation, (p is a diffeomorphism, i.e. a
n
bijective map between open subsets of E whose derivative Dp(z) has
maximal rank ( = n) at every z. Then from
/ o ( t ) = c(t) = / o 7 ( t ) ,
7
7(0 = ^(7(0), hence ?(t) = ^( (t))V(t)

7 (2.1.15)
and from
fob)) = f(z)
Figure 2.1.
we get
9ij(z) = ~9ki(<p(z))^(z)^j(z). (2.1.16)
From (2.1.15) and(2.1.16), we see
j
9iJ (7(0) (t)i (t) = hi m)) t W (*), (2-1.17)
and this shows again the equivalence of (2.1.5) and (2.1.1), and likewise
for the corresponding expressions of the energy. The important transfor
mation formula (2.1.16) shows how the metric tensor transforms under
coordinate transformations. This invariance property of L and E makes
it possible to express the length and energy of an arbitrary curve c on
M that is not necessarily contained in the image of a single chart as
follows:
One finds a subdivision
t = 0 < U < ... < t i

0 m <t =T
m
of [0, T] w i t h the property that
c([^_i,^])
is contained in the image of a single chart
U :U - v V v
for each v = 1 , . . . , m . Let (9ij(z)). j . ==1 n be the metric tensor of M

38 Geodesic curves
w.r.t. the chart / . Then

m
l
m f x
= E / (^(7,W)7iW7iW)'*
tiy 1
where c(t) = fvO~f (t) for t G

u By the preceding considerations,
this does not depend on the choice of charts / . For this reason, one
usually just says that for a curve c on M
T h
L(c) = f ( ^(t))f(t)jHt))
9ij dt, (2.1.18)
Jo
where 7 is the representation for c w.r.t. a local chart, and (9ij)ij=i,...,n
is the metric tensor of M w.r.t. this chart. Similarly
E(c) = \ j T ^ ( 7 ( * ) ) f ( * ) y (*)* (2-1.19)
We now assume that the charts for M are twice differentiable and return
to the question of finding shortest curves on M , for example between two
given points. By Corollary. 2.1.1, it is preferable to minimize E instead
of L , because a minimizer for E contains more information than one
for L ; namely, minimizers for E are precisely those minimizers for L
that are parameterized proportionally to arc-length. Thus, minimizing
E not only selects shortest curves but also convenient parameterizations
of such curves.
We now compute the Euler-Lagrange equations for E as given by
(2.1.19):
d
0= Eî for i = 1 , . . . , m
J
^ 0 = j t (2 (7(W (t))
9ij - ( ^ ) (7(*))7*(t)7>(*)
(the factor 2 in the first term results from the symmetry = gji)
f c
& 0 = 2 rf9i +2^y7 7> - îflyyV- (2.1.20)
We now introduce some further notation:
ij
(g ) ,
W / t,j = l,...,n
is the matrix inverse to (gij)ij i.e.

1 for i = k
9*9jk = 6k := for all i,
0 for i ^ k
and finally the Christoffel symbols
Equation (2.1.20) then becomes

il h k
0 = f 4- \g (2gi ,ki
j - 9kj ri V)
t
= f + ^ +Abu - ffiM) 7 * y
by using symmetries. Thus:
L e m m a 2.1.2. The Euler-Lagrange equations for the energy E for

curves on M are
i k
0 = f(t) + r ( (t)W(t)j (t)
jk 1 fori = l,...,n. (2.1.21)
The theorem of Picard-Lindelof about solutions of ordinary differential

equations implies:
n
L e m m a 2.1.3. For any z E U, v e K , the system (2.1.21) has a
unique solution y(t) with
7(0) = z , 7(0) = v for t E [e, e] and some e > 0.
Moreover, 7(2) depends differentiably on the initial values z, v.
Definition 2.1.2. The solutions of (2.1.21) are called geodesies on M.
Examples
Example 2.1.1. The sphere
in+l
is a differentiable manifold of dimension n. I n order to construct local

charts, we put
n n
:= S \ { ( 0 , 0 , . . . , 0 , l ) } , f i 2 :=S \ {(0,0,...,0,-1)}
40 Geodesic curves
and define
n n
gi : fix - E , g : fi -
2 2 E
as
and
+,
*" > - ( I ^ j t nSsr)
(#1 and g are the stereographic projections from the south and north
2
pole, respectively). We then obtain charts

1
/ 1 = j r : r - ^ \ { ( o 0,1)}
1 n
/ 2 = 2- :R"-5 \{(0,...,0,-l)}.
9
More explicitly, / i can be computed as follows:

With
1 n
/ 1 n , _ f X X \
2 n + 1 n + 1
1 = <* <* = * i(l
X X z z - " + ! ) 4-
x x X ,
hence
4-1
and then
2 z J
/0 = l1, . . . , n ^) .
1 4- ^
Thus
l n l {
i / 2z 2z zz - 1
,...,2
For the metric tensor, we compute
dfj_ 26 jk 4*V
2
dzk
~ (1 + z V ) '
Hence
()
9ij z = = ^-6ij. (2.1.22)
Actually, the metric tensor w.r.t. the chart f is given by the same 2
formula. I n order to compute the expression for geodesies, we also need

to compute the Christoffel symbols. I t turns out that adding a little
generality will actually facilitate the computations. We consider a metric
of the form
9ij = (2.1-23)
n +
where (\>: R E is positive and differentiable. Then
ij 2
g =<l> 8 . ij (2.1.24)
We also put
.
Then
_ % i _ _, _2__^0 _ _ r 2 dip
Next
kl
It,- = \g (9iu + 9ji,i ~ 9ij,i) (2.1.25)
= (9ikj + ?M - 9ij,k)
dip dip dip
= 6 i k 6 j k + 6 i j
- dzl~ ^ dz^'
Thus, vanishes if all three indices i, j , k are distinct, and for all i , j
T
)i T
= tj = - ^ j , a n d l ^ g fari^j. (2.1.26)
In the present case,
2
<p = log(l 4- | z | ) - log 2
hence
J
dip _ 2z '
j 2
dz~ " 1 + N "
42 Geodesic curves
Therefore, the equations for geodesies become

n n
i i
o = f + 2 r*,.(7) V' - r j ( ) 7 7 +
7 i 7 W W
(using the symmetry F\j = T^)
+
= V - 2
r r r V y V TTTP - - S W ( 2 L 2 7 )
1
1 + |7l j^x + 171
We now claim that the geodesic ( t ) through the origin, i.e. 7 ( 0 ) = 0,
7
n
with 7(0) = a R is given by
7() = aa(t), (2.1.28)
where a : E E then satisfies a ( 0 ) = 0, d(0) = 1. Making the ansatz

(2.1.28) in (2.1.27) leads to
3
i 2a a . .o 2a* a . 2
2 2 2
fr[l + a \a\ fril + a*\a\
2
if.. 2\a\ a . \ 2 . 1
1 L L
= a a - cr i = 1 , . . . , n.
2
V l + \a\ a* y
Since we may assume a ^ 0 (otherwise the solution with 7 ( 0 ) = a is
a point curve, hence uninteresting), this equation holds, if a(t) satisfies
the ordinary differential equation (ODE)
0 = a !-4a. (2.1.29)
2
1 4- |a| a
The theorem of Picard-Lindelof implies that (2.1.29) has a unique solu
tion in a neighbourhood of t = 0. We then have found a solution j(t) of
(2.1.27) of the desired form (2.1.28). The image of 7(f) is a straight line
through 0. By Lemma 2.1.3, we have thus found all solutions through 0.
The images of the straight lines under the chart / 1 are the great circles
n
on S through the south pole. We can now use a symmetry argument
n
to conclude that all the geodesic lines on S are given by the great cir
n
cles on S . Namely, the south pole does not play any distinguished role,
and we could have constructed a local chart by stereographic projection
n
from any other point on S as well, and the metric tensor would have
assumed the same form (2.1.22). More generally, one may also argue as
n
follows: We want to find the geodesic arc j(t) on S with 7 ( 0 ) = po,
n n
7 ( 0 ) = V for some p e S ,V
0 0 e T S. Let c (t) be the great circle on
0 Po 0
n
S parameterized such that Co(0) = po> co(0) = V . Co is contained in a
0
n + 1
unique two-dimensional plane through the origin in E . Let i denote
n + 1 n
the reflection across this plane. This is an isometry of R mapping S
n
onto itself. I t therefore maps geodesies on S onto geodesies, because
we have observed that the length and energy functionals are invariant
under isornetries, and so isometries have to map critical points to crit
ical points. Now i maps po and Vo to themselves. I f 7 were not invariant
under i , i o 7 would be another geodesic w i t h initial values po, Vo, con
tradicting the uniqueness result of Lemma 2.1.3. Therefore, 2 0 7 = 7 ,
and therefore 7 = c . 0
We draw some conclusions:

The geodesic arc through two given points need not be unique. Namely,
n
let p, q be antipodal points on 5 , e.g. north and south pole. Then there
exist infinitely many great circles that pass through both p and q.
n
We shall later on see that the first conjugate point of a point p S
along a great circle is the antipodal point q of p. One also sees by explicit
n
comparison that a geodesic arc on S ceases to be minimizing beyond
the first conjugate point, in accordance with Theorem 1.3.4.
2.2 F i e l d s of geodesic curves

d
Let M be an embedded, differentiate submanifold of E , or, more gen
3
erally, a Riemannian manifold of dimension nf, again of class C . Let
M q be a submanifold of M ; this means that Mo itself is a differentiable
d
submanifold of E , respectively a Riemannian manifold, and that the
inclusion i : M ^ M is a differentiable embedding. We assume that
0
3
M q has dimension n 1, and that i t is also of class C .
T h e o r e m 2.2.1. For any x M , there exist a neighbourhood V of

0 0
x in M, and a chart f : U * V with the following properties:

0
n
(i) U contains the origin o / E , / ( 0 ) = #o-
n
(ii) M nV
0 = f{UD{x = 0})
l
(Hi) The curves x = C\, C \ constant, i = l , . . . , n 1, are geodesies
n
parameterized by arc-length. The arcs 1 < x < 2 on any such
f We do not introduce the concept of an abstract Riemannian manifold here, but
some readers may know that concept already, and in fact it provides the natural
setting for the theory of geodesies. O n the other hand, the embedding theorem
of J.Nash says that any Riemannian manifold can be isometrically embedded into
d d
some Euclidean space E , hence considered as a submanifold of M . Therefore, from
that point of view, no generality is gained by considering Riemannian manifolds
d
instead of submanifolds of R .
44 Geodesic curves
n n
curve between the hypersurfaces x = 1 and x = 2 are M of
the same length 1 2
(iv) The metric tensor on U satisfies
9nn = 1, gin = 0 for all i = 1 , . . . , n - 1 (2.2.1)

l
(T/ie second relation means that the curves x = C\, i = 1 , . . . , n
n
1, intersect the hypersurfaces x = constant orthogonally.)
Proof. Since Mo is a hypersurface, for every p G Mo, there exist two

unit normal vectors n(p) to Mo at p, i.e.
n (p)eT M,
p
||n(p)|| = l
( n ( p ) , v) = 0 for all v G r M p 0 C T M.
P
In a sufficiently small neighbourhood V of #o, we may assume that such

a normal vector n(p) may be chosen so that it depends smoothly on
p G Mo f l V =: Vo. We assume that there is a local chart (p : Uo * VQ 0
71 1
for M q (Uo C M "" ), possibly choosing V smaller, if necessary. For every
p G Mo f l V , we then consider the geodesic arc 7 ( ) with P
7P(0) = P,
7P(0) = n ( p ) . (2.2.2)
This geodesic exists for |f| < e = e(p) by Lemma 2.1.3. By choosing
V smaller i f necessary, we may assume that e > 0 is independent of p.
Instead of 7 (), we write ~/(p,t). Since the solution of (2.2.2) depends
P
differentiably on its initial values (see Lemma 2.1.3), hence on p, the

map
/ :0b* (-e,)-M
(x, t) - > 7((^(x),f)

is likewise differentiable, where (p : C/o Vo is a local chart for Mo- We
may assume
x = y?(0),
0
by composing </? with a diffeomorphism if necessary. A t (0,0) G

Uo x (c, c), the Jacobian of / is spanned by the linearly independent vec
n n t e
tors a i ^ T ' M z ) ) ( o that 7(y>(a?),0) = y>(a?) and n(y?(x))
are orthogonal to all the vectors J-^r X ^ ^ M o , j = 1 , . . . , n 1). There

fore, by the inverse function theorem, / yields a chart in some neigh
bourhood U of (0,0) e Uo x (,). / obviously satisfies (i), (ii) (after
n
redefining V ) . (iii) also holds by construction (putting x = t). Next,
%
gnn = 1, since the curves x = c^, namely / ( c i , . . . , c _ i , ), t 6 (e, e), n
are geodesies parameterized by arc-length, hence g = = 1. nn
Finally, the system of equations for these curves to be geodesic is
-(dx
) + r * ^ dx
n 2 Sdx: t3 n n (* n
= *) ***=!,...,n.
Hence in particular
r* n = 0 for fc = l , . . . , n .
Now
1
r
1 M
n n = ^9 (l9nl,n ~ 9nn,l) = ^'ffnl.n,
since # n n = 1. Therefore
gnkjn = 0 for all k = 1,... , n .

1 n _ 1
Since furthermore ^ ( x , . . . , x , 0 ) = 0, because the geodesic arc
n l 1 n _ _ 1
x ~ t, x Ci ~ constant, is orthogonal to the surface ^ ( x , . . . , x )
1 n 1
= / ( x , . . . , x ~ , 0 ) , we obtain
9nk = 0.
DejRnition 2.2.1. T/ie coordinates whose existence is affirmed by The

orem 2.2.1 are called geodesic parallel coordinates based on the hyper-
surface M . 0
T h e o r e m 2.2.2. Let f : U V be a chart with the properties described

%
in Theorem 2.2.1. In particular, the curves x = Ci, Ci constant,, for
i = l , . . . , n 1 are geodesic arcs. Then any such curve is the short
est connection of its endpoints when compared with all curves contained
entirely in U and having the same endpoints.
Proof We consider the geodesic

n
7(t) = {x* = ^ , x = t, - e < t < c},
where U = / 7 x (-e, e). Let 7 ( f ) , t\ < t < t be another curve in U with

0 2
7 ( ^ 1 ) = 7 ( - e ) , 7 ( t ) = 7(e). We have to prove

2
Hi) > i(7), (2-2.3)

46 Geodesic curves
with strict inequality, unless 7 is a reparameterization of 7 . Now
(7) - ( 9ij ( 7 ( 0 ) 7 W ^ W + ( T ^ ) ) j (2.2.4)

/TL
V,J=I /
since # n n = 1, gi = 0 for i = 1,..., n - 1 by Theorem 2.2.1(iv),
n
> r n
i7 (o *>7 n
(<2)-7 n
(<i)=7 n
(e)-7 n
(-e)
= 1(7).
The first inequality is strict, unless 7* is constant for i = 1,... , n 1,

n
and the second one is strict, unless 7 ( ) is monotonic.
q.e.d.
Following Weierstrafi, we say that the geodesies

n
7() = {* = ,x
x C i = t,-e<t<e}
constitute a field of geodesies. Theorem 2.2.2 essentially says that any

geodesic arc i n this field is shorter than any other curve with the same
endpoints in the region covered by the field. Both properties are essential.
n
Namely geodesic arcs on S that are longer than a great semicircle show
that geodesies not embedded i n a field need not minimize the length
between their endpoints. A n d geodesic arcs on a cylinder, contained i n
meridians, but longer than a semicircle show that there may be shorter
curves not contained i n the field.
We observe that i f 7 ( f ) solves (2.1.21), so does y(\t) for A = constant.
We fix Zo G U and denote the geodesic arc 7 of Lemma 2.1.2 with
7(0) = s ,7(0) = t;0
by 7 . Then by the above observation
7*W=7A Q) V for A ^ O . (2.2.5)
Thus 7 A is defined on
v j], if 7 is defined on [c, e]. Since *y depends v
n
differentiably on v, and since v G E , \v\ = 1, is compact, there exists
0 > 0 with the property that for all v w i t h \v\ = 1, 7 is defined V
n
on [eo,eo]. From (2.2.5), we then conclude that for any w R with
M < eo, 7tu is defined on [1,1]. For later purposes, we also note that
by Lemma 2.1.3, CQ may be chosen to depend continuously on ZQ.
We now define a map

n
e = e Z0 : {w G E : \w\ < e } -+ U
0
w H- 7tu(l).
Then e(0) = z . We compute the derivative of e at 0 as
0
De(0)(v) = | 7 t (l)|
= ^7(%.o by (2.2.5)
= 7(0)
n
Hence, the derivative of e at 0 G E is the identity, and the inverse
mapping theorem implies:
n
T h e o r e m 2.2.3. e maps a neighbourhood of 0 G E diffeomorphically
l
(i.e. e is bijective, and both e and e~ are differentiable) onto a neigh
bourhood of ZQ G U. q.e.d.
We want to normalize our chart / : / V for M. First of all, we may

assume
z = 0 0 (2.2.6)
for the point zo G U under consideration. Secondly, the transformation

formula (2.1.16) implies that we may perform a linear change of coord
inates (i.e. replace / by / o A, where A G G L ( n , R ) ) in order to achieve
ffy(0) = fy. (2.2.7)
We assume that / : U V satisfies these normalizations. We then

n
replace / by / o e defined on {w G E : \w\ < e } . 0
T h e o r e m 2.2.4. In this new chart, the metric tensor satisfies
fti(0) = <i (2.2.8)
rj (0) - 0 -
fc ftj fc , (0) for all i, j , fc. (2.2.9)
Proof. B y (2.1.16), <ftj = 6ij holds, since the metric tensor w.r.t. the
chart / satisfies this property and De(0) is the identity by the proof
of Theorem 2.2.3. I n order to verify (2.2.9), we observe that in our new
n
chart, the straight lines tv (v G E , t \v\ < e) are geodesies. Namely, tv is
mapped to 7 ^ ( 1 ) = y (t) (see (2.2.5)), where *y (t) is the geodesic with
v v
48 Geodesic curves
initial direction v. We thus insert 7 ( f ) = tv into the geodesic equation

(2.1.21). Then 7 = 0, hence
i j k
T (tv)v v
jk =0 for i = l , . . . , n .
In particular, inserting t = 0, we get

n
Tj (0y f c = 0 for all v e E , i = 1 , . . . , n .
n
We use t; = e*, where {ei) l=1 n is an orthonormal basis of E . Then
17,(0) = 0 for a l i i a n d / .
l
We next insert v = \(ei 4- e ) , / ^ m . The symmetry TJ. = T - (which m fc h
l
directly follows from the definition of T - and the symmetry = g j) k k
then yields
rj (0)
m = 0 for a l i i , / , m .
The vanishing of gij^ for all i , j , then is an easy exercise in linear

algebra. q.e.d.
l n
D e f i n i t i o n 2.2.2. The local coordinates x ,...,x constructed before
a r e
Theorem 2.2.4 called Riemannian normal coordinates.
1 n
We let x , . . . , x be Riemannian normal coordinates. We transform
1 n l
them into polar coordinates r, y ? , . . . , (p ~ in the standard manner (e.g.
1 1 2 1
if n = 2, x = r c o s ^ , x = rsiny? ). This coordinate transformation
is of course singular at 0. We now express the metric tensor w.r.t. these
polar coordinates. We write g instead of gn, and we write g instead rr r(p
n
of gu, I = 2, . . . , n , and g instead of (9ki) ,i=2,...,d' * Particular, by
w k
Theorem 2.2.4 and the transformation rule (2.1.16)
ffrr(0) = l , f f r ( 0 ) = 0 . V (2.2.10)
The lines through the origin are geodesies by the construction of Rie
mannian normal coordinates, and in polar coordinates, they now become
1 n l
the curves (p ( y ? , . . . , <p ~ ) = constant; thus they can be written as
7(t) = (, (po) w i t h fixed (p . 0
Therefore, the geodesic equation (2.1.21) gives
r* = 0 r for all i
l
(where of course T rr stands for T ^ ) , i.e.
il
7}9 (29rt,r ~ g ,i) rr = 0 for a l i i ,
hence
2#w,r ~ 9rr,i = 0 for a l l / . (2.2.11)
Putting r I gives
grr = 0,
and w i t h (2.2.10) then

g rr = 1. (2.2.12)
Using this in (2.2.11) gives
hence w i t h (2.2.10) again

g rip = 0. (2.2.13)
We have thus shown:
T h e o r e m 2.2.5. In the preceding coordinates, so called Riemannian

polar coordinates, that are obtained by transforming Riemannian nor
mal coordinates into polar coordinates, the metric tensor has the form
(I 0 ... o x
Vo /
where # w stands for the (n 1) x (n 1)-matrix of the components of

1 71 1
the metric tensor w.r.t. the angular variables y ? , . . . , y? "" .
Note that this generalizes the situation for Euclidean polar coord
2
inates. The Euclidean metric on M , written in polar coordinates, e.g.
takes the form
Note that Theorem 2.2.5, in contrast to Theorem 2.2.4, is valid on the

whole chart, not only at the origin.
C o r o l l a r y 2.2.1. Riemannian polar coordinates are geodesic parallel

coordinates based on the hypersurfaces r = constant (r ^ 0, since r = 0
corresponds to a single point, and not a hypersurface).
Proof. By Theorem 2.2.5, all properties stated in Theorem 2.2.1 hold.

q.e.d.
50 Geodesic curves
By Corollary 2.2.1 and Theorem 2.2.1, the curves p = constant, n <

T < T2, are shortest connections between their end points among all
curves lying in the chart. We are now going to observe that this holds
even globally, i.e. also in comparison with curves that may leave the
chart:
T h e o r e m 2.2.6. For each p M, there exists Co > 0 with the property

that Riemannian polar coordinates centered at p may be introduced with
domain
{(r,^):0<r<e }, 0
Q may be chosen to depend continuously on p. We denote the subset of

M corresponding to this coordinate domain by B(p,o). For any e with
0 < e < o and any q G dB(p,e), there exists precisely one geodesic of
shortest length e) from p to q. Namely, if q has coordinates (e,(po),
this geodesic arc is given by y(t) = (t, po), 0 < t < e.
Proof. The first claim follows from Theorem 2.2.3, since Riemannian
polar coordinates are based on the diffeomorphism e (see the construc
tions before Theorems 2.2.4 and 2.2.5). As already noted before Theo
rem 2.2.3, Lemma 2.1.3 implies that .we may choose e as a continuous
0
function of p. I n order to verify the second claim, let c(t) be a curve from
p to g, w i t h c(0) = p. Let
t : = sup{ > 0 : c(r) G B(p, e)

0 for 0 < r < t}.
Since w.l.o.g. e > 0 and c is continuous, to is positive. We are going to

show that
(2.2.14)
Since the curve (,y>o)> 0 < f < e, has length e as easily follows from
Theorem 2.2.5, this will imply the claim. I n order to verify (2.2.14), we
proceed as follows:
(identifying cj [0 j w i t h its coordinate representation)

by Theorem 2.2.5 and since g w is positive definite (writing c(t) =

(r(t)Mt)))
\ \r\dt, again by Theorem 2.2.5

Jo
> / rdt = r{t ) = e. 0
Jo
Here, equality only holds i f g ifiip = 0, i.e. (p(t) = constant, r > 0, i.e. i f
w
C|[0 j is a straight line through the origin. The second claim now easily
follows. q.e.d.
Corollary 2.2.2. If M is compact, there exists e > 0 with the property

0
that for every p G M, there exist Riemannian polar coordinates with

domain
{(r,^):0<r<e }. 0
Proof. This follows from Theorem 2.2.6, since the constructions em

ployed for polar coordinates depend continuously on p (see essentially
the construction of the diffeomorphism e). q.e.d.
2.3 T h e existence of geodesies

Definition 2.3.1. Let M be a connected differentiable submanifold of
d
Euclidean space M , or, more generally^, a connected Riemannian man
ifold. The distance between p,q G M is
d(p,q) : = i n f { L ( c ) | c : [a, 6] M
rectifiable curve with c(a) = p, c(6) = q}.
T h e o r e m 2.3.1. Let M (as in Definition 2.3.1) be compact. There

exists eo > 0 with the property that any two points p, q G M with
d(p, q) < to
can be connected by a unique shortest geodesic arc (i.e. of length d(p,q)).
This geodesic arc depends continuously on p and q.
Proof. We take eo as described in Corollary 2.2.2. This gives a unique

shortest geodesic arc from p to q which furthermore depends contin
uously on q. Exchanging the roles of p and q then yields continuous
dependence on p, too. q.e.d.
f See footnote on p. 43.
52 Geodesic curves
We now proceed to establish a global result:
T h e o r e m 2.3.2. Let M be a compact connected differentiable subman-

ifold ofW*, or, more generally, a compact connected Riemannian man
ifold. Then any two points p, q G M can be connected by a shortest
geodesic arc (i.e. of length d(p,q)).
Proof. Let ( c ) N be a minimizing sequence. We may assume w.l.o.g.

n n
that all c are parameterized on the interval [0,1] and proportionally to

n
arc-length. Thus
Cn(0) = p, c ( l ) = q, n
L(c )
n > d(p, q) for n oo.
For each n, we may find
^0,n = 0 < i ?n < ... < , = 1 m n
with
w i t h e given by Theorem 2.3.1. By Theorem 2.3.1, there exists a unique

0
shortest geodesic arc between c (tj-i, ) =: P j - i , and c (tj ) = : P j , .

n n n n jTl n
We replace c \ n [t ^ j by this shortest geodesic arc and obtain a

t
new minimizing sequence, again denoted by c , that now is piecewise n
geodesic. Since the length of the c are bounded because of the mini
n
mizing property, we may actually assume that m is independent of n .

Since M is compact, after selecting a subsequence of c , the points pj n yTl
converge to limit points p^, ( j = 0 , . . . , m) as n oo. c j ^ the n [t i t
unique shortest geodesic arc between P j _ i , and P j , , then converges to n n
the unique shortest geodesic arc between Pj-\ and pj (for this point, one
verifies that limits of geodesic arcs are again geodesic arcs, that limits
of shortest arcs are again shortest arcs, that d(pj-\,pj) < Q, and one
uses Theorem 2.3.1). We thus obtain a piecewise geodesic limit curve c,
with c(0) = p, c ( l ) = g, and
L(c) = l i m L ( c ) , n
noo
since we have for the geodesic pieces
L c = L c
( iii-..,.) B ^ ( i^-.-^-i)
for all j (tj = l i m ^ , ) . Since the c constitute a minimizing sequence,

n n
noo '
L(c) = d(p,g),
and c thus is of shortest possible length. This implies that c is geodesic.

c
Namely, otherwise we could find 0 < s\ < s < 1 w i t h L ( | ) < e,
2 [ s i ) S 2 ] 0
but w i t h C |
( S Inot being geodesic. Replacing c j by the shortest{ai 8 2 )
geodesic arc between c(s\) and c(s ) would yield a shorter curve (cf.
2
Theorem 2.2.6.), contradicting the minimizing property of c.

q.e.d.
Thus, any two points on a compact M may be connected by a shortest

geodesic. We now pose the question whether they can be connected by
n
more than one geodesic, not necessarily the shortest. On 5 , for example,
this is clearly the case. Actually, the answer is that i t is the case on any
compact M. That result needs a topological result that is not available
to us here, however. Therefore, we will restrict ourselves to a special
case which, however, already displays the crucial geometric idea of the
construction for the general case, too.
T h e o r e m 2.3.3. Let M be a differentiable submanifold of Euclidean

space W*, (or more generally^, a Riemannian manifold), diffeomorphic
2
to the sphere S . The latter condition means that there exists a bijective
map
2
h:S ~*M
that is differentiable in both directions. Then any two points p, q G M

can be connected by at least two geodesies.
2
Proof. M is compact and connected since diffeomorphic to S which is
compact and connected. Let us assume p ^ q. We leave it to the reader
to modify our constructions in order that they also apply to the case
p = q. (In that case, T h m 2.3.3 asserts the existence of a nonconstant
geodesic c : [0,1] > M w i t h c(0) = p = c(l).) One may then construct
a diffeomorphism
2
ho : S - M
w i t h the following properties:

2 2 3 3
Let S = { ( \ , ) G M : |x| = l } . Then
p = M0,0,l), g = MO, 0,-1)

and a shortest geodesic arc c : [0,1] M w i t h c(0) = p, c ( l ) = q is
given by
c(t) = M0,sin7r,cos7r).
f See footnote on p. 43.

54 Geodesic curves
Let us point out that these normalizations are not at all essential, but
only convenient for our constructions. We look at the family of curves
7(,s) = /io(sin27rssin7r, cos27rssin7r,cos7r), 0 < s,t < 1. (2.3.1)
Then
7 ( * , 0 ) = 7 ( M ) = c(t) for all t
and
7(0, s) = c(0), 7(1, s) = c ( l ) for all s.
We find some number K w i t h
L{n(-,8))<K for all s. (2.3.2)
Redefining the parameter t, we may also assume that all curves 7(-,s)
are parameterized proportionally to arc-length. By Theorem 2.3.1, there
exists o > 0 such that the shortest geodesic between any p, q G M, w i t h
d(p, q) < o is unique. Let
0 = t < *i < ... < t

0 m = 1
be a partition of [0,1] w i t h
h ~ tj-i < for j = 1 , . . . , m . (2.3.3)
Let another partition ( T I , . . . , r ) satisfy m
r
To = t < l < h < T2 < < T
0 M < tm = T m + i
and
T j T j 1 < :
~ ~ ? forj = l,...,m + l. (2.3.4)
I f 7 : [0,1] M is any curve parameterized proportionally to arc-length

with
L(l) < K,
we then have for j = 1 , . . . , m
d ( 7 ( ^ - 1 ) ,7(*j)) < I ( T | [ t j _ l ! ( j ) ) < * f =
Therefore, by Theorem 2.3.1, the shortest geodesic from y(tj-i) to y(tj)

is unique. We then define 7*1(7) to be that piecewise geodesic curve for
which r i ( 7 ) j [t i t ] coincides w i t h the shortest geodesic from 7(j_i)
to 7(^j), j = 1 , . . . , m . Likewise, we let ^ ( 7 ) by that piecewise geodesic
curve for which r (7) | . 2 [r r j coincides w i t h the again unique short
est geodesic from 7 ( T J _ I ) to 7 ( T J ) , j = 1 , . . . , m + 1. We now observe:
L e m m a 2 . 3 . 1 . Suppose d(7(fy)7(fy-i)) < e and d(y(Tj),7(r _i)) 0 7 <

Q / o r a// j .
r(7) : = r o r i ( 7 ) 2
sais/ies
L(r( ))<L( ) 7 7 (2.3.5)
equality iffy is geodesic.
Proof. By uniqueness of the shortest geodesic between y(tj-\) and 7(j),

we have
i(ri (7))<i(7)
w i t h equality only i n case
7*1(7) = 7-
Likewise, for every curve 7', L (7,' ] < Q for all j ,
L(r ( '))<L(V)
2 7
w i t h equality only in case
; ;
r (7 ) = 7 -
2
Therefore
L(r( ))<L( ) 7 7
w i t h equality only i f
r(7) = 7.
If r ( 7 ) = 7, however, 7| [t and 7 | [t I T J are geodesic for every j ,

= s
and hence 7 is geodesic itself. (If 7*1(7) 7, then 7 i piecewise geodesic
w i t h corners at most at the j, and i f r ( r i (7)) = r\ (7), then r i ( 7 ) is
2
geodesic w i t h corners at most at the Tj. Thus, i f r(y) = 7, 7 cannot have

any corners at all.)
q.e.d.
L e m m a 2.3.2. Let 7 : [0,1] + M be a curve parameterized propor

n
tionally to arc-length and with L("y) < K. Then a subsequence of> (7)
( = ro...or(7)) converges uniformly to a geodesic with the same endpoints
as 7.
56 Geodesic curves
N
Proof. Each curve r ( 7 ) , n N , is a piecewise geodesic w i t h corners
n N N N
r 7 ( n ) , . . . , r 7 ( r ) and endpoints r 7 ( r ) = 7 ( 0 ) , r 7 ( r + i ) = 7 ( 1 ) .
m 0 m
The individual segments are the unique shortest connections between

these points. Therefore, each such curve is uniquely determined by the
m-tupel
n
A n : = = ( n r 7 ( T l ) > _ ? r 7 (r )) GMx... x M .
m
m times
n
Since M is compact, a subsequence of A converges to some limit
( p i , . . . , P m ) 6 M x ... x M .
N
r ( 7 ) then converges uniformly towards the piecewise geodesic 7 0 w i t h
endpoints 70(0) = 7(0)>7o(l) = 7(1) and nodes 70(r<) = Pi (i =
1,... , m ) w i t h segments 7 | 0 { T i t } being the shortest geodesic arcs be
tween their endpoints. This follows from the continuous dependence of
the occurring geodesic arcs on their endpoints (Theorem 2.3.1). We de
N
note the convergent subsequence of ( r ( 7 ) ) N N by ( 7 J , ) . For all v N
g N
then
n M
7 l / + 1 = r 7 with n[y) 6 N .
By the minimizing property of the subsegments of the 7 ,
L r
OHT,-!,^)
= d
( l v ( i ~ i ) 7i/ fo)) >
hence
ra+1
L
(7^) = <*(7/ ,7 (rj)).
Since ^ ( T J ) converges to pj = 70 (T7), L(7i,) converges to

ra+1
L
(7o) = d (7o ( T J - I ) , 7 0 fa-))
for v oo. Then also
n
L ( ) = lim L (
7 o 7 l / + 1 ) = l i m L(r <"> ) 7
i/oo 1/oo
< l i m L(7i/) by Lemma 2.3.1
v>oo
= i(7o),
and equality has to hold throughout. Moreover, r ( 7 ) converges to r ( 7 ) , 0

and
L(r( o))=
7 Urn L ( r ( ) )7
Voo
N
> lim L f r ^ 7 ^ ) by Lemma 2.3.1 again
= (7o).
Lemma 2.3.1 then implies that 7 0 is geodesic.
q.e.d.
We now return to the proof of Theorem 2.3.3:

We apply the preceding curve shortening process to all curves 7 ( - , $ ) ,
N
s [0,1], simultaneously. For each 5, a subsequence of r 7 ( - , s ) then
converges to a geodesic from p to q. We want to exclude the situation
that all those limit geodesies coincide w i t h c. Let
K 0 := L(c),
and
N
K\ : = sup lim L(r 7(-,s)).
N
Since 7 ( - , 0) = c(-) is geodesic, r 7 ( - , 0) = 7 ( - , 0) for all n, hence K\ > K . 0
We distinguish two cases:
(1) K\ > K 0
N
Since 7 ( - , s) is continuous in s, so is r 7 ( - , s) for every n G N . We
now claim:
Whenever
N
s u p L ( r 7 ( - , s ) ) < KI + e (2.3.6)
there exists s n [0,1] w i t h
n n + 1
L ( r ( - , *)) - i ( r
7 7 ( - , )) < 2c (2.3.7)
and
n
L(r (-, ) ) > ! - .
7 (2.3.8)
Indeed, otherwise
n + 1
supL(r 7 (-,)) < i -c,
N + 1
contradicting the definition of K\ (note that sup L ( r 5 7 ( - , $))
58 Geodesic curves
is monotonically decreasing in n by Lemma 2.3.1). By definition

of K i , there exists a subsequence ( e ) N - * 0 w i t h n n
n
s u p L ( r 7 ( - , 5 ) ) < fti + e . n
3
n
A subsequence of ( r 7 ( - , s ))neN has to converge to some limit
n
curve c as above, and because of (2.3.7) w i t h e = e , we conclude n
as in the proof of Lemma 2.3.2 that
L(r(c)) = L(c),
and c is hence geodesic by Lemma 2.3.1. Because of (2.3.8) and

continuity of L in the limit as in the proof of Lemma 2.3.2, we
get
L(c) K\.
Since c and c are both defined on [ 0 , 1 ] and have different lengths,

they have to be different curves. Thus, c is the desired second
geodesic.
(2) K\ = K0
We are going to show that in this case, there even exist infinitely
many geodesies from p to q. For that purpose, we consider the
curve
This is a closed curve w i t h 7 ( 0 ) = 7 ( 1 ) = c(\) (see Figure 2.2).

n
Since ho is a diffeomorphism and r 7 ( t , s) is obtained through a process
that can easily be made continuous from
7(^,5) = /io(sin27T5sin7rt,cos27T5sin7rt,cos7rt),
n
r 7 ( , s ) has to map [ 0 , 1 ] x [ 0 , 1 ] surjectively onto M . Therefore, for
every n N and every 5 G [ 0 , 1 ] , there exists cr (s) w i t h n
n
7(5) G r 7 ( . , a ( ) ) =:7n,s(0
n 5
n
(in other words, r 7 ( - , o - ( s ) ) is a curve passing through 7 ( 5 ) ) . 7 ,s(*)
n n
then is a curve with
7n,a(0) = C ( 0 ) = p , 7 n , ( l ) = c ( l ) = q,
and because of K \ = KQ, we obtain

n
lim L ( , ( . ) ) 
Figure 2.2.
After selection of a subsequence, (7n,s(*))nN again converges to some

limit curve c (-) w i t h
s
c (0)
s =p,c (l) = q
a
and
By (2.3.5),
and since K,Q is the infimum of the energies of all curves from p to q
(o = L(c), and c is minimizing), c (-) is a minimizing curve itself,
s
hence geodesic.
Therefore, we have shown that for every 5, there exists a geodesic
from p to q that passes through 7(5). Hence there exist infinitely many
geodesies from p to g, as claimed.
q.e.d.
Remarks:
(1) Lemmas 2.3.1 and 2.3.2 do not need that M is diffeomorphic to

2
S . Compactness suffices.
60 Geodesic curves
(2) We may construct the curves 7 ,s(*) at the end of the proof also
n
in case K\ > K . I n that case, however, limits of such curves need

0
not be geodesic anymore.

(3) See Section 3.1 for an abstract version of the argument at the end
of the preceding proof.
Exercises
2 2 2 2
2.1 For curves j(t) = (j\j ) : R -+ {(x\x ) e E |x > 0 } , con
sider
Compute the Euler-Lagrange equations and determine all solu

tions.
2.2 For curves
d
1 d d d 2
7 ( t ) = ( 7 , , 7 ) : R - {{x\ . ..,x ) R\ 5 > ' ) < 1},
1=1
consider
Compute the Euler-Lagrange equations and determine all solu

tions.
2.3 Determine all geodesies between two given points on a cylinder
3 2 2
{(x,y,z)eK :x + y ==l}.
3
2.4 Let E be a surface of revolution in E , i.e.
3 2 2
Z = {(x,y,z)em :x +y =f(z)}
for a smooth, positive / : E E. What can you say about

geodesies on E? For example, are the curves (x, y) = constant
geodesies? When are the curves z = constant geodesies?
n
2.5 Determine Riemannian polar coordinates on the sphere S with
a domain of definition that is as large as possible.
2.6 Let p be the center of Riemannian polar coordinates on M , w i t h
d
domain of definition {v G E : ||v|| < g}. Let c : [0, e] M be
a geodesic with c(0) = p that is parameterized by arc-length,
0 < e < Q. Show that c([0,c]) does not contain a point that is
conjugate to p.
Exercises
d
2.7 Let M be a differentiable submanifold of R that is diffeomor-
2
phic to S . Show that for any p G M , there exists a nonconstant
geodesic c : [0,1] M with c(0) = c ( l ) = p.
2.8 Try to find other topological classes of manifolds w i t h the prop
erty that there always exists more than one geodesic connection
between any two points.
3
Saddle point constructions
3.1 A f i n i t e d i m e n s i o n a l e x a m p l e
d 1
Let F : E E be a function of class C which is bounded from below
and which is 'proper' in the following sense:
F{x) -+ oo for |x| - oo. (3.1.1)
Since F is bounded from below, (3.1.1) is equivalent to: For every s G E,

d
{x G E : F(x) < $} is compact. (3.1.2)
Therefore, F assumes its infimum. Namely, we take any
s > inf
0 F(x).
d
xR
Then
d
{x G E : F(x) < so}
is compact and nonempty, and since F is continuous, it has to assume

its infimum on that set. We now assume that F even has two relative
d
minima, x i , #2 in E , and that they are strict in the following sense: For
x = # i , #2, we have
3<5 Vt/
0 with 0 < \y-x\ < 6 : F(y) > F(x).
0 (3.1.3)
T h e o r e m 3 . 1 . 1 . Under the above assumptions, F has a third critical

point 3 (i.e. VF(xs) = 0) with
F(x ) 3 > m a x ( F ( x i ) , F ( x ) ) = : o
2
d
Proof. We consider curves 7 : [0,1] R with
7(0) = x i , 7 ( 1 ) = * 2 . (3.1.4)
62
We first observe that there exists a > 0 with the property that for any
such curve, there exists t (0,1) with
0
F^(t ))>K 0 0 + a. (3.1.5)
In order to verify this, we may assume w.l.o.g.
F(xx) < F(x ). 2
We then choose <5 w i t h
0 < 6 < min((5 , ^ | x i - x | ) . 0 2 (3.1.6)
For every y with |y - x | = <5 then by (3.1.3)

2
F(y) > F(x ), 2
and since {\y x \ = 6} is compact, F assumes its minimum on this set,

2
hence for some a > 0
min F(y)> F(x )+a 2 = K + a.

0 (3.1.7)
\y-x \=S
2
Since for every curve 7 w i t h (3.1.4) we have
| ( 1 ) - x \ = 0, | ( 0 ) - x \ = \xi -
7 2 7 2 x \,
2
there has to exist some t [0,1] with

0
\7(t ) - x \ = 6
0 2 (recall (3.1.6)) .
By (3.1.7) then
F ( ( ^ o ) ) > o 4- a,
7
and (3.1.5) follows indeed.

We now define
K\ : = inf sup F(i(t)),
7
*[0,1]
d
where 7 again is a curve in R with 7(0) = x i , 7(1) = x . By (3.1.5) 2
i > K . 0 (3.1.8)
Our intention now is to find a critical point X3 of F w i t h
F(x ) = i. 3
Since
F(xi),F(x )< , 2 0
64 Saddle point constructions
#3 will then be necessarily be different from X\ and #2- As a step towards

the existence of such a point 3 , we claim
Ve > 0 3<5 > 0 V curves 7 w i t h 7(0) = # 1 , 7 ( 1 ) = x 2
with
sup F(7(0) <n\+6 (3.1.9)
t[0,l]
31 0 [0,1] with:
F(-y(t ))0 > i - 6 (3.1.10)
|(VF)( (*o))|<c.
7 (3.1.11)
Suppose this is not the case. Then
3to > 0 Vn N 3 curve 7 between Xi and x2 with
supF(7(t)) KI -

0 0 (3.1.13)
|(VF)(7(o))|> . 0 (3.1.14)
For s > 0, we define a new curve 7 i S by
7,.(<) := 7n() - (VF)(7(*))-
Since x\ and #2 are minima, VF{x\) 0 = VF(#2)> and so
7n,*(0) = xi,7 , (l) n 5 = x , 2
so that the curves 7 are valid comparison curves. By our properness

n > s
d
assumption (3.1.2) and (3.1.12), 7 ( ) stays in a bounded subset of E ,
n
and VF will then be bounded on that bounded set, and hence for any
So > 0 and all 0 < 5 < so, the curves 7 ,s(0 stay in some bounded n
set, too. This set is independent of n (as long as 0 < 5 < 5 , for fixed 0
SQ > 0). By Taylor's formula
F(7n,,(0) = F( (t)) ln - sVF( (t)) ln V F ( ( ) ) + o(s).

7
Since F is continuously differentiable and 7 , s ( ) is contained in a bound n
ed set, 0(5) can be estimated independently of n and t (as long as 0 <

n
s < s ) . * Particular, after possibly choosing s > 0 smaller,
0 0
2
F(7n,.W) < ^(7(*)) - I |VF(7n(<))| (3.1.15)
for all n, s w i t h 0 < s < s , and t w i t h 0
|VF( (*))| > c .

7 n 0 (3.1.16)
Thus, in particular,
F( ,s (t))
ln 0 < F( (t)) ln - ^4 (3.1.17)
for all such t and all n. We now simply choose n so large that
i < f 4 (3.1.18)
Then by our assumption, all t w i t h F ( 7 ( ) ) > i - e

0 n 0 0 satisfy (3.1.14),
and hence for all such to
2
F(ln,s (to))
0 < F( (*o)) -
7 |e
<i + - - ? o by (3.1.12) (3.1.19)
< i by (3.1.18).
Having proved (3.1.19), there are now various ways to construct a path
7 from X\ to X2 w i t h
F(7(*)) < i for all * [0,1]. (3.1.20)
One way is to refine the above construction by letting s depend on t as

follows: we choose a smooth function
v(t):[0, l]-[0,* ] 0
with
a(t) =0 whenever F(y (t))

n < K\ e 0
and
a(t) = s 0 whenever F(y (t))

n > K\- ~.
We then look at the path 7 ( f ) = 7 ,<7(t)(0- Then for t w i t h F ( 7 ( ) ) <

n n
?
F(7(0) = i ( 7 n ( 0 ) 
K l
~ 2
F( (t)) = F(7n,. (*))<!
7 0 (cf- (3.1.19)).
Thus, (3.1.20) holds indeed. This, however, contradicts the definition of

i . Therefore, the assumption that our claim was not correct led to a
contradiction, and the claim holds. I t is now simple to prove the theorem.
Namely, we let e > 0 for n oo, and for e = e , we find <5 = 6 as i n
n n n
the claim. We than choose a curve 7 from x\ to x% with n
sup F ( 7 ( * ) ) Ki-e
n n (3.1.22)
|(VF)( (* ))|<e .
7 n n n (3.1.23)
/
After selection of a subsequence, ( y (tn))neN then converges to some
n
point x , because of (3.1.2) and (3.1.21). x$ then satisfies by continuity

3
of F and V F
F ( x ) = i
3 (3.1.24)
V F ( x ) = 0. 3 (3.1.25)
Thus, 3 is the desired critical point.

q.e.d.
Theorem 3.1.1 may be refined as follows:
T h e o r e m 3.1.2. Let F as above again have two relative minima, not

necessarily strict anymore. Then either F has a critical point x$ with
F(x )
3 > max(F(xi),F(x )) = , 2 0
or it has infinitely many critical points.
Proof. For the argument of the proof of Theorem 3.1.1, we only need
inf sup F ( 7 ( * ) ) > , 0 (3.1.26)

7
t[0,l]
d
where the infimum again is taken over curves 7 : [0,1] E with 7(0) =
# i , 7(1) = #2. So, suppose that (3.1.26) does not hold. We then want to
show the existence of infinitely many critical points. As in the proof of

Theorem 3.1.1, we may assume
F(x )
x < F(x ). 2
The argument at the beginning of the proof of Theorem 3.1.1 then shows
that (3.1.26) holds if x is a strict relative minimum. I f x is a relative
2 2
minimum, which is not strict, for all sufficiently small 6 > 0, say <5 < <5Q,
we have
F(x )2 < F(x) for all x with \x - x \ < <5

2 0 (3.1.27)
and there always exists some x$ with 0 < \xs x \ < 6 and 2
F(x )
6 = F{x ). 2 (3.1.28)
We then put 8\ = <5 /2. Then xs is a relative minimum of F by (3.1.27),

0 x
(3.1.28), hence a critical point. Having found a critical point xs w i t h n
0 < \x - x \ < \x _
6n 2 - x | , we put
6n 1 2
<Wl = \ \x6 n - X\2
and find a critical point xs n+1 with
0 < \xg n+1 - x \ < <5 i.

2 n+
Thus, xs n+1 is a critical point of F different from all preceding ones.

q.e.d.
Remark. I t is not very hard to sharpen the statement of Theorem 3.1.2

from 'infinitely many' to 'uncountably many'.
3.2 T h e construction of L y u s t e r n i k - S c h n i r e l m a n
In this section, we want to prove the following theorem, in order to ex
hibit some important global construction in the calculus of variations, in
troduced by Lyusternik-Schnirelman. The result presented is much more
elementary than the theorem of Lyusternik-Schnirelman, which says
that on any surface w i t h a Riemannian metric, e.g. a surface embedded
in some Euclidean space, diffeomorphic to the two-dimensional sphere,
there exist at least three closed geodesies without self-intersections. The
more elementary character of our setting allows us to bypass essential
geometric difficulties encountered in a detailed proof of the Lyusternik-
Schnirelman Theorem.
Figure 3.1.
1
T h e o r e m 3 . 2 . 1 . Let 7 be a closed convex Jordan curved of class C in
2
the plane E . (7 then divides the plane into a bounded region A, and an
unbounded one, by the Jordan curve Theorem. That 7 is convex means
that the straight line between any two points of 7 is contained in the
closure A of A.) Then there exist at least two such straight lines between
points on 7 meeting 7 orthogonally at both end points (see Figure 3.1).
Proof. We start by finding one such line. Let C be the set of all straight
lines / in A w i t h dl C 7. We say that a sequence (l )neN C converges
n
to / E , i f the end points of the l converge to those of /. I n order to

n
have a closed space, we allow lines to be trivial i.e. to consist of a single

point on 7 only. We denote the space of these point curves on 7 by - 0
We let / : = [0,1] be the unit interval. We consider continuous maps
v:I-*C
with the following two properties:
(i) v(0) = v(l).

(ii) To any such family, we may assign two subregions A\(t) and A (t) 2
of A in a certain manner. Namely, we let A\(t) and A (t) be the 2
two regions into which v(t) divides A. Having chosen A\(0) and
A (0), A\(t) and A (t) then are determined by the continuity
2 2
d
t A closed Jordan curve is a curve 7 : [0, T] R with 7 ( 0 ) = 7(T") that is injective
on [0, T ) . Cf. the definition of a Jordan curve on p. 35.
Figure 3.2.
requirement. We then require
A (1)
1 = A (0).
2
We let Vi be the class of all such families v.

The construction is visualized in Figure 3.2. (0 corresponds to 0 J,
/to\J/ to I / / / to | , 1 to 1)
Actually, in order to simplify the visualization, if v(0) is a point curve
(on 7), i) may be relaxed to just requiring that v(l) also is a point curve
(on 7), not necessarily coinciding with v(0) (see Figure 3.3). Namely, any
point curves can be connected through point curves, i.e. w i t h vanishing
length.
We denote by L(l) the length oil C and define
K\ : = inf
v y
supL(v(t)).
î tei
Figure 3.3.
We want to show that
i > 0.
For this purpose, let p > 0 be the inner radius of 7, i.e. the largest p for
which there exists a disc
B(x p) 0l C A
2
for some XQ G A (B(xo,p) := {x G E : \x - x | < p})- Then
0
i > i : = inf sup L(v(t) P\ B(XQ, p)).
We let Aft) : = ^ ( t ) n B(xo,p), i = 1,2. Because of (ii) and the

f
continuous dependence of J4*() and hence also of A (t) on t, there exists {
some to I w i t h
Area ( t ) ) = Area
0 (A' (t )).
2 0
Thus v(to) divides B(xo,p) into two subregions of equal area. v(to) then
has to be a diameter of B(xo,p), i.e.
L(v(to)nB(x ,p))=2p.0
Therefore
i > i = 2p > 0
and Ki is positive indeed. We are now going to show by a line of reasoning

already familiar from Sections 2.3 and 3.1 that K\ is realized by a critical
point / of L among all lines w i t h end points in 7, i.e. by / meeting 7
orthogonally (see Theorem 1.4.1). For that purpose we shall assume for
3
the moment that 7 is of class C . Later on, we shall reduce the case
1
where 7 is only C to the present one by an approximation argument.
We now claim
Ve>0 3(5 > 0 : Vv Vi w i t h
supL (v(t)) < Ki+6

tei
3 to G I with L (v(to)) > i - c
and |cos(ai (v (t )))\

0 , |cos ( a {v (t )))\
2 0 < c,
where a\(l) and a (l) 2 are the angles of / at its endpoints w i t h 7.

Otherwise
3e > 0 : V n G N
0 3v n G V\ w i t h
s u p L ( v ( t ) ) K I
0 n o
e
|cosai (v (*o))| > o
n

or |cosa2 (vn(*o))| > o-
The idea to reach a contradiction from that assumption is simple, once

the following Lemma is proved:
3
L e m m a 3 . 2 . 1 . For every planar closed Jordan curve 7 of class C ,
2
there exists (3 > 0 with the following property: Whenever x G E satisfies
dist(x,7) : = inf \x y\ < (3

ye-y
there exists a unique y G 7 with dist(#,7) = \x y\.
Proof. We consider 7 as an embedded submanifold of the Euclidean

2
plane E . 7 is then covered by the images of charts / : U V of the
2
type constructed in Theorem 2.2.1. Here, U and V are open in E , and
7 n v = f (u n {x = 0 } ) .
2
1
Furthermore, the curves x = constant in U correspond to geodesies, i.e.
straight lines in V perpendicular to 7, and they form shortest connec
tions to 7 fl V. By shrinking U, i f necessary, we may assume that i t is of
the form ( - , ) x (-77, rj), w i t h > 0, rj > 0. Since 7 is compact, i t can
be covered by finitely many such charts
: y
fi ( - 6 , 6 ) x (-WiVi) ~* i , i = l,...,m.
x 1 1
If we then restrict fi to ( - 6 , 6 ) ("f ' ^J, lines x = constant,
2
~ k < x < ^ , then correspond to shortest geodesies to 7, since the part
of 7 not contained in V{ is not contained in the image of fi, and hence has
distance at least ^ from the image of the smaller set ( & ) x , ).
This is indicated in Figure 3.4 where the broken lines correspond to
2
x = ^ and this is depicted for two different indices i.
Therefore, (3 : = min ( ^ ) satisfies the claim.
i=l,...,n
q.e.d.
Saddle point constructions
Figure 3.5.
We now return to the proof of Theorem 3.2.1:
Without loss of generality eo < 0 < Assume e.g.
cosai (v (to)) > e .

n 0
The following construction is depicted in Figure 3.5. Choose si(to)

v (to)
n with
|ai(*o)-Pi(to)|=0,
where p\(to) is the endpoint of v (to) where i t forms the angle a\(to)
n
w i t h 7. We replace the subarc v^(to) of v (to) between p i ( t ) and si(o)n 0
f
by the shortest line segment v (to) from s\(to) to 7. By the theorem of
n
Pythagoras and the convexity of 7

l
L (v' (t )) < L (y (t )) s i n a i (v
n 0 n 0 n (t ))
0
<L(vi(t ))yfl^.
0
We then let
Vn(to)
be the straight line from the second endpoint p (to) 2 of v (to)n to the
f
point where v (to) meets 7. Then, letting v(to)
n denote the segment of
v (to)
n between s\(to) and p (to), 2 by the triangle inequality
2
L(v* (t ))<L(v' (t ))
n 0 n 0 + L{v (t ))
n 0
<L{vl{t*))J^l + L{vl(t ))Q
2
= f 3 ^ 4 + L(v (t )).
n 0
Since L (y\ (to)) = /?, we have
2
L(v (t ))n 0 < KI-/? +
We then choose n so large that
Py/l ~4 + Ki-0+^<Ki-ri
for some rj > 0. Hence
M < (*(>)) <*!-*/.

We now continuously select points si(), 52(t) on v ( t ) for every t e I n
with
\Pi(t) - Si(t)\ = 0, whenever L (v (t)) > Ki - 0

n
and
Pi(t) = Si(t), whenever L (v (t)) < K\ 20

n i = 1,2
and
\Pi(t) - Si(t)\ < 0 foralU.
We then choose again the shortest lines from Si (t) to 7 and replace v (t) n
by the straight line v (t) between those points, where these two shortest
n
lines meet 7. By our geometric argument above

s
L (v (t)) < K\ rj for some rj > 0,
n
whenever
L(v n (t)) > KI - e. 0
Since also always

L K (t)) < L (v n (t)),
we may then construct a family i V\ w i t h
sup L (t)) <K - X min(77, (3)

tei
contradicting the definition of K \ . Consequently, our claim is correct. We
then find a sequence (t ) C / and (v )
n neN C V\ w i t h n neN
supL(v n (t)) < K \ + -

tei n
L(v n (t ))
n > K\ - -
n
|cos ( a i (v (t )))\, n n |cos ( a (v (t )))\

2 n n <
n
A subsequence of (v (t )) ^
n then converges to a straight line l\ in A
n ne
of length K\ meeting 7 orthogonally at its endpoints.
In order to construct a second line l meeting 7 orthogonally at its 2
endpoints, we proceed as follows:

We denote by V the class of all continuous maps
2
v : J x J -+ C
with
v({0} x / ) and v({l} x I) C C 0 (3.2.1)
and w i t h the following property:

For all continuous maps
T(8) = (h(8)MB))
f -o
2 t2 = 1/4 t 2 = 1/2 f = 3/4
2 r -l
2
Figure 3.6.
with
t i ( l ) = 1 - t i ( 0 ) , t ( 0 ) = 0 , t ( l ) = 1,
2 2
(3.2.2)
we have
Let us exhibit an example of such a v 6 V (see Figure 3.6). We consider 2
the v\ G V\ of Figure 3.5 where i>i(0) and V were point curves on
7, and we rotate v\ via the parameter t so that at 2 = 1 we have 2
the same picture as at t = 0, but w i t h t\ interchanged w i t h 1 t i .

2
Equation (3.2.2) then holds.

We note that I x I becomes a Mobius strip, when we identify the
parameter t\ on the line t = 1 w i t h the parameter 1 t\ on the line
2
t =0.
2
We define
K2 := inf sup L (v (t)).
vV [22te
Then
K2 > Ki
and K again is realized by some straight line l in A meeting 7 orthog

2 2
onally at its endpoints. We consider two cases:
(1) K > K\.2
Then L(l ) = K > K \ = L ( i i ) , and l hence is different from l\.

2 2 2
(2) K = tti.
2
We claim that in this case, we even get infinitely many solutions

of our problem, i.e. lines in A meeting 7 orthogonally. Namely,
we let VQ G V be any critical family, i.e. satisfying
2
supL(i>o (t)) = ^ 2 -
ten
(It is not hard to see that in the present case such a VQ G V 2
indeed exists.)
2
We then have for any r : J -+ I w i t h (3.2.2)
sup L (v (T (s))) < K .

0 2 (3.2.3)
sei
On the other hand, since VQOT eVi,
KX < sup L ( v o ( r (*))), (3.2.4)

sei
and since K \ = K , we have equality in (3.2.3) and (3.2.4). This means

2
that VOOT is a critical family for i , and i t then has to contain a solution
l of our problem.
r
Let S C {(s,t) e I x 11 L(i;o(s,)) = K ) } denote the set in J x J 2
corresponding to all solutions induced by vo. After carrying out the

identification prescribed by (3.2.2), which makes I x I into a Mobius
strip, we see that the complement of S in this Mobius strip is not path
connected. Namely, otherwise we could find r satisfying (3.2.2) for which
T ( J ) avoids 5, and for such a r , vo o r would then not contain a solution,
as S is the set of all solutions in the family VQ. This, however, contra
dicts what has just been said (see Figure 3.7). I n fact, S has to carry
a one dimensional cyclef on the Mobius strip. Otherwise, S would be
contractible (in the Mobius strip) and one could reparameterize VQ on
2
I so that the set of solutions corresponds to a finite number of points.
But this is incompatible w i t h K = K \ as we have just seen. Since for
2
each path r as in (3.2.2) w i t h r ( J ) C S, vo o r G V\ is nonconstant by

(3.2.1) and (3.2.2), we obtain an uncountable number of solutions.
3
We thus have shown our result if 7 is of class C . I f 7 is only of class
1 3
C , we choose a sequence of curves y of class C approximating 7 . This n
means that there are parameterizations 7 U ( T ) , 7 ( T ) by arc-length w i t h
lim s u p ( | 7 ( r ) - 7 ( r ) | +
n = 0.
noo r \
We then let li, and Z ,n be the corresponding solutions for j . After

n 2 n
selection of subsequences, l\ and l then converge to solutions Zi, l for

in 2yU 2
7 , and those l\ and l realize the critical values K \ and K , respectively.

2 2
Since the argument to produce infinitely many solutions in case K \ K 2
f We have to employ here some constructions from algebraic topology. A reference is

any good book on that subject, e.g. M.Greenberg, Lectures on Algebraic Topology,
Benjamin, Reading, Mass., 1967, pp. 33-45, 186. While this is somewhat technical
we strongly urge the reader to try to understand the essential geometric idea of
the preceding construction.
^
ti
Figure 3.7.
did not depend on a higher differentiability assumption on 7, i t is still

applicable here, and we thus can complete the proof as before.
q.e.d.
The variational content of Theorem 3.2.1 is that we produce two

2
geodesies in E that meet a given convex Jordan curve orthogonally
In fact, this statement generalizes to any closed convex Jordan curve on
some surface, enclosing a domain homeomorphic to the unit disk.
In Sections 2.3, 3.2, we could only treat variational problems that
could be reduced to finite dimensional problems, because we did not
yet develop tools to show the existence of critical points of functionals
defined on infinite dimensional spaces. We shall develop such tools in
Part I I , and consequently in Chaper 9 of Part I I , we shall be able to
present general results about the existence of unstable critical points in
the spirit of the preceding results. The crucial notion will be the Palais-
Smale condition that guarantees that the type of reasoning presented in
Section 3.1 extends to certain functionals defined on infinite dimensional
spaces. Also, the reasoning employed in Section 3.2 that infinitely many
critical points can be found i f two suitable critical values coincide will
be given an axiomatic treatment in Section 9.3 of Part I I .
Exercises
3.1 Let F C ^ M j R ) ( M an embedded, connected, differentiable
d
submanifold of R ) be bounded from below and proper (i.e. for
all 5 R, {x M : F(x) < s} is compact), and suppose F has
two relative minima X i , x . Let2
KO : = m a x ( F ( # i ) , Ffa))-
Show that F either possesses a critical point 3 w i t h F(x^) > ô,

or that i t has uncountably many critical points.
x d
3.2 Let F C ( R , R ) be bounded from below and proper, and
suppose i t has three strict relative minima i , # 2 , 3 . Try to
identify conditions under which F then has to possess more
than two additional critical points, e.g. three or four.
2 3
3.3 Let A be a compact convex subset of the unit sphere S C R ,
and suppose OA is a smooth curve 7; the convexity condition
here means that for any two points in A, one can find precisely
one geodesic arc inside A that connects them. Show the existence
of at least two geodesic arcs in A that meet 7 orthogonally at
both endpoints.
4
The theory of Hamilton and Jacobi
4.1 T h e canonical equations

We let t be a real parameter varying between t\ and t . We consider the 2
variational integral
2 1 n l n
I = f L fax ^),... ,x (t),x (t),... ,x (t)) dt (4.1.1)
Jti
1 n
for the unknown functions x(t) = (x ^),..., x (t)) w i t h fixed endpoints
x(t\) and x(t ). Here,
2
.i _ <W_
X :
~ dt'
2
We assume that L is of class C . The Euler-Lagrange equations for /
are
^-L -L =0
i xi (i = l , . . . , n ) . (4.1.2)
We assume the invertibility condition
detL ij ^0. (4.1.3)
2
As shown in 1.2, this implies that solutions of (4.1.2) are of class C .
(4.1.3) also implies that we may perform a Legendre transformation.
Namely, by the implicit function theorem, we may then locally solve
Pi = L i (4.1.4)
l
w.r.t. x , i.e.
i
x =x*(t,x,p) ( p = (pi,...,Pn)). (4.1.5)
79
80 The theory of Hamilton and Jacobi
The expressions pi are called momenta. The Hamiltonian H is defined

as
%
H(t,x,p) := x pi - L ( , x , x ) . (4.1.6)
We obtain
H x i p j L i J L x i
- M ~ d x ^ ~ >
and with (4.1.4) then
H% x = ~L i. x
and with (4.1.2) and (4.1.4) then
Hi x = -p.. (4.1.7)
Also
n
Pi Pj r X L/ jX ,
dpi dpi
and thus again with (4.1.4)
H =x\ Pi (4.1.8)
(4.1.7) and (4.1.8) constitute a so-called canonical system. We are going

to see that (4.1.7) and (4.1.8) also arise as Euler-Lagrange equations of
the variational problem obtained by expressing L in (4.1.1) through H
via (4.1.6). Namely,
2 j
I = j [x Pj -H(t,x,p)) dt, (4.1.9)
Jti
where the unknown functions are x(t) and p(t), has Euler-Lagrange
equations (4.1.7) and (4.1.8), and so does
2 j
J = - [ (x pj+H(t,x,p))dt. (4.1.10)
Before proceeding, we observe that i f H does not depend explicitely on

t, i.e. H = H(x,p), then i f is a constant of motion, i.e. constant along
any solution x(t) of the equations, Namely,
l
~H (x(t),p(t)) = H ix x + H pi Pi = 0 (4.1.11)
by (4.1.7) and (4.1.8).

4-2 The Hamilton-Jacobi equation 81
Example. For L = | |x| V(x), we have
1 2
H= -\p\ + V(x),
and the canonical equations become
x =p
P=-V . x
This example, which describes the Newtonian motion of a particle of

unit mass subject to a potential F , is helpful for remembering the signs
in the canonical equations.
4.2 T h e H a m i l t o n - J a c o b i equation
n + 1 1 n
Assumption. There is given a set fi C M = { ( ^ x , . . . , ) } with
1 71
the property that for any points A, B fi, A = (a, ft , . . . , ft ), B =
1 n n
(s, q ,..., q ), there is a unique solution x(t) = ..., x (t)) of
(4.1.2) contained i n fi w i t h (a, x(o~)) = A, (s,x(s)) = B.
Thus, fi is covered by solutions of (4.1.2), and those can be considered
as functions of their endpoints. Thus
n n
= (t;s,q\...,q ;a^\...^ ) (4.2.1)
and also
n l n
Pi = gi(t;s,q\...,q ;o-,K, ,...,K, ) = L i. (4.2.2)
In particular,
1 1 71
= (a; s , ? , . . . , ^ ; a, ft ,..., ft ) (4.2.3)
i 1 n 1 n
Q* = f (s\s,q ,...,q ;a,K, ,...,K, ).
We also define
l n 1 n
ipi := gi(a; 5 , q ,..., q ; a, ft ,..., ft ) = L i (a, k ft, ft) (4.2.4)
x n 1 n
Vi := gi(s; s,q ,... ,q ;a, ft ,..., ft ) = L^(s,q,q).
In the sequel, / * etc. will mean a derivative w.r.t. the first independent
l
variable, f etc. a derivative w.r.t. the second one. Inserting (4.2.1),
(4.2.2) into i " , we obtain
J = I(s,q,a,K) (4.2.5)
and call this expression the geodesic distance betweeen A and B. I n this
connection, I is called eiconal. Recalling (4.1.9), we may write
I = j ( < -H(t,x,p))dt.
p< (4.2.6)
We want to compute the derivatives of I(s, q, a, K).
{
I s = q
Vi - H(s, q, v) + j f {g'J + f' 9i - H f' Xi - H gty
Pi dt
Equations (4.1.7) and (4.1.8) yield H i = fa, H x Pi = /*, and thus
Is = - H{s, q,v) + J (oif) dt

l I ts
1
= ^<f - H(s,tf, v) + gif
\t = (T
Equation (4.2.3) yields
p + f = 0 for t = s
f = 0 for t = a,
and thus altogether
l
h = Vi<j - H(s,q,v) - Vitf
= -H{s,q, v)
i
= L(s,q,q)-q L^. (4.2.7)
Next
q3 f + ]
~l [W *W~ *W
t=s
9i^\ again by (4.1.7), (4.1.8)
1 n
^(s; 5, g , . . . , q ; 1
G , K , . . . , K) N
by (4.2.3)
and : = 0, ~P~~~t = 6{j. 13

dqi dq3
Thus
I q j = Vj =L#(s,q,q). (4.2.8)
4-2 The Hamilton-Jacobi equation
Analogously,
I a = H(a, ft, ip) = L(o, ft, ft) 4- l

k L x.k (4.2.9)
4* = ->j = -L (o-,K,,k).
kJ (4.2.10)
Inserting (4.2.8) into(4.2.7), we obtain
h + H( (4.2.11)
Thus, the geodesic distance as function of the endpoint satisfies (4.2.11),

a Hamilton-Jacobi equation. I n the present context that equation then
is also called eiconal equation. We observed at the end of Section 4.1
that H is constant along solutions i f it does not depend on t explicitly.
In that case, (4.2.11) implies that / then depends linearly on 5. I t may
be useful for understanding the preceding formulae if we derive them
without the use of the Legendre transformation. Thus
1= f L(t,x(t),x(t))dt= f L(t,fJ)dt
J a J o
and
The Euler-Lagrange equations give
and so
= L(s,q,q)+ L if

As before, we obtain from (4.2.3)
p =-f fort = 5
/*' = 0 for t = a,
hence
I s = L(s,q,q) - L#q\
i.e. (4.2.7). Likewise,
3 x 3
dq dq
I t= s
dp
rT"
L +i
i.e. (4.2.8). Thus, the Hamilton-Jacobi equation (4.2.11) is

i
I -L(s,q,q)
s + Iq qi =0. (4.2.12)
We have seen in the preceding how solutions of the canonical equa

tions yield solutions of the Hamilton-Jacobi equations. We now want to
establish a converse result.
1 n
Let (p(t, x , . . . , x ) be a solution of the Hamilton-Jacobi equation
which we now write as
n
po + H(t, x\ . . . , x , P l , . . . ,p ) = 0 n (4.2.13)
with
Po = >t
Pi = <Px*-
D e f i n i t i o n 4 . 2 . 1 . If
1 n 2
<p = G(t, x , . . . , x , A i , . . . , A ) n with G e C (4.2.14)
and
det(G x % ). . = 1 n ^ 0 (4.2.15)
is a family of solutions of (4-2.13) depending on n parameters A i , . . . , X , n
we call
n
<p = G(t, x\ . . . , x , A i , . . . , A ) + A n (4.2.16)
(where A is a free real parameter) a complete integral of (4-2.13).
We have the following theorem of Jacobi:

1 n
T h e o r e m 4 . 2 . 1 . Let <p = G(t, x , . . . , x , A i , . . . , A ) - f A be a complete n
integral of (4.2.13). Then one may obtain a family of solutions of the

4-2 The Hamilton-Jacobi equation 85
canonical equations
{
H =x Pi (4.2.17)
H. x = -Pi (4.2.18)
1 n
depending on 2n parameters A i , . . . , A , / i , . . . , / i n by solving
G Xi = V? (4.2.19)
Gi =x P i . (4.2.20)
Proof. Because of (4.2.15), (4.2.19) may be solved w.r.t. x*,

l 1 7 1
x = *(, A i , . . . , A n , / / , . . . , / / ) .
Inserting this into (4.2.20) then yields

1
Pi =Pi(t, Aij-.-jAn,// ,...,//").
%
We have to show that x and i satisfy the canonical equations. For this
P
l
purpose, we differentiate (4.2.13) w.r.t. x and obtain:
Gtxi + H G k Pk x xi + H i = 0. x (4.2.21)
Differentiating (4.2.13) w.r.t. A^, we obtain
G +H G =0,
tXi Pk xkXi (4.2.22)
since the terms containing ^ cancel by (4.2.21). Differentiating (4.2.19)

w.r.t. t, we obtain
k
dx
GXit + G =0. Xixk (4.2.23)
Comparing (4.2.22) and (4.2.23) and recalling (4.2.15) yields (4.2.17).

Differentiating (4.2.20) w.r.t. t, we obtain
^ = G xH + G x i x ^ . (4.2.24)
Comparing (4.2.24) and (4.2.21) and using the relation (4.2.17) just
derived, we then obtain (4.2.18).
q.e.d.
The canonical equations are a system of ODE whereas the H a m i l t o n -

s t
Jacobi equation is a 1 order partial differential equation (PDE). The
preceding considerations show the equivalence of these equations. While
in general, one may consider a P D E as being more difficult than a system
of ODE, in applications, one may often find a solution of the canonical
equations by solving the Hamilton-Jacobi equation. Here, it is typically

of great help that the Hamilton-Jacobi equation does not depend on the
unknown function itself, but only on its derivatives.
Let us consider the following example of geometric optics:
2
/ = / </?(, x) y/l + x dt (</?(, x) > 0 ) ,
Jti
already explained in Example (3) of Section 1.1 in a slightly different
notation. The physical meaning is that x(t) is considered as the graph
of a light ray travelling in a medium with light velocity ^ ^ y , where c
is the velocity of light in vacuum. I n this example, putting
2
L(t, x, x) = <p(t, x) \ / l + x , (4.2.25)
we have
p= L = ipx
x
2
VTTx
2
H = px-L = - vV-P - (4.2.26)
7(s, g, cr, K) here is the time that a light ray needs to travel from A =
(a, K) to B = (5, q). The Hamilton-Jacobi equation I + H (5, g, I ) = 0 s q
becomes the eiconal equation
2 2 2
I +I =<p .
s q (4.2.27)
The surfaces 7(5, q) = constant are called wave fronts.

Another simple example comes from a quadratic
2 2
L(t, x, x) = i ( x + a x ) (a = constant). (4.2.28)
Then
2 2
p= L x = x,H = p x - L = \(p - ax ), (4.2.29)
and the Hamilton-Jacobi equation becomes
2 2
/ + i(/
( x - a x ) = 0. (4.2.30)
2
If we substitute I = p(t)x , we are led to the Riccati equation
2
p + 2p -^=0. (4.2.31)
4.3 Geodesies 87
If we substitute I = Xt + ip(x) w i t h a parameter A, we obtain from

(4.2.30)
1
- A + ^ (' lJ<\2
( x ) - a x )2\ = 0 ,
n
2 2
i.e.
2
i/>'(x) = \ A * x + 2A
and a solution
x
f
2
I = -\t+ v /a + 2Ad. (4.2.32)
Jo
The equation
I\(t, x, A) = \i
means
2 1
7o Vx /<a f + 2A
This can be solved for x; let us assume for example a < 0; then the
solution is
x = y sin (v'a ( 4- /^)).
x of course solves the Euler-Lagrange equation for (4.2.28)
x = ax.
A physical realization is the harmonic oscillator, where x(t) is the dis

placement of an oscillating spring, with a ~ (rn mass, k = spring
constant). Since
p= I x , I + H(x,I )
t x = 0,
we obtain from (4.2.32)
A= H(x,p),
i.e. A is the energy of the spring.
4.3 Geodesies
We consider the case where L is homogeneous of degree 1, i.e.
L = L&x\ (4.3.1)
Then
det L i x xj =0, (4.3.2)
and we cannot perform a Legendre transformation as in Section 4.1. We

have
H = ~ L + x % i =0, (4.3.3)
and the computations of Section 4.2 yield (writing L i instead of pi etc.) x
i
I, = L(8,q,q)-q L i=0 4 (4.3.4)
A n example are the geodesic lines considered in Chapter 2. Here,
L = s/Q
with
n i j
Q = g (x\...,x )x x .
ij (4.3.5)
The Euler-Lagrange equations are
Q Q = 0
s(^j ")-75 " ( 4 3
- ' 6 )
Since t does not occur explicitely in (4.3.5) and since I is invariant under
transformations of t, we may choose t such that
Q = 1, (4.3.7)
i.e. that solutions are parameterized by arc-length. Equation (4.3.6) then

becomes
j Qi*
t ~ = 0. (4.3.8)
Conversely, along a solution of (4.3.8), we have Q = constant, justifying

our choice of t. Namely, Q is homogeneous of degree 2 w.r.t. the variables
1
x , hence
Q iX x
l
= 2Q. (4.3.9)
Differentiating (4.3.9) w.r.t. t along a solution,
Q i + Q i i = 2 Q = 2 Q x i i + 2 Q x i
{lt ) ^ Jt ^
and (4.3.8) indeed yields
~Q = 0 along a solution.
dt
4-4 Fields of extremals 89
As already demonstrated in 2.1, (4.3.8) are the Euler-Lagrange equa

tions for
E = x t
\ja Q( ( )^(t))dt=^^ gijixityxWxi^dt. (4.3.10)
We recall (Lemma 2.1.1) that the Schwarz inequality implies
j y/Qdt <(s-a) (^j Qdt^j

w i t h equality precisely i f Q = constant, and the extremals of E are pre
cisely those extremals of I parameterized proportionally to arc-length.
In contrast to J, E is no longer invariant under transformations of t.
Therefore, for solutions of the Euler-Lagrange equations corresponding
to E, the parameterization is determined up to a constant factor. The
Hamiltonian for E is
{
H = Q ix x -Q = Q because of (4.3.9) . (4.3.11)
Moreover,
j
Pi = Q =2 x . xi gij (4.3.12)
Thus
ij 1
H = \g Pi Pj (with g** = (gij)- ). (4.3.13)
The Hamilton-Jacobi equation becomes
ij
Et + \g E <E x XJ = 0 cf. (4.3.13), (4.2.11), (4.3.10) (4.3.14)
and the canonical equations are
ij
x* = ^g Pj cf. (4.1.8), (4.3.13) (4.3.15)
ki
1 da
cf 4 1 7 4 3 N 4 3 5
pi = - j - ^ r P f e P i -1 - - )' ( - - )> ( - - )-
As observed at the end of Section 4.2, E depends linearly on t.
4.4 F i e l d s of extremals
n + 1
Let Q C M satisfy the assumptions of 4.2, T C ^ f y R ) . The equa
tion
T(a,K,\...,K, ) N
= 0 (4.4.1)
then defines a possibly degenerate hypersurface E (assume E ^ 0). Given

1 n 1 N
B = (s, g , . . . , q ) G fi, we seek A = (a, K , . . . , K ) G E that minimizes
n
I(s,q\...,q ,a, K \ . . . , K U
)
1 n
as a function of (a, K , . . . , tt ) satisfying (4.4.1). A t such a minimizing
A, we have with some Lagrange multiplier A
I a + \T a = 0 (4.4.2)
4;+AT^=0 (j = l , . . . , n ) .
Unless the situation is degenerate (A = 0 or T = T t = 0 for all i), this a K
means that the vector (I , I i,..., aI n) is proportional to the gradient

K K
of T, hence orthogonal to E. From (4.2.9), (4.2.10), we then obtain
-H(cr, K, <p) = L(a, K, k) - kiL . k = XT a (4.4.3)
1 n
These are equations for the tangent vector ( K , . . . , k ) of the solution
from A to B. A solution satisfying (4.4.3) is called orthogonal to E. We
want to use the following:
A s s u m p t i o n . Through each point of fi, there is precisely one solution

orthogonal to E.
1 n
For each B (s, g , . . . , q ), we thus find a unique
A = (a (5, q), K (5, q)) G E minimizing 7(s, g, a, K). We call
J ( 5 , g) : = 7 (5, g, a (5, g ) , K ( S , g))
the geodesic distance from the hypersurface E.
T h e o r e m 4 . 4 . 1 . Given such a field of solutions orthogonal to E, the

geodesic distance satisfies
J9 = -H(s,q,L ) 4 (4.4.4)
and
JJ Q =Ly, (4.4.5)
hence also the eiconal equation
J +H(s,q,J )=0.
s q (4.4.6)
44 Fields of extremals 91
Proof.
J =Is
s + I<rO-s + I iK
K s (4.4.7)
7 - T T T D
^
T(a(s, q), ft(s, q)) = 0 implies
aT s a + fc*T^ = 0
and likewise
If we then use (4.4.2), we obtain in (4.4.7)
Js Is
=
Jqi Iqi
and the result follows from (4.2.7), (4.2.8), (4.2.11).

q.e.d.
Conversely
2
T h e o r e m 4.4.2. If J(s,q) is a solution of (4-4-6) of class C , there
exists a field of solutions orthogonal to the hypersurfaces J(s, q) = con
stant, and J is the geodesic distance from the hypersurface J = 0.
Proof. Let J satisfy (4.4.6). We put
Vi'.= JAs,q). (4.4.8)
The following system of ODE

l
q = H (s,qi,J ) Pi qJ (4.4.9)
then defines an n-parameter family of curves. By (4.4.8), we have along

any such curve
3
Pi = Jqi + JqiqJQ , S
and (4.4.6) gives

{ i 3 % =
Jq
S + Hq + HpjJq q -
Recalling (4.4.9), we obtain
P i = ~H . qi (4.4.10)
Equations (4.4.9) and (4.4.10) state that the curves q(s) constitute a
field of solutions. (4.4.6) and (4.4.8) yield
-H = J S
Pj = Jqi
This means that (4.4.3) is satisfied for T = J with A = 1, and the

solutions are orthogonal to the hypersurfaces J = constant.
q.e.d.
Theorem 4.4.1 gives solutions of the Hamilton-Jacobi equation (4.4.6)

1 n + 1
depending on an arbitrarily given function T G C ( M ) (namely, we
obtain those solutions that start on T = 0), whereas Theorem 4.4.2
implies that all solutions are obtained in that way. The surfaces J =
constant are called parallel surfaces of the field. I n the special case where
the hypersurface T 0 degenerates into a point, we recover the consid
erations of Section 4.2.
4.5 Hilbert's invariant integral and Jacobi's theorem

1 n
For a solution J(t, x , . . . , x ) of the Hamilton-Jacobi equation, we put
again
: = 1 n
Pi Jx^ti x , > x ).
1 n 1 n
If A = (a, K , . . . , K ) and B = (s, q ,..., q ) are connected by an arbi
trary differentiable path x*(r), the integral
J(B)-J(A) = Jj J(T,x(T))dT
t
dx
r * r \ j
-h Ji
x +J T J dr
does not depend on this particular path, but only on the end points A
and B. We rewrite this integral as
i
' dx \
Pi -^-H(r,x(T),p(T)))dT (4.5.1)
and call it Hilbert's invariant integral. Conversely now let functions

1 n n + 1
Pi(r, x , . . . , x ) be given in a region ft C E for which the integral
(4.5.1) does not depend on the path x ( r ) connecting A = (a, x (a)) and
4-5 Hilbert's invariant integral and Jacobi's theorem 93
B = (s,x (s)). Thus, we may define J : fi > R by
J(B) - J{A) = ( p I ^ - - H (T, x ( r ) ,p ( T ) ) ) dr. (4.5.2)
Since this integral does not depend on the path connecting A and B, we
must have
J i =
x P i (4.5.3)
J =
t -H(t,x,p).
J then solves the Hamilton-Jacobi equation. By Theorem 4.4.2, any so
lution of the Hamilton-Jacobi equation is the geodesic distance function
for a field of solutions of the canonical equations. Thus, any invariant
integral of the form (4.5.1) yields a field of solutions.
Let us now reconsider Jacobi's Theorem 4.2.1. Let
1 n
? = G(t,x ,...,x ,A ,...,A ) + A
v 1 n (4.5.4)
be a complete integral of
n
p + H(t, x \ ..., x , P l , . . . ,p ) = 0 n (4.5.5)
(with p = (pt, i P (p i);

x in particular
det(G < ,)^0. x A (4.5.6)
Jacobi's theorem says that we obtain a 2n-parameter family of solutions

of the canonical equations by solving
Gi
x = Pi,
n
where the parameters are A i , . . . , A , fi\ . . . , /Li . For fixed values of A i , . . . ,
n
A , A, G determines a field of solutions of the canonical equations, and

n
by the preceding consideration, i t is given by the corresponding invariant

integral
G(B) - G(A) = (G . ^-H^Jdr

x (4.5.7)
= | l (r, ^ ( r ) , *(r)) + - ( r ) ) L J dr,
l
where x (r) now denotes the derivative in the direction of the solution
l
and not in the direction of the arbitrary curve x (r) connecting A and B.
We now vary A i , . . . , A , but keep the curve x*(r) fixed. Then the field
n
of solutions varies, and so then does X (T). We also determine A so that1
G(A) = 0. Differentiating (4.5.7) then yields
a < 4 5 8 )
* - -
In the same way as G(B), this expression only depends on B (A is kept
3
fixed for the moment) but not on the particular x (r). For each J3, we
find Bo on the surface
1 n
G(t,x ,...,x ,A ,...,A ) = 0 1 n
that can be connected with B by a solution of the canonical equations.

Along such a solution, we have
3
dx ,
and the integrand in (4.5.8) thus vanishes along this curve. Instead of
integrating from A to B, i t therefore suffices to integrate from A to JBo,
and we obtain
GXi = ii\ (4.5.9)
% %
w i t h \i being the value of the integral from A to Bo- Thus, \i can be
considered as a constant for the solution passing through BQ.
l J
If, conversely, (4.5.9) defined a family of curves x (t, \j,fjt ) (the family
is locally unique because of (4.5.6)), then, since G\ is constant, the 3
integrand in (4.5.8) has to vanish along any curve of the family Thus
3
/dx \
l -x>)L , =0 x Xi (t = l , . . . , n ) . (4.5.10)
In our field we have (cf. (4.2.8))
hence by assumption (4.5.6)
detL .XJX = detG XJXl + 0.

Equation (4.5.10) then implies
3
dx .,
this means that the curves defined by (4.5.9) are solutions of the canoni
1 n
cal equations contained in the field defined by G(t, x , . . . , x , A i , . . . , A ) . n
We also observe that the parameter A is only used for specifying the sur
face G = 0 and has no geometric meaning beside that.
4-6 Canonical transformations 95
4.6 C a n o n i c a l transformations
We want to find transformations, i.e. diffeomorphismsf
2 n 2 n
ip : R -> R
H-> (,7r),
that preserve the canonical equations
x = H p
(4.6.1)
This means that = ( x , p ) , TT = 7r(x,p) satisfy
(4.6.2)
with #*(f^(x,p),7r(x,p)) = H(t,x,p).

Equation (4.6.1) constitutes a system of ODE and i f the assump
tions of the Picard-Lindelof theorem are satisfied, a solution exists for
given initial values x(to) = XQ, p(to) = po on some interval [ t o , t i ] .
w
For any i E [ ^ o ^ i ] e then obtain such a transformation by letting
5
(x,p) = x ( f ) , 7r(x,p) = p(i) where (x(t),p(t)) is the solution of (4.6.1)

w i t h x(t ) = x,p(to) = p. Thus, the evolution of (4.6.1) in time t, the
0
so-called Hamiltonian flow, yields 'canonical t r ansfor mat ions'. However,

the concept of canonical transformations is more general as we now shall
see.
Since
b y ( 4 6 1 }
- -
f A diffeomorphism is a bijective map that together with its inverse is everywhere

differentiable.
and
i
dx dpi
H r + HPi
d7T4 On,
i
dx dpi
we obtain the conditions
dpi dp
dnj ~
i
dx dp
drci dpi
dpi dwj
dp dx*
i
dx diTj
(4.6.3)
dp dpi'
or in matrix notation
07T
dp (r -() (4.6.4)
T
l-(ff)
T
where A denotes the transpose of a matrix A. Obviously, this is a con
dition that does not depend anymore on the particular Hamiltonian H.
Definition 4.6.1. A diffeomorphism ip : R R , (x,p) H-> (,7r),

2 n 2 n
satisfying (4-6.3) (or equivalently (4*6.4)) is called canonical transfor

f
mation.
Canonical transformations can often be used to simplify the canonical

equations. Before we return to that topic, however, we interrupt the
discussion of the Hamilton-Jacobi theory in order to describe some basic
points of symplectic geometry (for more information on that subject, we
refer to D.Mc Duff, D.Salamon, Introduction to Symplectic Topology,
Oxford University Press, Oxford, 1995). We denote the (n x n) unit
matrix by I and put
n
Then obviously
J2
= ~hn. (4.6.5)
Equation (4.6.4) may then be written as
1 T
(DI/J)~ = -J(Dil>) J, (4.6.6)
or equivalently
T
(Di)) JD<4> = J. (4.6.7)
In this connection, a satisfying (4.6.7), i.e. a canonical transformation,

is also called symplectomorphism. Prom these relations, one also easily
sees that ^ is a canonical transformation iff is.
In terms of J , the canonical equations (4.6.1) can also be written as
z = -JVH(t,z) (4.6.8)
where z = (x,p), VH(t,z) = (H ,H ). X P
For a reader who knows the calculus of exterior differential forms, the
following explanation should be useful. We consider the two-form
1 2 n
u = dx A dpi on E
1
(here, as always, we use a summation convention: dx A dpi means
l
Y17-1 dx Adpi). According to the transformation rules for exterior differ
3 1
ential forms (i.e. d^ = J^dx etc.), we have, for = ( # , p ) , n = 7r(x,p),
J x
d A dnj = y ^ i g ^ ~ ~g~[ j d> A dpk-
Thus, UJ remains invariant under the transformation ip, i.e.
dtf A diTj = dx* A dpi (4.6.9)
precisely i f ^ is a canonical transformation. I n fact, this is often used as

the definition of a canonical transformation. I f UJ is left invariant under
so is
n n z l n
uj : = ( J A - A ( J = n\(l) ^ ^'dx A' - A d x A d p i A - -Adp . (4.6.10)
n
n times
Since
x n l n
d A A d AdTTi A Ad7T = (det Dip)dx n A Adx Adp A A d p ,
x n
we conclude Liouville's:
2 n 2 n
T h e o r e m 4 . 6 . 1 . Every canonical transformation ip : R > R satis
fies
det Dip-zl. (4.6.11)
q.e.d.
One also expresses this result by saying that a canonical transforma
l n
tion is volume preserving in phase space as dx A A dx A dpi A A dp n
2n
can be interpreted as the volume form of R . By what was observed in
the beginning of this section, this applies in particular to the Hamilto-
nian flow which constitutes Liouville's original statement.
After this excursion and interruption, we return to our canonical equa
tions (4.6.1) and t r y to simplify them by suitable canonical transfor
mations. Canonical transformations may be easily obtained from the
variational integral
I I L(t,x,x)dt
with
L(t, x, x) x p - H(t, x,p) (p = L ).

x
If W is any differentiable function, then
dW\
* = J* (^L(t x,x)dt +J ~)dt
)
has the same critical points as / , because
I*=I + W(t )-W(ti), 2

so that / * and / differ only by a constant independent of the particular
path x(t). Thus, we may for example take any function W ( , x , ) and
require that for all choices of x, , x,
dW
x-p- H(t,x,p) = 7T - H , T T ) + . (4.6.12)
Then, with
differs from 7 only by a constant. Thus, i f x(t) is a critical path for 7,

(x(t),p(t)) then becomes a critical path for I * . Since

dW
= W t + W .x x + W Z,
r
(4.6.12) becomes
x-(p-W )--(n
x + Wz)-H + H*-W =0 t (4.6.13)
Since (4.6.13) is required to hold for all choices of x, , x, , we obtain:
T h e o r e m 4.6.2. Given an arbitrary (differentiable) function W(t, x, ),

a canonical transformation (transforming (4-6.1) into (4-6.2)) is ob
tained through the equations
P^W X
7T = -\\\ (4.6.14)
H* H , i.e. W =0.
t
Wt = 0 of course means that W = W ( x , ).
I n the same manner, we may also take a function W(t,p, ), W(t, x, 7r)
or W ( t , p , 7r). I n the first case, we obtain for example the equations
x = W p
H* H , i.e. W = 0. t
Here and above, of course H* = H*(t,,ir).

We may now easily explain Jacobi's method for solving the canonical
equations. We t r y to find W(x,) satisfying
H(t,x,W (x,Q)
x = H*(Q, (4.6.15)
i.e. reduce the Hamiltonian to a function of the variable alone. We

have to require that
detW t x v 7^0. (4.6.16)
This ensures that the equation 7r = W^ determines x, and p then is

determined from p = W . I f (4.6.15) holds, (4.6.2) becomes
x
i =o
7r = - i f | . (4.6.17)
This implies that are constants of motion (i.e. independent

of t), or so-called integrals of the Hamiltonian flow. A system for which
n independent integrals can be found is called completely integrable.

Thus, if we can find a so-called generating function W(x, ) of the above
type reducing the Hamiltonian to a function of alone, the canonical
x n
system is completely integrable. Clearly, since in this case , . . . , are
constant in t, the relation 7r = if|() can then be used to determine
7 T i , . . . , 7r . I n other words, a completely integrable canonical system may
n
be solved explicitly through quadratures. Actually, one may show in this

x 1 n n
case that the sets T = { = c , . . . , = c } for a constant vector
c
1 n
c = ( c , . . . , c ) are n-dimensional tori, if compact and connected. Thus,
2n
the so-called phase space { ( x , p ) G R } is foliated by tori that are
invariant under the motion, and on each such torus, the motion is given
by straight lines.
It should be pointed out, however, that completely integrable dynam
ical systems are quite rare, in the sense that the complete integrabil-
ity usually depends on particular symmetries, and their dynamical be
haviour is quite exceptional in the class of all Hamiltonian systems.
The invariant tori may disappear under arbitrarily small perturbations.
By way of contrast, the Kolmogorov-Arnold-Moser theory asserts that
these invariant tori persist under sufficiently small and smooth pertur
bations if the coordinates of H are rationally independent and satisfy
certain Diophantine inequalities, and if the matrix H^ of second deriva
tives is invertible.
In the older literature, the notion of 'canonical transformation' is usu
2n 2n
ally applied to any transformation ip : R > R that preserves the
form of the canonical equations, i.e. (4.6.1) is transformed into (4.6.2),
but without requiring that
H*(t,,ir)=H(t,x,p).
A n example of a canonical transformation in this wider sense is
= 2x , TV = p
with H * = IE.
If we now take a generating function W(t, x, ) as above, the Hamiltonian

is transformed into
H* = H + W t (4.6.18)
while the first two relations of (4.6.14), i.e.
p W x , 7T = W$ (4.6.19)
still hold. This may be used to explain Jacobi's theorem once more, as
we now shall see.
1 n
Let I(t,x ,..., x , A i , . . . , A ) be a solution of the Hamilton-Jacobi
n
equation
J + #(*,x,J )=0,
t x (4.6.20)
depending on parameters A i , . . . , A and satisfying as usually

n
det/ i x A j ^0. (4.6.21)
We now choose
n 1 n
W{t,x\...,x ,t ,...,S )
1 n = I(t,x ,...,x ,tu...,Z ).
n
The corresponding transformation then is
P= h
7T = -It: (4.6.22)
H*(t,t,n) = H(t,x,p) + It.
Because of (4.6.20),
H* = 0 .
Thus, the new canonical equations are just
i =o
7T = 0.
Solutions are of course
= A = constant
7r = I\ = fi = constant.
We have thus obtained the statement of Jacobi's Theorem 4.2.1, namely

that from a solution of (4.6.20) w i t h (4.6.21), we may obtain solutions
of the canonical equations by solving
h =
Ix = P
1 n
w i t h parameters A = ( A i , . . . , A ) , fi = ( / i , . . . , /L* ).
n
Classical references for this chapter include:
C.G.J. Jacobi, Vorlesungen iiber Analytische Mechanik (ed. H . Pulte),

Vieweg, Braunschweig, Wiesbaden 1996,
C. Caratheodory, Variationsrechnung und partielle Differentialgleich-
ungen erster Ordnung, Teubner, Leipzig 1935,
R. Courant, D. Hilbert, Methoden der Mathematischen Physik II,
Springer, Berlin, 2nd edition, 1968.
The global aspects are developed in
V . I . Arnold, Mathematical Methods of Classical Mechanics, GTM60,

Springer, New York, 1978.
A recent advanced monograph is
H. Hofer, E. Zehnder, Symplectic Invariants and Hamiltonian Dynam

ics, Birkhauser, Basel, 1994.
That text will give readers a good perspective on the present research
directions in the field.
Exercises
4.1 Discuss the relation between the canonical equations for the
energy functional E and the equations for geodesies derived in
Chapter 2.
4.2 (Kepler problem) Consider the Lagrangian
2 3
L(x,x) = \ \x\ + r i r for x G E .
2 |x|
Compute the corresponding Hamiltonian and write down the
canonical equations. Show that the three components of the
angular momentum x A x are integrals of the Hamiltonian flow.
2 n
4.3 For smooth functions F,G : E E, define their Poisson
bracket as
X ]
' ' dxidpj dpjdxi'
1 n
where z = (x,p) = ( x , . . . , x , p \ , . . . ,p ) are Euclidean coordi
n
2 n
nates of E . Let z(t) = (x(t),p(t)) be a solution of a canonical
system
x = H p
P = ~H X
Exercises 103
for some Hamiltonian H(x,p) that is independent of t. Show

that for any (smooth) F : E 2 n
R
j F{z{t))
t = {F,H}.
Show that the Poisson bracket is antisymmetric, i.e.
{F,G} = -{G,F}
and satisfies the Jacobi identity
{{F, G},L} + {{G, L},F} + {{L, F}, G} = 0
for all smooth F, G, L .

2 n 2 n
Show that a diffeomorphism ip : R ~> R is a canonical trans
formation if
{F,G}o^ = {Fo^,Go^}
for all smooth F, G.

5
Dynamic optimization
Optimal control theory is concerned with time dependent processes that

can be influenced or controlled via the tuning of certain parameters.
The aim is to choose these parameters in such a manner that a desired
result is achieved and the cost resulting from the intermediate states
of the process and from the application or change of the parameters
is minimized. I n some problems, the control parameter can be applied
only at discrete time steps, while other problems can be continuously
controlled. As we shall see, however, the discrete and the continuous
case can be treated by the same principles. Since the end result may be
prescribed, and the value of a parameter at some given time influences
the state of the system at subsequent times and therefore typically will
also contribute through this influence to the cost of the process at those
later times, the determination of the optimal control parameters is best
performed in a backward manner. This means in the discrete case that
one first selects the best value of the control parameter at the last stage,
whatever state the system is in at that time, then the value at the
second-to-last stage, so that at this step the contribution of the value
of the control parameter at the last stage to the total cost function is
already determined and one only needs to optimize the cost function
w.r.t. the second-to-last parameter value, and so on.
5.1 Discrete control problems

d
We consider a process with n states x i , . . . , xn M . A t each state # i ,
we may choose a control parameter
K e Ai, (5.1.1)
104
5.1 Discrete control problems 105
c
where A* is a given control restriction (A* C M ) to determine
Xi+i = (pi(xi,\i) (5.1.2)
w i t h cost
ki(Xii A$).
The total cost of the process starting at the initial state x v is

n
K (x ,\ ,...
u u u ,\ )
n := ^ f c t ( x j , A i ) , with x\ i+ = (fi(xi, A<). (5.1.3)
We wish to minimize the total cost of the process and define the Bellman
function
Iu(x ):= u inf #(, A,...,A) (i/ = 1 , . . . , n). (5.1.4)
T h e o r e m 5.1.1. The Bellman function satisfies the Bellman equation
I (x )
v v = inf (k (x , A) + I +i
v v v ((ft, (x, A))) forv=l,...,n
AA
(5.1.5)
(here, we put I +i = 0). Furthermore, ( A , . . . , A ) G A^ x x A ,
n n n
( x , . . . ,x ) with (5.1.2) are solutions of (5.1.4) iff

n
Ij(xj) = ^(XJ^XJ) + I + (x +i)

j 1 j for j = i / , . . . , n . (5.1.6)
Proof. Since
K. (x \ v v \ , . . . , A ) = k (x ,
v n v v \ ) -h i ^ - i - i ((p (x ,
v v v \v)\ A | / 4 - i , . . . , A ) ) ,
n
we get
J,/(x) = inf jFf(x;A,...,A ) n
i = t/,...,n
= inf 1 inf K (x \A,...,

v v A )I
n
= inf \k (x ,\ )+
v v v inf K +1 ( < / v ( ^ A); A + i , . . . , A ) ]
v n
A e A
AA \ i i /
= inf (k (x i\ ) v V u + I +i(<p {x i\ )))

v v V v i
AA
which is (5.1.5). For ( A , . . . , A) G A x x A , Xj+i = n <PJ(XJ,\J) for

106 Dynamic optimization
j = i / , . . . , n,
< kvixv, X ) -f- -f- k (x ,

v n n A ) = i (x ,
n v v A , . . . , A ).
n n
If the infimum w.r.t. Xj G A j (j = v,..., n) is realized, we must have

equality, and (5.1.6) follows.
q.e.d.
Corollary 5.1.1. ( A i , . . . , A) A i x x A , ( x i , . . . , x ) with (5.1.2) n n
is a solution of (5.1.4), iff for all v 1 , . . . , n, ( A , . . . , A ) G A x x n
A ( x j , , . . . , x ) with (5.1.2) is a solution of (5.1.4)-

n ? n
Corollary 5.1.2. (Bellman's method) An optimal solution of the pro

cess can be calculated as follows:
For any value of x , compute A ( x ) minimizing (5.1.5) for
n n n
v = n. Having computed XJ(XJ) for j = v + l , . . . , n , com

pute A(x) for any value of x as to minimize (5.1.5) and put
u
x +\ = Kp (x , X (x )).
v v For an arbitrary initial value x\, an
v v v
optimal process thus is given by:
Ai : = A i ( x i ) , x 2 := ^ i ( x i , A i ) , A = A (x ),... 2 2 2 .
5.2 Continuous control problems

We want to minimize
d
K(h,x(ti)) for a path x : [ t , * i ] 0 R
under the following conditions: We have the initial condition
x(t )
0 = x 0
and the final condition
x(h) G B t
d
w i t h a given set B\ G R - We have the control equation
x(t) = f(t, x(t), X(t)) for almost all t G (* , h) 0
for a piecewise continuous control function X(t) satisfying
X(t) G A
5.2 Continuous control problems 107
c
for some given A C R . Pairs ($t),x(t)) satisfying all these restrictions
are called admissible, and the set of admissible pairs is called P(to,xo).
We put
I{t ,x ):=
0 0 inf K(ti,x(ti))
(A(t),x(t))P(to,*o)
(Bellman function).
L e m m a 5.2.1.
(i) I(t\, x$ = K(ti, xi) for all xi G Bi

(ii) For any path (\(t),x(t)) G P(t ,xo), I(t,x(t)) 0 is a monotonically
increasing function oft [to, t i ] .
Proof, (i) is obvious. For (ii), if t < T\ < r < t\, the set of all admiss
0 2
ible paths from ( T 2 , X ( T 2 ) ) to (ti,Bi) can be considered as a subset of

those ones from ( T I , X ( T I ) ) to (ti,x(ti)). Namely, if we have any path
from ( T 2 , X ( T ) ) to (ti,xi)
2 for some xi G S i , we may compose i t w i t h
x(t)\
[ri r a ]to obtain a path from ( T I , X ( T I ) ) to (ti,xi). Thus, every end-
point in Bi that can be reached from ( T 2 , X ( T ) ) by an admissible path2
can also be reached from ( T I , X ( T I ) ) by an admissible path. This implies

monotonicity.
q.e.d.
T h e o r e m 5.2.1. (\(t),x(t)) is a solution of the problem, if I(t,x(t))

is constant in t. Moreover, if there exist a function J(t,x) that satisfies
J(ti,xi) = K(ti,xi) for all xi G Bi and is monotonically increasing
along any admissible path, and an admissible path (\(t),x(t)), along
which J is constant, then that path is a solution of the problem.
Proof. For a solution,
I(t ,x )0 0 = K(t x(ti))

u = I(t x{ti)
u (x = x ( t ) ) ,
0 0 (5.2.1)
I(t,x(t)) then is constant by Lemma 5.2.1 (ii). I f I(t,x(t)) is constant,

then (5.2.1) holds, and by Lemma 5.2.1, we have a solution. Given J as
described, by the monotonicity of J, for any admissible path
J(to,xo) < K(ti,x(ti)) and for the path (\(t),x(t)),
J{t ,x )
0 0 = J(h,x{ti)) = K(ti,x(ti)),
and optimality follows.

q.e.d.
Lemma 5.2.1 implies that for those t for which I(t,x(t)) is differen
tiable ((\(t),x(t)) e P(t ,x )) 0 0
I (t, x(t)) + I (t, x(t))f(t,

t x x(t), A(0) > 0.
For an optimal (A(), #()), we have by Theorem 5.2.1 then
I {t, x(t)) + I (t, x(t))f(t,

t x x(t), \(t)) = 0.
C o r o l l a r y 5 . 2 . 1 . (Bellman equation) Let t [o>*iL Assume

that for every A A, there exists an admissible pair (\(t),x(t)) with
\(r) = A, x(r) = . Then
inf: ( / t ( r , O + / ( r , O / ( r , f , A ) ) = 0 .
Proof. This follows from the proof of Lemma 5.2.1. Namely, the assump
tion implies that we may select A such that the path is optimal at the
point (r, ) under consideration.
q.e.d.
Example. We want to minimize the integral

2
[ t l
(u (t) + \ (t))dt
2
Jt0
with the initial condition

u(t )
0 = 1i 0
and the control equation
u(t) = au(t) + p\(t) with given a, (3 e E. (5.2.2)
In order to express this problem as a control problem, we introduce a

new dependent variable v(t) as solution of the equation
2 2
v(t) = (t) u + X (t) , v(t ) = 0. 0 (5.2.3)
We then want to minimize
v(h).
Given p : [to,t\] E with
p{h) = 0
and satisfying the Riccati equation
p(t) = -2ap(t)+0 p (t)-l, i 2

(5.2.4)
we put
2
J(t,u,v) = p(t)u (t) +v(t).
Then
J(h,u{t ),v(t ))=v(t )
1 1 1
and from (5.2.2), (5.2.3), (5.2.4)
2 2 2 2 2
^-J(t,u(t),v(t)) = (3 p u + 2p(5u\ + A = (/Jpu + A ) > 0,
and this expression vanishes precisely i f
X(t) = -f3p(t)u(t). (5.2.5)
By Theorem 5.2.1, x(t) = (u(t),v(t)) and X(t) = ~(3p(t)u(t) yield an

optimal solution.
If we substitute X(t) through the control equation (5.2.2) in the vari

ational integral, we obtain the integral
which is essentially the same as the one considered at the end of 4.2 with
integrand given by (4.2.28). We recall that the latter one had also been
reduced to a Riccati equation.
Equation (5.2.5) expresses the control parameter as a function of the
state of the system. We just have a feedback control: knowing the state
at a given time determines the control needed to reach an optimal state
at the next time.
5.3 T h e Pontryagin m a x i m u m principle

We consider the control problem
(x
(5.3.1)
with the control conditions
x(t )
0
(5.3.2)
x(t)=f(t,x(t),X(t))
w i t h controls
X(t) A C R C
and the end condition
g(t ,x(t ))=0.

1 1 (5.3.3)
Here, X(t) is required to be piecewise continuous, and x(t) to be continu

ous. (Equation (5.3.2) then has to be interpreted as an integral equation
x(t) = x + f* / ( r , X ( T ) , A ( r ) ) d r . ) F , / , and g are required to be of class
0 Q
1
C . Also, to is fixed, whereas t\ > to is variable subject to the restriction
(5.3.3). We define the Pontryagin function
H(x, A , p , t , / / ) :=p-
0 f(t,x,\) - ii F(t,x,\).
0
We now state the Pontryagin maximum principle
T h e o r e m 5 . 3 . 1 . If (x(t),\(t)) is a solution of the control problem,

D
there exist fi > 0, a = (OJI, . . . , ay) M. (a ^ 0 i f HQ 0) and a
0
continuous p = ( p i , . . . ,p^) on [fo, f i ] swcft f/iof of a// points where X(t)

is continuous, we have
H(x(t), A(f),p(f), *,W>) = max W ( x ( 0 , A,p(t), W>) (5.3.4)
ond
p = -H x , x = Hp (5.3.5)
and of f/ie end poinf f i , we /love the trans vers ality condition
j
da
P(h) = --:(tux(t ))-a . 1 j (5.3.6)
There also exists a continuous function rj : [ f o ^ i ] * R swc/i f/iof of o//

points where X(t) is continuous
ri(t)=H(x{t),\(t),p{t),t,iM>) (5.3.7)
and
r)(t) = H t (5.3.8)
V(h) = ^ ( t u x ^ a j . (5.3.9)
i4/so, one mo?/ always achieve HQ = 0 or 1.

Remarks:
(1) The equation x = H p is just the control equation
x = f(t,x(t),$t)).
c
(2) I f A = M , then (5.3.4) becomes
H (x(t),\(t),p(t),t,t )=0-
x M)
(3) I f we want to guarantee a fixed end time ?i, we simply introduce

an additional variable
d+1
x =t
with control conditions

d+1
x = 1
d+l
x (t )=t 0 0
and end condition

d+
x \h) = h.
We now want to exhibit the Hamilton-Jacobi theory as a special

case of optimal control theory. Concretely, we want to derive the Euler-
Lagrange equations which are equivalent to the canonical equations of
Chapter 4 from the Pontryagin maximum principle. We thus consider
the variational problem
L(t, x(t), x(t))dt min
d
with x(to) = #o, #(î) = x i , x : [to, t i ] R and where x(t) is required
to have piecewise continuous first derivatives. We introduce the control
variable through the control equation
X(t) = x(t)
d
with A = M , i.e. no constraint imposed. We have g(ti,x(ti)) = x\
x(t$. The Pontryagin function of this problem is
H(x, A,p, t, no) = p A - n L(t, 0 x, A).

d
By Theorem 5.3.1 there exists fi = 0 or 1, a G R
0 (a ^ 0 for fi = 0)
0
0 d
a n d p e C ( [ t , t i ] , R ) with
0
P = -n x
p(ti) = a
H(t,x(t),$t),p(t),ii ) 0 = max(t,x(t),A,PW,W>)
and 77 C ( [ t , t i ] , R ) w i t h
0
r?(t)=W(t,rc(t),A(t),p(t), ) W)
q(ti)=0.
We now want to exclude that /LÔ = 0. I n that case, we would have
rj = Ht=0 , hence 77 = 0 since rj(t$ = 0
and
p = Hx=0 , hence p = a since p(t\) = a.
Thus
W = a A,
and since H = 0, ( x ( t ) , A(t),p(t), t, 0) = 0, and thus

t
a = 0,
contradicting the statement of the theorem that a ^ 0 in case ^ 0 = 0.

We may thus assume == 1-
The Pontryagin maximum principle then gives the Weierstrafi condi
tion
d
L ( t , x ( t ) , A) - L(t,x(t),x(t)) > p . (A - (t)) for all A e R (5.3.10)
and
W (t,x(t),A(t),p(t),l)=0
A (5.3.11)
and the Legendre condition
L (t,x(t),x(t))
xx is positive semidefinite. (5.3.12)

P= L , x
and together with

we obtain the Euler-Lagrange equations
~M = L .
x (5.3.13)
at
A basic reference for the variational aspects of optimization and control
theory where also a detailed proof of the Pontryagin maximum principle
together with many applications is given is
E. Zeidler, Nonlinear Functional Analysis and its Applications, I I I ,
Springer, New York, 1984, pp. 93-6, 422-40.
Part two
Multiple integrals in the calculus of
variations
1
Lebesgue measure and integration theory
1.1 T h e Lebesgue measure and the Lebesgue integral

In this section, we recall the basic notions and results about the Lebesgue
measure and the Lebesgue integral that will be used in the sequel. Most
proofs are omitted as they can be readily found in standard textbooks,
e.g. J. Jost, Postmodern Analysis, Springer, Berlin, 1998, pp. 151-97 and
209-15.
d
Definition 1.1.1. A collection E of subsets ofR is called a a-algebra
d
(on R ) if
d
(i) R G E
d
(ii) / / A G E, then also R \ A G E
A
(iii) / / A G E, n = 1,2,3..., then also JJ^Li n e E.
n
The Borel a-algebra is the smallest a-algebra containing all open sub
d
sets of R . The elements of the Borel a-algebra are called Borel sets.
Easy consequences of (i)(iii) are
(iv) 0 G E
(v) I f A n G E, n = 1,2,3 . . . , then also (XLi 6 E.
(vi) ISA, Be E, then also A - B := A \ (A f l B) G E.
Definition 1.1.2. Let E be a a-algebra. A measure [i on E is a count-

ably additive function
fi: E R+ Pi {oo}.
117
118 Lebesgue measure and integration theory
1
'Countably additive here means that
oo \ oo
(n=l
U A n
)
/
=
TO=1
E
/ o r on?/ collection of mutually disjoint (AmCiA = 0 form ^ n) elements n
of 12. A measure defined on the Borel a-algebra is called a Borel measure.

A Borel measure \i is called a Radon measure if n(K) < oo for every
d
compact K cM. and fi(B) sup{fi(K) \ K C B, K compact} for every
Borel set B.
A measure / i o n E enjoys the following properties:
(vii) jx(0) = 0
(viii) I f A , B G E, A c B , then fi(A) < fi(B)
(ix) I f A G E, n = 1,2,3,... and A C A + 1 for all n , then
n n n
00 \
1J A = lim
n fi(A ).
n
/ n>oo
n=l /
d
T h e o r e m 1.1.1. There exist a (unique) a-algebra E on R and a
(unique) measure [i i n E satisfying
d
(x) i4nt/ open subset ofR is contained in E (i.e. E contains the Borel
a-algebra)
(xi) For
d d j
Q := [ x = (\x . ..,x )eR \ aj < x < bj , j = 1 , . . . , d] ,
/or numbers a\,..., a^, 6 i , . . . , bd, we have

d
b a
v(Q) = Yl( J~ A
i=i
d
(xii) (translation invariance) For x G E , A G E we /love
x + A : = { # + i/1 i/ G A} G E and + A) = n(A)
(xiii) / / A C B, B G E, //(JB) = 0, f/ien A G E (and, consequently,
This ji is called Lebesgue measure, and the elements of E are called

(Lebesgue) measurable.
In later chapters, we shall however write meas in place of fi for Lebesgue
measure.
One should note that the a-algebra of (Lebesgue) measurable sets is

larger than the Borel a-algebra.
d
We say that a property holds almost everywhere in A C E i f i t holds
on A \ B for some B C A w i t h n(B) = 0. We say that two functions
f,g : A E U { 0 0 } are equivalent if f(x) = g(x) for almost all x A.
A set contained in a set of measure 0 is called a null set.
We usually write meas A instead of n(A) for a measurable set A.
d
D e f i n i t i o n 1.1.3. Let A C R be measurable. A function
f : A E U { o c }
is called measurable if
{xeA\ f(x)<\}
is measurable for every A G E .
If / , n G N , are measurable, c G E, then / i + / , c / i , / i / , m a x ( / i , / ) ,

n 2 2 2
m i n ( / i , / ) , l i m s u p _ / , liminf _>oo f are likewise measurable. Any

2 n o c n n n
continuous function / is measurable, because in that case { / ( # ) < A}

is open in its domain of definition. We have the following important
composition property:
c c
T h e o r e m 1.1.2. Let g : A E be measurable (i.e. g = (#\..., # ),
3 c
and each component g is measurable), y : E E continuous. Then
y o g is measurable.
d
The characteristic function \A of A C E is defined as
1 if x A
otherwise.
Thus, A is measurable if and only if its characteristic function \A is
measurable.
More generally, s : A E is called a simple function or a step function
if it assumes only finitely many values, say s(A) { A i , . . . , A^}, and if
all the sets {s(x) = Xi} are measurable. Thus
k
T h e o r e m 1.1.3. / : A E is measurable if and only if it is the point-

wise limit of a sequence of simple functions. If f : A * E is measurable
and bounded, then it is the uniform limit of a sequence of simple func
tions.
D e f i n i t i o n 1.1.4.
d
(1) Let AcR be measurable with fi(A) < oo,
k
a simple function on A. The Lebesgue integral of s is

k
s(x)dx : = ] T \in({s(x) = Xi}).
i=l
(2) Let A be as in (1), f : A E measurable and bounded. Let s : n
A E 6e o sequence of simple functions converging uniformly to

f according to Theorem 1.1.3. The Lebesgue integral of f then is
I f(x)dx := l i m / s (x)dx
n
N
JA -* JA
(this integral is independent of the choice of the sequence (s ) ^).

n n
(3) A as in (1), / : i - ^ E U { 0 0 } measurable. Put
{ m
n
f(x)
if f(x) < m
iff(x)>n
ifm<f(x)<n.
We say that f is integrable if
x d:i
lim / fmA )
exists. That limit then is called the Lebesgue integral f A f(x)dx

off.
d
(4) A c l measurable, / : i - > R U { 0 0 } measurable, f is called
integrable if for any increasing sequence A\ C A2 C C A
of measurable subsets of A with fi(A ) < 00 for all n, f n is XAn
integrable on A and n
lim / f(x)xA (x)dx n
exists. That limit then is independent of the choice of (A ) and n
called the Lebesgue integral J f(x)dx of f. A
T h e o r e m 1.1.4. The Lebesgue integral is a linear nonnegative func

X
tional on C (A), the vector space of Lebesgue integrable functions on a
measurable set A, and it satisfies:
X
(1) If f e (A), and if f g almost everywhere on A, i.e.
{xeA\ f(x)^g(x)}=0,
X
then g e C (A), and
I f(x)dx = I g(x)dx.
JA JA
In particular,
f f(x)dx = Oiffi(A) = 0.
JA
l l
(2) Iffe C (A), then \f\ e C (A), and
\[ f{x)dx\< [ \f(x)\dx.
\JA I JA
1
(3) If f C ^), h: A-> R U { 0 0 } measurable with \h\ < / , then
l
h C {A) and
I h(x)dx < I f(x)dx.

JA JA
(4) If 11(A) < 00, / : A R measurable with m < f < M, then

1
feC (A), and
mfi(A) < [ f(x)dx < Mfi(A).

JA
(5) / / (-A )nN is a sequence of mutually disjoint (A n A = 0 for

n m n
m^n) measurable sets, A := U^Li A , f ^(An) for every n, n
and if
00 .
/ \f(x)\dx<oo,
j A
n=l n
l
then f e C (A), and
[ f(x)d X = f; / f(x)dx.
jA
JA n=i n
X
Conversely, if f e C (A), then this equation holds for any such
sequence (A ) ^.
n ne
l
(6) Iff G (A), then for every e > 0, there exists
d d
<p G C$(R ) := {g :R -+R continuous] {x \ g(x) ^ 0} bounded}
with
r
\f{x) - (p{x)\dx < e.
c d
T h e o r e m 1.1.5 ( F u b i n i ) . Let A C R , B C R be measurable, and
write x = (,77) G A x B. If f : A x B R U { o c } is integrable, then
j f(x)dx= j ( [ f(a,r))dri)dt
JAxB JA \JB /
ffor,)di) dn.
IB [JA
(Here, for example j B /(, n)dn exists for almost all A)
l
For / C (A), we put
We then have Jensen's inequality:

d
T h e o r e m 1.1.6. Let A C R be bounded and measurable, f a convex
l
function. Then for all ip G C (A)
1.2 Convergence theorems

In this section, again no proofs are given, and the reader is referred to
J. Jost, loc. cit., pp. 199-208.
d
T h e o r e m 1.2.1 ( B . L e v i ) . Let A C R be measurable, and let f : n
A R U { 0 0 } be a monotonically increasing sequence (i.e. f (x) < n
/ i ( x ) for all x G A, n N) of integrable functions. If

n +
lim / f (x)dx
n < 00,
n-ôoJ A
then f := l i m _ > n 00 fn (pointwise limit) is integrable, and
/ f(x)dx = lim / f (x)dx.

n
n
JA - JA
1.2 Convergence theorems 123
d
C o r o l l a r y 1.2.1. Let A C R be measurable, f n : A R+ U { 0 0 }
(nonnegative and) integrable. If
00 ~
< OC,
Y] / fn(x)dx
s
then YlnLi fn i> integrable, and
J A ^ ^ J A
d
T h e o r e m 1.2.2 ( F a t o u ) . Let A C R be measurable, f : A RU n
{ztoo} integrable for n G N . Assume that there exists some integrable

F :A R U { 0 0 } with
fn>F for all n G N ,
L
1A
f (x)dx
n < K < 00 for some K independent of n.
Then lim infnôo f n is integrable, and
/ l i m i n f f (x)dx
n < liminf / f (x)dx.
n
n
JA N
^ ^ JA
d
T h e o r e m 1.2.3 (Lebesgue). Let A C R fee measurable, f : A * n
R U { o o } o sequence of integrable functions converging pointwise almost

everywhere on A to some function f : A R U { 0 0 } . Suppose there
exists some integrable F : A R U { 0 0 } with
\fn\<F for all n.
Then f is integrable, and
/ f(x)dx = lim / f (x)dx.

n
N
JA ^ JA
Thoerem 1.2.3 is called the theorem on dominated convergence.

Let us consider an example that shows the necessity of the hypotheses
in the previous results:
f : [0,1] R is defined as
n
U(x)
t ( \
:=<
fn for 1/n < x <
1 - -
2/n 1 , ^ ^
(n>2)
10 otherwise.
Then
lim fn = 0,
and
lim / f (x)dx
n = 1^0= / f(x)dx.
n Jo Jo
The f do not form a monotonically increasing sequence so that B.Levi's
n
theorem does not apply, and they are not bounded by some integrable
function that is independent of n so that Lebesgue's theorem does not
apply either. Considering f instead of / , we finally obtain a sequence
n n
for which Fatou's theorem does not hold.

As a corollary of Theorem 1.2.3 one has (approximate the derivative
by difference quotients):
Corollary 1.2.2 (Differentiation under the integral). Let I C R

d
be an open interval, A C R measurable, and suppose f : A x I *
R U { 0 0 } satisfies
(i) for any t I , / ( , t) is integrable on A
(ii) for almost all x A, / ( # , ) is differentiable on I
(Hi) there exists an integrable (j>: A R U { 0 0 } with the property that
for all t I and almost all x A
57/OM) <<Kx).
Then
is a differentiable function oft I , with
q.e.d.
2
Banach spaces
In this chapter, we present some results from functional analysis that will
be needed in the sequel, in particular in the next chapter. A l l proofs are
supplied. As a reference, one may use any good book on functional anal
ysis, e.g. K . Yosida, Functional Analysis, Springer, Berlin, 5th edition,
1978, pp. 52-5, 81-3, 90-92, 102-28, 139-45 or F. Hirzebruch, W . Schar-
lau, Einfuhrung in die Funktionalanalysis, Bibliograph. Inst., Mannheim,
1971, pp. 60-88, 107-12. (These were also our main sources when com
piling this chapter.)
2.1 Definition and basic properties of B a n a c h a n d H i l b e r t

spaces
Definition 2.1.1. A vector space V overR is called a normed space if
there exists a map
||-|| : V R, called norm,
satisfying
(i) | M | >0 for alive V, vÔ

(ii) ||Av|| = |A|\\v\\ for all A G R, v e V
(iii) \\v + w\\ < \\v\ \ + ||w|| for all v,w G V (triangle inequality)
A sequence (f )nN C
n V is said to converge to v V if
lim \\v n v\\ = 0.

noo
(In order to distinguish the notion of convergence just defined from
the notion of weak convergence to be defined in the next section, we
sometimes call it norm convergence or strong convergence.)
125
126 Banach spaces
A sequence (v )neN C V is called a Cauchy sequence if for every e > 0

n
we may find N G N such that for all n,m > N
\\v -Vm\\
n < .
A normed space (V, ||-||) is called a Banach space if it is complete w.r.t

the notion of convergence just defined, i.e. if every Cauchy sequence
converges to some v G V.
Examples
(1) Every finite dimensional normed vector space is a Banach space,
d
for example R w i t h its Euclidean norm |-|.
d
(2) Let K C R be compact. C(K) := {/ : K R continuous},
: = x r
ll/lloo P Z G A : l / ( ) l f f C(K)> defines a Banach space.
S U
m
If we equip C (K) := {/ : K R m-times continuously dif
ferentiable}, m G N , w i t h the norm I H I ^ , i t is not a Banach
space, because it is not complete. Namely the convergence w.r.t.
s u n r m
IN loo * i f convergence, and while the uniform limit of con
tinuous functions is continuous, in general the uniform limit of
differentiable functions is not necessarily differentiable.
(3) Let (V, 11-||) be a Banach space, W cV & linear subspace that is
closed w.r.t. ||-|| i.e. if (u> )nN C W converges to v G n
V^limn-ôo\\w - v\\ = 0), then v G W. Then (W, ||-||) is a Ba

n
nach space itself.
Definition 2.1.2. A Hilbert space is a vector space H overR equipped

with a scalar product, i.e. a map (-,-): H x H -*R satisfying
(i) (v, w) = (w, v) for all v,w G H

(ii) ( A i ^ i + \ v ,w)2= Ai(i>i,w) + A (t>2,w) for all A i , A
2 2 2 G R,
vi,v ,w
2 GH
(iii) (v, v) > 0 for alive H\ { 0 } .
In addition, we require
(iv) H is complete w.r.t. the norm \ \v\\ : = (v, i.e. a Banach space.
In order to justify the preceding definition, we need to verify that

II?;11 = (v, v ) i defines indeed a norm in the sense of Definition 2.1.1.
Since the properties (i), (ii) of Definition 2.1.1 are clearly satisfied, we
only need to check the triangle inequality:
2.1 Basic properties of Banach and Hilbert spaces 127
L e m m a 2.1.1. Let (-,-): H x H -> R satisfy (i)-(iii) of Definition

2.1.2. Then we have the Schwarz inequality: \(v,w)\ < \\v\ \ ||w|| for all
v,w G H, with equality if and only if v and w are linearly dependent.
Proof. We have for v, w G H, A G R
(v + \w, v + Xw) > 0 by (iii) .
Inserting A = and expanding with the help of (i), (ii) yields the
Schwarz inequality
2 2 2
Since + w | | = (v + w, v + w) = \\v\\ + | | w | | + 2 ( v , i u ) , the Schwarz
inequality in turn implies the triangle inequality.
q.e.d.
Definition 2.1.3. Let V be a vector space (overR, as always). M CV

is called convex if whenever x,y G M, then also
tx + (1 - t)y G M for allO <t< 1.
Example 2.1.1. Let (V, ||-||) be a normed space. Then for every \i < 0,
:= {# G V | ||x|| < fi} is convex. Namely if x,y G i.e. |x| <
< A*, then for 0 < t < 1
|to + (1 - t)y| <t\x\ + (l-t)\y\ <n,
hence -h (1 t)y G
The following definition contains a sharpening of the convexity of the

balls B^ I t will be formulated only for fi = 1, but by homogeneity ((ii)
of Definition 2.1.1), it implies an analogous condition for any fi > 0.
Definition 2.1.4. A normed space (V,||-||) is called uniformly convex

if for all e > 0 there exists 6 > 0 with the property that for all x,y eV
with \\x\\ = \\y\\ = 1, we have
> 1 \x - y\\ < e. (2.1.1)
Remark 2.1.1. A n equivalent form of the implication (2.1.1) is
x-y\\ > 1 => <l-6 (2.1.2)
(again for ||x| = 1).

128 Banach spaces
Example 2.1.2. I n a Hilbert space (H, (, )), we have the parallelogram

identity
2 ^ + 2/) (2.1.3)
which follows by expanding the norms in terms of the scalar product.

Therefore, any Hilbert space is uniformly convex.
L e m m a 2.1.2. In Definition 2.1.3, the condition \\x\\ = \\y\\ 1 may

be replaced by
Nl<i, llvll<i.
Proof. I n the situation of Definition 2.1.3, for eo > 0, we may find So > 0
such that for all z,w with \\z\\ = \\w\\ = 1, we have
> l-S =>

0 \\z-w\\ <e .0 (2.1.4)
Let now c > 0, ||x|| < 1, \\y\\ < 1. I f for 6 < \
1 - 6 <
then
II^H > 1 - 26 , | | y | | > l - 2 .
In particular, x, y ^ 0, and by the triangle inequality,
x
2(lWI +
IMl) > 2 ^ + ^)
R llyll
> 1 - 3<5.
We apply (2.1.4) with z = w Q = . I f 36 < <5Q, we then get
x y e
R llyll <2-
Now
x x
1^ < y
Ml + IMI +
by the triangle inequality
<4* + f.
Choosing 8 = min(3<5o, e/8), we have shown the implication
>l-6=>\\x-y\\<e
for ||*|| < 1, ||y|| < 1.

q.e.d.
L e m m a 2.1.3. Let (V, ||-||) be a uniformly convex Banach space. Let

(^n)nN C V be a sequence with
lim sup | | x | | < 1 for all n G

n (2.1.5)
and
lim 2 (n
x
H" %m) l. (2.1.6)
n,moo
Then (x ) n converges to some x G V with \\x\\ = 1.
Proof. Let e > 0. (2.1.5) and (2.1.6) imply l i m | | x | | = 1. Therefore,

n
by replacing x n by jjf^jp we may assume w.l.o.g | | x | | = 1. Because of

n
(2.1.5), we may apply Lemma 2.1.2. By (2.1.6), we may find N G N such

that for n , m > N
1
2 (nx
H" #m) > 1 " ,
w i t h (5 determined by Lemma 2.1.2. We obtain
|\%n %m j j ^ ^
i.e. ( # ) n e N is a Cauchy sequence. Since (V,||-|| is a Banach space, i t

n
has a limit x, and
\\x\\ = lim | | x | | = 1. n
q.e.d.
In order to formulate the Hahn-Banach theorem, a fundamental ex

tension result for linear functionals from a linear space to the whole
space, we need:
D e f i n i t i o n 2.1.5. Let V be a (real) vector space.
+
p : V -+ R+ (R : = {t G R | t > 0 } )
130 Banach spaces
is called convex if
(i) p(x + y) < p(x) + p(y) for all x,y eV

(ii) p(Xx) = \p(x) for all x eV, A > 0
Example 2.1.3. The norm on a normed vector space.
Let VQ be a linear subspace of the vector space V , /o : Vb R linear.

A linear / : V R is called an extension of fo if
f\v0 = /o-
T h e o r e m 2.1.1 ( H a h n - B a n a c h ) . Le Vo be a linear subspace of the

vector space V, p : V R convex. Suppose that /o : Vo R *5 linear
+
and satisfies
fo(x)<p(x) forallxeVo. (2.1.7)
T/ien there exists an extension f : V R o/ /o m^/i
/ ( x ) < p(x) for all xeV. (2.1.8)
Remark 2.1.2. We shall need the Hahn-Banach theorem only in the case
where V possesses a countable basis, i.e. is separable (see p. 130).
r
Proof. We may assume VQ ^ V. Let v G V \Vb, V\ be the linear subspace
of V spanned by Vo and v, i.e.
Vi : = {x + tv \ x G Vb, , t G l } .
We shall now investigate how /o can be extended to f\ : V\ R with
h(x)<p(x) forallxGî. (2.1.9)
We put fi(v) =: a. Then as an extension of /o, f\ satisfies
fi(x + tv) = f (x)

Q +ta.
Equation (2.1.9) requires
f (x)
0 + ta< p(x 4- tv). (2.1.10)
For t > 0, this is equivalent to
o<p(f+)-/ (f), 0 (2.1.11)
and for t < 0 to

(2 L12)
^-pHH-Mf)- -
w e
For # i , # 2 V, have
h(x ) 2 ~ fo{xi) < p{x 2 - xi)

= p((x 2 + v) - ( x i +v))
< p(x 2 + v) + p(~xi - v),
hence
~fo(x )
2 +p(x 2 + v) > -fo(xi) - p ( - x i - v). (2.1.13)
Thus
a2 : = inf ( - f {x )
0 2 + p ( x - f f))
2
> ai : = s u p ( - / o ( x i ) - p{-xi -v)).
Therefore, any a w i t h
ot\ < cx < a 2
satisfies (2.1.11) and (2.1.12), hence (2.1.10). Thus, the desired extension
/ i exists. I f V possesses a countable basis, we may use the preceding
construction to extend /o inductively to all of V.
If V does not possess a countable basis, we need to use Zorn's lemma
to complete the proof. For that purpose, let
<!>:={</?: W E extension of /o to some

linear subspace W, Vo C W C V ,
satisfying </?(#) < p(x) for all x G W}
On <, we have an obvious ordering relation (namely, for <>i : Wi > E,
i = 1,2, we have (pi < (p if Vî C W and 2 = <i), and every 2
totally ordered subset 3>o of $ possesses a maximal element, namely

<Po defined on the union of the domains of all (p <fio and coinciding
with each such (p on its domain of definition. By Zorn's lemma, $ then
contains a maximal element / . Let W be the domain of definition of / .
/ then extends fo to W. I f W were not the whole space V, we could
use the preceding construction to extend / to a larger subspace of V ,
contradicting the maximality of / . Therefore, / furnishes the desired
extension of /o.
q.e.d.
C o r o l l a r y 2 . 1 . 1 . Let Vo be a linear subspace of the normed vector space

(V, | | . | | ) , A > 0, f : V -> E linear with
0 0
\fo(x)\ < X\\x\\ for allxe V.0

132 Banach spaces
Then there exists an extension f : V R of /o with
\f(x)\ < X\\x\\ forallxeV.
T h e o r e m 2.1.2 ( H e l l y ) . Let (V, ||-||) be a Banach space, / i , . . . , / n
linear functionals V R t/iot ore continuous w.r.t. the norm conver

gence, / i , c * i , . . . , a G R. Suppose that for any X\,...,
n A G R n
(2.1.14)
Then for each e > 0, there exists x G V with e
fi(xi)=ai for t = 1 , 2 , . . . , n (2.1.15)
and
| | X | | < / * + .

Proof. Let m < n be the maximal number of linearly independent fi,

i = l , . . . , n . I t suffices to consider m linearly independent fi, w.l.o.g.
/i /m since the remaining ones are easily seen to be taken care of by
(2.1.14). F(x) : = ( / i ( x ) , . . . , fm(x)) may then be considered as a linear
m m
map onto R . We equip R w i t h its Euclidean structure. Let
Bp :={xeV\
+e \\x\\ < + }.
Then F(B^ ) e is a convex set containing 0 as an interior point. Also,

m
F(B^ )+e is balanced in the sense that w i t h p G R i t also contains p.
We now assume that a i , . . . , a is not contained i n F(JE? ). Be
m M+C
cause of the properties of F(B ) just noted, we may then find A =

fl+e
( A i , . . . , A ) with
m
^2 ^ iOLi
S U
P $>/<(*)
2=1 =1
TO
0* + ) 5 > / <
contradicting ( 2 . 1 . 1 4 ) . Thus (ai,...,a ) m G F(JB M + C ), implying the

claim. q.e.d.
2.2 D u a l spaces and weak convergence

Let V be a vector space. The linear functionals
f:V-+R
then also form a vector space. I f (V, ||-||) is a normed vector space, we
define the norm of a linear functional / : V E as
+ : = s u p I M G | + u W (2.2.1)
x*o IFII
L e m m a 2.2.1. A linear functional f :V E is continuous if and only

W I I . < o .
The easy proof is left to the reader. (See also Lemma 2.3.1 below.)
q.e.d.
D e f i n i t i o n 2.2.1. V* := {/ : V E linear with \\f\\^ < oo} equipped

with the norm (2.2.1) is called the dual space of (V, | | - | | ) . (It is easy to
verify that (2.2.1) defines a norm on V* in the sense of Definition 2.1.1.)
L e m m a 2.2.2. (V*, | | - | | ) is a Banach space.

#
Proof. Let (f )neN C V* be a Cauchy sequence. For every e > 0 we may

n
then find N G N such that for n, ra G N
\\fn-fm\l <.
By (2.2.1), this implies that for every x G V
\fn(x) ~ fm(x)\ < C.
Therefore, since E is complete, (f {x)) eN converges for every x G X.

n n
We denote the limit by f(x). f : V E then is a linear functional. I t is

an easy consequence of the triangle inequality that \\f\\* < oc and that
l i m _ o o | | / n - / I I * = -
n
T n i s
implies that (f )nen converges to / G V*, n
and (V*, I H U therefore is complete, hence a Banach space.

q.e.d.
Remark 2.2.1. We did not assume that V itself is a Banach space.
We now consider
(V*)* = : V**,
the dual space of V*, w i t h norm denoted by | | - | | . Any x G V defines ++
a linear functional
i(x) :V* -+R

i(x)(f) := (/,*) :=/(*).
134 Banach spaces
L e m m a 2.2.3. = Thus, the linear functional i(x) : V* >

E is contained in V**, i.e. we have a linear isometric map i : V V**.
Proof. We have
l ( / , * ) l < l l / I U M I ,
and therefore
INI> sup \ f ^ = \\i( \\. x) (2.2.2)

fev*
Conversely, let x G V. Let
f (tx)
0 := t\\x\\ for t G E.
By the Hahn-Banach theorem (Corollary 2.1.1), we may extend / 0 from

{ x | G E } t o V a s a linear functional / with
= 1
and
l ( / , * ) l = N I -
Therefore
l l ^ ) I L = sup i ^ > | | x | | . (2.2.3)
Equations (2.2.2) and (2.2.3) imply the result.

q.e.d.
D e f i n i t i o n 2.2.2. A normed linear space (V, ||-||) is called reflexive if
i :V -> V**
is a bijective isometry (i.e. \\x\\ = \\i(x)\\^^ for all x G V).
Remark 2.2.2.
(1) Since (V**, | | - | | ) is a Banach space by Lemma 2.2.2, any reflexive

##
space is complete, i.e. a Banach space.

(2) By the remark before Definition 2.2.2, the crucial condition in
that definition is the surjectivity of i.
D e f i n i t i o n 2.2.3.
(i) Let (V, 1 b e a normed linear space. (x )nGN C V is said to be n
weakly convergent to x G V if f(x ) converges to f(x) for all

n
f G V*, in symbols:
v
XJI x.
(ii) Let (V*, be the dual of a normed linear space. (f )nen C V* n
is said to be weak* convergent to f G V* if f (x) converges to n
f(x) for all x G V.
T h e o r e m 2 . 2 . 1 . Let V be a separable] normed linear space. Let

(/n)nGN C V* be bounded, i.e. | | / | | * ^ constant (independent of n).
n
Then (f ) contains a weak* convergent subsequence.

n
Proof. Let (y^^^n by a dense subset of V. Since (f (yi))neN is bounded,

n
a subsequence (fn(yi)) of (fn(yi)) converges. Having iteratively found

a subsequence (f) of ( / ) for which ( / ^ ( ^ ) ) n N converges for 1 <
n
+ 1
v < m, we may find a subsequence ( / ^ ) of (f) for which also
+1
( / r ( 2 / m + i ) ) n N converges. The diagonal sequence (/)nN then con
verges at every y , v G N , and since (y )v^
v is dense in V ,
v (fn{x)) en n
has to converge for every x G V. Thus, we have found a weak* convergent

subsequence of ( / ) e N . n n
q.e.d.
Remark 2.2.3.
(1) The argument employed in the preceding proof is called Cantor

diagonalization.
(2) Theorem 2.2.1 remains true without the assumption that V is
separable, and so does the following:
C o r o l l a r y 2 . 2 . 1 . Let (V, ||-||)6e a separable reflexive Banach space.

Then every bounded sequence (# )nGN contains a weakly convergent sub
n
sequence.
Proof. By (2.2.2) or reflexivity, (i(x )) n is a bounded sequence in

n ne
V** and therefore contains a weak* convergent subsequence. Since V is
f Separable means that V contains a countable subset {y )veN that is dense w.r.t.
u
1 i . e . for every y 6 V , e > 0 there exists y with \\y y \\ < e.

u u
136 Banach spaces
reflexive, the limit is of the form i(x) for some x G V. Thus
f(x )n = (/, x ) -> ( / , ar) = f(x)

n for every / G F *
so that ( x ) N converges weakly to x.

n n G
q.e.d.
T h e o r e m 2.2.2. Am/ weakly convergent sequence (# )nN n o, Banach

space is bounded.
Proof. We shall show that i(x ) n is uniformly bounded on

{feV*\ H/IL < 1 } . Then also
I M = I I ^ ) I I = sup (2.2.4)
fev* 11/11*
is bounded (see Lemma 2.2.3 for the first equality). Since i(x ) is linear, n
it suffices to show uniform boundedness on some ball in V*. Otherwise,

we find a sequence Bj of closed balls,
Bj = { / G V* | ||/ - fj\\ < Qi} for some ft e V* , Qj > 0
with
Bj+x C JBJ and l i m Qj = 0
j*oo
and a subsequence (x' ) of (x )

n n with
|(/,x;)|>i for all fsBj. (2.2.5)
By construction, (fj)jeN forms a Cauchy sequence and therefore con

verges to some /o G V*, with
oo
/o e n -Si-
Because of (2.2.5), we have
|(/o,Ol > J foralljGN.

f
This is impossible since (fo,x ) n converges because (#{JnGN converges
weakly. q.e.d.
Example 2.2.1.
(1) I n a finite dimensional normed vector space (which automati

cally is complete, i.e. a Banach space), weak convergence is just
componentwise convergence and therefore equivalent to the usual
convergence w.r.t. the norm.
(2) I n an infinite dimensional reflexive Banach space (V, | | - | | ) , this is

no longer so, because one may always find a sequence ( e ) N C V n n G
w i t h ||e$|| < 1 for all i and ||e* ej\\ > 1 for i ^ j . Such a
sequence cannot converge w.r.t. ||-||, because i t is not a Cauchy
sequence, but i t always contains a weakly convergent subsequence
according to Corollary 2.2.1 (we have shown Corollary 2.2.1 only
under the assumption of separability, but i t holds true in general).
L e m m a 2.2.4. Let (V, ||-||)6e a separable normed space. Then V* satis

fies the first axiom of countability w.r.t. the weak* topology, i.e. for each
f G V*, there exists a sequence (U^^jq of subsets of V* that are open
in the weak* topology such that every U that is open in this topology
and contains x is contained in some U . Consequently, if (V, ||*||)is also
n
reflexive, then V* satisfies the first axiom of countability w.r.t. the weak
topology.
Proof. Let f eV*. Every neighbourhood of / w.r.t. the weak* topology

contains a neighbourhood of the form
U , ,..., (f)-={geV*\
t vl Vk \g(vi) - f(v )\ t < fori = l,...,fc}.
Since V is separable, there exists a sequence (w ) n C V that is dense n ne
w.r.t the 1 t o p o l o g y . We claim that the neighbourhoods of the form
u ... (f)
tWilt tWik
form a basis of the neighbourhood system of / of the required type, i.e.

every U ,,, (/)
e;Vu yVk contains some such U. w }... iWik ( / ) For that pur
pose, we choose n w i t h ^ < e and , . . . , Wi w i t h \VJ Wi | < ^ for
k j
j = 1 , . . . , k. For g G U. Wii ^ . . ^ ( / ) , we then have
\9(VJ) - f(vj)\ < | $ K ) - / K ) | + \(g - f)(vj - wij)\ < i + ~ < 6,
i.e. g G v . ., (f)
Vlim as required.
iVk
Finally, i f V is reflexive, then the weak* and the weak topology of V*

coincide.
q.e.d.
We now present some further applications of the Hahn-Banach theo

rem that will be used in Chapter 3.
L e m m a 2.2.5. Let (V, ||-||) be a normed space, Vo a closed

linear subspace. Then VQ is also closed w.r.t. weak convergence.
138 Banach spaces
Proof. By the Hahn-Banach theorem (Corollary 2.1.1), for every XQ G

V \ Vb, we may find a continuous linear functional fo : V R w i t h
/o(x ) = 1
0
/ok=0.
Thus, xo cannot be a weak limit of a sequence in Vb-

q.e.d.
L e m m a 2.2.6. Let (V, ||-||)6e a reflexive Banach space, VQ a closed

linear subspace. Then Vb is reflexive.
Proof We may identify VQ** w i t h a subspace of V * * , by putting v(f) =

v(f\ )
Vo for f eV*, v e Vb**. Let v G Vb**. Since V is reflexive, there
exists x G V with
!>(/) = / ( * ) for all / G T .
We claim x G Vb- Otherwise, by the Hahn-Banach theorem (Corollary

2.1.1), there exists f eV* with
/(*) o
/ k = o .
Since / ( x ) = v(f\y ) by the above, this is impossible. Since every fo G

0
VQ can be extended to / G V*, again by Hahn-Banach, we conclude
v(fo) = fo(x) forall/GVo*.
Thus, v = i(x). This implies VQ* = i(Vb), i.e. reflexivity of Vb.

g.e.d.
Corollary 2.2.2. yl Banach space (V, ||-||)is reflexive if and only if its
dual (V*, \ is reflexive.
Proof. I f V = F**, then also F* = V***. Thus, i f V is reflexive, so is

V*. Consequently, i f conversely V* is reflexive, so then is V**. Since V
can be identified w i t h a closed subspace of V** by Lemma 2.2.2, Lemma
2.2.6 then yields reflexivity of V.
q.e.d.
L e m m a 2.2.7. Let (V, ||-||)6e a normed space, and suppose that

( x ) N C V converges weakly to x G V . Then
n n G
\\x\\ < l i m i n f | | x | | .
n
Proof. After selection of a subsequence, we may assume that | | x | | con n
verges (see Theorem 2.2.2). Assume
||x|| > l i m | | x | | . n
n>oo
As in the proof of Lemma 2.2.3, we may find f eV* with
11/11. = i
l/(x)| = I N I -
But then
| / ( x ) | > lim | | x | | > l i m s u p | / ( x ) | ,
n n
n*oo >oo
n
while the weak convergence of ( # ) e N to x implies

n n
f(x) = lim f(x ).

n
n+oo
This contradiction establishes the claim.
q.e.d.
T h e o r e m 2.2.3 ( M i l m a n ) . Any uniformly convex Banach space is re

flexive.
Proof (Kakutani). Let (V, ||-||)be a uniformly convex Banach space, and
let XQ* G V**. We need to show that there exists some x V w i t h 0
t ( x ) = x*,*
0 (2.2.6)
(see Remark 2 after Definition 2.2.2). We may assume w.l.o.g. that
11*5*11 = 1- (2-2.7)
For every n N , we may then find / V* w i t h | | / | | = 1 and n
1 - - < x* *(f ) 0 n < 1. (2.2.8)

n
We now claim that for every n G N , we may find x n G V with
fi(x )=x* *(fi)

n 0 fort = l , . . . , n (2.2.9)
and
M<||xS*|| + 1+ -. (2.2.10)
n
For any A i , . . . , A G R, we have
n
4* U > / i < iixoir

|i=l
140 Banach spaces
and so the claim follows from Helly's Theorem 2.1.2. Since in addition
to (2.2.10) also
I K H = ||/n|| I k n l l > fn(Xn) = %0*(fn) > 1 ~ K
we must have
lim | | x | | = 1. n
n*oo
For ra > n, we have
2 2
2 - - < fn(Xn) + fn(Xm) < ||x + X \\ < \\x \\ + | | x | | < 2 + - . m m m
n n
By Lemma 2.1.3, ( # ) n N is a Cauchy sequence and converges to some
n
xo eV, satisfying
I |xo|| = 1 (2-2.11)
and
/<(*o)=x5*(/<) f o r i = 1,2,3,... (2.2.12)
The solution XQ of (2.2.11), (2.2.12) is unique. Namely, if there were
f
another solution x , on one hand, we would have
0
I N + Xoll < 2 (2.2.13)

by uniform convexity. On the other hand
f o r a 1 1
fi(x 0 + x' ) = 2x1*(fi)
0 h
hence
2 - T < 2xl*(fi) = fi(x 0 + x' ) < \\x + 4 | | ,

0 0
hence
2
I N + 4 1 1 > -
This contradicts (2.2.13), and so x 0 is unique. We now claim that
fo(x ) 0 = *5*(/ ) 0 for any / 0 G V\ (2.2.14)
so that XQ* = i(xo), proving the theorem. Let this /o G be given.

In the above reasoning, we replace the sequence / i , / 2 , / 3 , . . . by
/o, fi, / 2 , / 3 , We then obtain a?Q G V with
and
/<K)=*S*(/*) for i = 0 , l , 2 , 3 , . . . (2.2.15)
Since the solution x of (2.2.11), (2.2.12) was shown to be unique, how

0
f
ever, we must have x = x$. Equation (2.2.15) for i = 0 then is (2.2.14).
0
q.e.d.
Corollary 2.2.3 ( R i e s z ) . Any Hilbert space ( H , (, )) can be identified

with its dual H*.
Proof. Since a Hilbert space is uniformly convex, Therem 2.2.3 implies

H = H**. On the other hand, any x G H induces an f G H* by x
f {y)
x '= {x,y) for y G H.
We have
\\fx\\ = sup (x,y) < \\x\\

llvll=i
2
and f (x)
x (x,x) = \\x\\ , hence
ll/xll = I N | .
Thus, H is isometrically embedded into H*. For the same reason, H* is

isometrically embedded into H**, and since H = i f * * , one readily verifies
that these embeddings must be surjective, hence H = H* = H**.
q.e.d.
Let M be a linear subspace of a Hilbert space H. The orthogonal

L
complement M of M is defined as
L
M := {x e H : (x, y) = 0 for all y e M} .
L
It is clear that M is a closed linear subspace of H. M need not be
closed here, but the orthogonal complement of M is the same as the one
of its closure M in H.
Corollary 2.2.4. Let M be a closed linear subspace of the Hilbert space

H. Then every x G H can be uniquely decomposed as
1
x = xi + x 2 with xi G M , x 2 G M -.
Proof. By the proof of Corollary 2.2.3, x G H corresponds to f x G H*

with
f (y)
x = (x,y) for all y G H.
We let f^f be the restriction of f x to M. M , since closed, is a Hilbert

142 Banach spaces
space itself, and f*f is an element of the dual M*. By Corollary 2.2.3,
it corresponds to some X\ M , i.e.
/ f ( y ) = ( * i , y ) for all y G M .
We put #2 = x x\. Then for all t/ M ,
(a? - = f (y) x - f**(y) = 0 since = / f on M.

x
Therefore, #2 M . Thus, we have constructed the required decom
position. Concerning uniqueness, i f
w n 1
x X\ + #2 = # i 4- #2 ^ # i Af, #2, #2 ^ Af" ,
then for all y M
Or, j/) = =
and by Corollary 2.2.3 applied to M , # i = x'j, and therefore also x 2 = x' .

2
q.e.d.
Of course, the reader knows the preceding result in the case where H
is finite dimensional, i.e. a Euclidean space. x\ is interpreted as the
orthogonal projection of x onto the subspace M , and therefore Corollary
2.2.4 is called the projection theorem.
The next result will be needed for Sections 4.2 and 4.3 when we estab
lish the existence of minimizers for lower semicontinuous, convex func
tionals.
T h e o r e m 2.2.4 ( M a z u r ) . Suppose (x )nm converges weakly to x in n
some Banach space V. For every e > 0, we may then find a convex
combination
N N
]P A n X n ( A > 0, n ]T A n = 1)
71=1 71=1
with
N
< e. (2.2.16)
Proof. We consider the set Co of all convex combinations of the x , i.e. n
( N N \
X n X n w i t h
Co : = I > 0, ^ A = 1> .
n
<n=l n=l
Replacing all x by x x\ and x by x - # i , we may assume 0 Co. I f

n n
(2.2.16) is not true, then there exists e > 0 w i t h
\\x-y\\>e for all y G C .

0 (2.2.17)
Ci := {z G V : ||z - y\\ < | for some y G C } 0
is convex and contains the ball with radius | and center 0. We consider
the Minkowski functional p of C\ defined by
p(z) : = i n f { A > 0 ; A - ^ G C i } .
p is convex i n the sense of Definition 2.1.5 since C\ is convex, and contin

uous since C\ contains the ball of radius | > 0 about 0. Since, because
of (2.2.17),
||x z\\ > ^ for every z G C i ,
we have
p(x) > 1.
More precisely, there exists yo w i t h

_1
x = A i/o, 0 < A< 1
p(yo) = 1.
We consider the linear subspace
V = {Mo,tiR}C
0 V
and the linear functional
/o() = fionV .
0
Then
fo <p on Vb,
and by the Hahn-Banach Theorem 2.1.1, there exists an extension / of

fo to all V w i t h
Since p is continuous, / is also continuous (see Lemma 2.2.1). We have
sup f(y) < sup f(y) < sup p(y) = 1

yeC0 yCi yCi
1
< A " = fiX-'yo) = /(*).
144 Banach spaces
This, however, contradicts the fact that ( x ) j v C Co converges weakly

n n
to x. Thus, (2.2.17) cannot hold, and (2.2.16) is established.

q.e.d.
2.3 L i n e a r o p e r a t o r s b e t w e e n B a n a c h spaces
The results of this section will be used in Chapter 8. I n Section 2.2, we
considered linear functionals
in the beginning, V was a normed linear space, w i t h norm denoted by

1 a n d later, we also assumed that V was complete, i.e. a Banach space.
In the present section, we replace the target E by a general Banach space
W, with norm also denoted by ||*||. We thus consider linear operators
T:V ->W,
and we put
\\Tx\
|r||:=supiL-i E+U{oo}. (2.3.1)
IF!I
L e m m a 2 . 3 . 1 . The linear operator T : V W is continuous if and
only if\\T\\ < oo.
Proof. I f | | T | | < oo, then the inequality
IW<||r|||N| (2.3.2)
implies that T is continuous. (Of course, this uses the linearity of T.)
Conversely, if T is continuous, we recall the usual e 6 criterion for
continuity, and so for e = 1, we find some 6 > 0 with the property that
\\Ty\\ < 1 if Ili/H < 6.
For x G V \ { 0 } , we then have with y = SjAr (\\y\\ < 6)

PIT
\\Tx\ [
Ty
Thus
l|T||<<oo.
q.e.d.
The space of continuous linear operators T : V W between the

normed spaces (V, ||-||) and (W, ||-||) is denoted by L(V, W). I t becomes
a normed space w i t h norm | | T | | .
L e m m a 2.3.2. If(W, ||-||) is a Banach space, then so is (L(V, W ) , | | - | | ) .
The proof is the same as the one of Lemma 2.2.2, simply replacing (R, | |)
b y W I N D .
Remark 2.3.1. Again, (V, ||-||) need not be a Banach space here.
L e m m a 2.3.3. Let T G L(V, W). Then
k e r T : = {x G V : Tx = 0}
is a closed linear subspace of V.

- 1
Proof, ker T = T (0) is the pre-image of a closed set under a continuous
map, hence closed.
q.e.d.
In the sequel, we shall encounter bijective continuous linear operators
T :V ->W
between Banach spaces. I t is a general theorem in functional analysis,

- 1
the inverse operator theorem, that the inverse of T, denoted by T , is
then continuous as well. Here, however, we do not want to prove that
1
result, and we shall therefore frequently assume that T" is continuous
although that assumption is automatically fulfilled in the light of that
theorem.
L e m m a 2.3.4. Let
T:V ->W
be a bijective continuous linear map between Banach spaces, with a con

- 1
tinuous inverse T . If S G L(V,W) satisfies
IIT-SIKpijjp (2.3.3)
l
then S is bijective, and S~ is continuous, too.
Proof. We have
1
S = T(Id-T~ (T-S)).
146 Banach spaces
As with the geometric series, the inverse of S then is given by
1
^(r~ (r-s)H T~\ (2.3.4)
^=0
provided that series converges. However,
l
l
Y^{T- {T-s)y <J2\\(T- (T-s)y\\
vm
< ^(WT-'WWT-SWY,
and since |J^ 11J \\T - S\\ < 1 by assumption, the series satisfies the
Cauchy property and hence converges to a linear operator w i t h finite
norm.
q.e.d.
If V is a vector space, we say that V is the direct sum of the subspaces
v = Vi e v 2
if for every x G V, we can find unique elements X \ G V i , x G V , with 2 2
x = X\ + x.
2
We then also call V\ and V complementary subspaces of V . Easy lin

2
ear algebra also shows that if V\ possesses a complementary subspace of

finite dimension, then the dimension of that space is uniquely deter
mined, i.e. i f Vi 0 V = Vi 0 V ', then dim V = dim V .
2 2 2 2
We now consider a normed vector space (V, | | - | | ) . Then every finite

dimensional subspace Vo is complete, hence closed. We also have:
L e m m a 2.3.5. Let Vo C V be a finite dimensional subspace of the

normed vector space (V, | | - | | ) . Then Vo possesses a closed complemen
tary subspace V\, i.e. V = Vo 0 V\.
Proof. Let e i , e n be a basis of V , /Q : Vo * R be the linear function

0
als with
e s
fo( i) = ij (hj = l , . - , n ) .
3
By Corollary 2.1.1, we may find extensions p : V R with f ^ = f^.
We define 7 r : V > V as
n
7T is continuous, with ir(V) = Vb.
Vi : = ker 7r
then is closed as the kernel of a continuous linear operator (Lemma

2.3.3), and every x V admits the unique decomposition
x = n(x) + (x n(x))
w i t h ir(x) Vb, x ir(x) V i , because 7r O TT = 7r.
D e f i n i t i o n 2 . 3 . 1 . Lef T : V -+ W be a continuous linear operator

between Banach spaces (V, ||-||) ond! (W, | | - | | ) . T is co//ed! o Fredholm
operator if the following conditions hold:
(i) Vb = k e r T is finite dimensional. Consequently, according to

Lemma 2.3.5, there exists a closed subspace V\ ofV with
V = VbSVi. (2.3.5)
(ii) There exists a finite dimensional subspace WQ of W, called the

cokernel ofT (cokerT) giving rise to a decomposition ofW into
closed subspaces
W = W 0 W\ (2.3.6)
with
Wi = T(V) =: R(T) {range ofT).
Thus, T yields bijective continuous linear operator T\ : V\ > W\.

We finally require
(iii) T~ : W\ > V\ is continuous.
l
For a Fredholm operator T, we call
ind T : = dim V - dim W

0 0 ( = dim ker T - dim coker T )
the index ofT. The set of all Fredholm operators T : V > W is denoted
by F(V,W).
Remark 2.3.2. Question to the reader: Why is F(V,W) not a vector

space?
148 Banach spaces
Remark 2.3.3. As mentioned, condition (iii) is automatically satisfied as

a consequence of the inverse operator theorem.
1 o u r
Remark 2.3.4- I * conventions, the cokernel of T is only determined
up to isomorphism, i.e. any Wo satisfying (2.3.6) w i t h W\ = T(V) is a
cokernel. Usually, one defines the cokernel as the quotient space WjW\,
but here we do not want to introduce quotient spaces of Banach spaces.
T h e o r e m 2.3.1. Let V,W be Banach spaces. F(V,W) is open in

L(V,W), and
ind : F(V, W) -> Z
is continuous, hence constant on each connected component of F(V, W).
Proof. Let T : V > W be a Fredholm operator. We use the decomposi

tions
V - Vo 0 Vi w i t h Vo = k e r T
W = Wo 0 Wi w i t h W = coker T 0
of Definition 2.3.1. For S G L(V, W ) , we define a continuous linear op

erator
S' :V xWo^W
x
(x, z) v-* Sx - f 2,
and we obtain a continuous linear operator
L(V, W ) > L ( V i x i y , W ) 0
;
Since T\ : V\ > VFi is bijective w i t h a continuous inverse, T is also
bijective w i t h a continuous inverse, and by Lemma 2.3.4 this then also
holds for all S in some neighbourhood of T. For such 5, S'(Vi) is closed
as Vi is closed and S' is continuous, and we have the decomposition
f
w= s'(v )es {w ),
1 0
;
and since 5 ( V i ) = 5 ( V i ) also
/
W = 5(V"i)0 5 ( i y ) , o (2.3.7)
and since Wo is finite dimensional, so is S'(Wo). Then S(V) D S(V\) is

also closed since S(V\) is closed and possesses a complementary subspace
of finite dimension.
Finally, the dimension of the kernel of S is upper semicontinuous.
Namely, i f S is in our above neighbourhood of T , then since S is bijective,

S is injective on V i , and hence the kernel of S is contained in some
complementary subspace of V i , and as observed above, the dimension of
such a subspace equals the one of Vo- Thus
dim ker S < dim ker T (2.3.8)
if S is in a suitable neighbourhood of T in L(V, W).

Altogether, we have verified that S is a Predholm operator i f i t is
sufficiently close to T.
Prom the preceding, we see that there exist finite dimensional sub-
spaces VQ = ker S and V " of V with 0
v = v 'ev "eVi,
0 0
and thus
dim V ' + dim V " = dim V

0 0 0 (V = ker T ) .
0 (2.3.9)
S thus is injective on V " V i , and since S coincides w i t h S' on V i , we

0
get a decomposition
w= s(v )s(v)w^
1
with WQ = cokerS and from (2.3.7)
dim S(V ") + dim W = dim S'(W )

0 0 = dim W 0 (2.3.10)
f
since S is bijective.
Consequently
ind S = dim ker S dim coker S

/
= dimV 0 -dimV^
/;
= (dim V - dim V ) - (dim JV - dimS(V ")) by (2.3.9), (2.3.10)
0 0 0 0
/;
= dim Vo - dim W 0 since S is injective on V 0
= indT.
for S in some neigborhood of T .

q.e.d.
The following result motivates the definition of a Predholm operator:
T h e o r e m 2.3.2 ( F r e d h o l m a l t e r n a t i v e ) . Let V be a Banach space,

T : V V a Fredholm operator of index 0. We consider the equation
Tx = y. (2.3.11)
150 Banach spaces
Either
(i) Either Tx = y is solvable for all y, and thus T is surjective, hence

also injective as i n d T = 0, and so the solution x is uniquely
determined by y,
or
(ii) Tx = y is only solvable if y is contained in some proper subspace

ofV (with a finite dimensional complementary subspace), and for
each such y, the solutions x constitute a finite dimensional affine
subspace.
Proof. A direct consequence of the definition.

q.e.d.
2.4 C a l c u l u s in B a n a c h spaces
In this section, we collect some material that will only be used in Chap
ters 8 and 9.
Definition 2.4.1. Let (V, \\-\\v), (W, ||-||w) be Banach spaces, F :V ->
W a map. F is called differentiable (in the sense of Frechet) at u V
if there exists a bounded linear map
DF(u) : V -+ W
with
lim \\nu + v)-F(u)-DF um\\ { w =
0
<:;o> IHIv
/ is called differentiable in U CV if it is differentiable at every u EU.
1
f is said to be of class C if DF(u) depends continuously on u. f is
2
said to be of class C if DF(u) is differentiable in u and the derivative
2
D F(u) := D(DF)(u) depends continuously on u.
It is easy to show that a differentiable map is continuous.

We now wish to derive the implicit and inverse function theorems in
Banach spaces that will be used in Chapter 8. We shall need a technical
tool, the Banach fixed point theorem:
L e m m a 2.4.1. Let A be a closed subset of some Banach space (V, | | - | | ) .

Let 0 < q < 1, and suppose G : A A satisfies
\\Gyi - Gy \\ < q\\yi - y \\

2 2 for all y y
u 2 e A. (2.4.2)
2.4 Calculus in Banach spaces 151
Then there exists a unique y G A with
Gy = y. (2.4.3)
If we have a continuous family G(x) where all the G(x) satisfy (2.4-2)
(with q not depending on x), then the solution y = y(x) of (2.4-3) de
pends continuously on X.
Proof. We choose yo G A and put iteratively
Vn := Gy -\. n
We have
n n
G _ 1 Gi l
Vn = (W - W - 0 +2/0 = 2 ( ' yi - ~ Vo) + Vo- (2-4.4)
i=l i=l
We obtain from (2.4.2)
1 1
E l l G ' - V - G^j/oll < X y - | | y i - ftll < j - Wvi ~ Vo\\
9
t= l i=l
Consequently, the series y in (2.4.4) converges absolutely and uniformly

n
to some y G A, noting that A is assumed to be closed and the limit

function y = y(x) is continuous. We have
V = l i m Gy n = G (lim y) n = Gy,
noo \noo /
hence (2.4.3). The uniqueness of a solution of (2.4.3) follows from (2.4.2),

since q < 1.
q.e.d.
T h e o r e m 2.4.1 ( I m p l i c i t F u n c t i o n T h e o r e m ) . Let Vi,V ,W be Ba 2
nach spaces with all norms denoted by \\-\\, U C V\ x V open, (xo>S/o) G 2
1
U, F C (J7, W ) , i.e. F is continuously differentiable. For purposes of
normalization solely, we assume
F(x ,y )=0. 0 0 (2.4.5)
We also suppose that
D F(x ,y ):V ^W,

2 0 0 2
the derivative of F(XQ, ) '- V ^ W at y = yo, is invertible. By our differ

2
entiability assumption, D F(xo yo)2 is continuous, and we assume that

1
152 Banach spaces
its inverse is likewise continuous. Then there exist open neighbourhoods

U\ of XQ, U of yo with U\ x U U, and a differentiable map
2 2
(p:U Û 1 2
with
F(x,(p(x))=0 (2.4.6)
and
1
D<p(x) = - ( D F ( x , ifix)))-
2 o D F(x 1 1 tp(x)) for all xeUx
(2.4.7)
(D F(-,y)
x : Vi -+ W is the derivative of F(-,y) : Vi -+ I V ) . I n /ac*, /or
et;er?/ x C/i, <p(x) is the only solution of (2.4-6) in U . 2
The content of the implicit function theorem is that the equation
F(x,y)=0
can be solved locally uniquely for y as a function of x, i f the derivative

of F w.r.t. y is continuously invertible.
Proof. The idea is to transform the problem into a fixed point problem
for which the Banach fixed point theorem is applicable. We put
l:=D F( yo)-
2 X(h
W i t h this notation, our fixed point equation is

1
$(x,y):=y-r F(x y)=y 1 (2.4.8)
which clearly is equivalent to our orginal equation F(x, y) = 0. For every

x, we thus want to find a fixed point of
&(x,y).
l
Using l~ o / == id (note that / is invertible by assumption), we get
x
*(ar, yi) - y ) = l~ (D
2 2 F(x 0j y )(yi
0 - y ) - (F(x, yi) - F(x,
2 y )).
2
In Lemma 2.4.1, we take q = ^, and by the differentiability of F at

x 1
( Oiyo) and the continuity of Z"" , we may find 6* > 0, > 0 with the
property that for
f
\\x xo\\ < 6
and
112/1 ~ 2/o 11 < II2/2 - 2/011 < ( hence also - y \\ < 2e ) ,

2
we have
| | * ( r r , y i ) - *(a?,ifc)|| < ~ | | y i - y \\.

2
Furthermore, we may find <5" > 0 w i t h the property that for
II*-soil
we have
||*(ar,2/o) - * ( o , 2 / o ) | | < | -
Since $(sco, t/o) = 2/o by assumption, we then have for ||y j/o|| <
- voll < - * ( z , y ) l l + ll*(,yo) - * ( a o , > ) | |

0
< ^lly-yoll + l <e
whenever x | | < <5 : = min(<5',6"). This means that i f \\x #o|| < <5,
0
$ ( x , y) maps the closed ball
A : = {y Vi : | | y - l / o | | <
onto itself. By Lemma 2.4.1, for every x with \\x x \\ < <5, there exists 0
a n
a unique y =: tp(x) w i t h ||y y | | < d 2/ = $ ( # , t/), i.e. F(x,y)
0 = 0.
Moreover, t/ depends continuously on x. We consider the open balls
Ux : = { x : H * - soil < } , ^2 : = { 2 / : ||y - l/o|| < e } .
(<&(sc, ) also maps the open ball t/2 onto itself.) By choosing <$, > 0
smaller, i f necessary, we may assume
U1XU2C U.
I t remains to show that <p(x) is differentiable and that its derivative is

given by (2.3.7). We consider
(x <p(xi))
u eUxX U 2y
and abbreviate y\ := <p(x\). We put
h := DiF(xi,yi),l 2 := D F(x ).
2 uyi
Since F is differentiable, we may write
F(x,y)=h(x-xi)+ l (x - x ) + r(ar,y)
2 2
where the remainder term satisfies

r
lim Tl 4Ml M = 0
- 2 4
( - - ) 9
V-+V1
154 Banach spaces
Since F ( x , <>(#)) = 0 for x U\ by construction of we obtain

X
ip(x) = -l2 h(x - x i ) + y i - Z^VC^, <p(x)). (2.4.10)
By (2.4.9), we may find rj,p> 0 such that for
I k - x i H < rj, | | 2 / - 2 / i | | v)\\ < || i||(lk-*i|| +

2 r \\y-yill)-
Thus
I K * , <p(x))\\ < (\\x - \\

Xl + \\<p(x) - ( x i ) | | ) .
V (2.4.11)
By (2.4.10), (2.4.11),
| | ^ ( x ) - < ^ i ) l l < WtfhW\\x - x i | | + \ \\x - x i 11 + ~\\<p{x) - V (xi)||,
hence
||(/?(x) ^ ( x i ) | | < c\\x x\\\ for a constant c.
We abbreviate r (x) 0 : = - Z ^ r ^ , </?(#)) and rewrite (2.4.10) as
ip(x) - ip(xi) = -l^hix - xi) + r ( x ) ,

0 (2.4.12)
with
r
lim , , ^ = 0 from (2.4.9). (2.4.13)
x-+xi \\x Xi\\
(2.4.12) and (2.4.13) yields the differentiability of (p and (2.4.7).
q.e.d.
C o r o l l a r y 2.4.1 (Inverse F u n c t i o n T h e o r e m ) . LetV,W be Banach

spaces, U C V open, yo G U. Let f : U W be continuously differen
tiable, and assume that the derivative Df(yo) is invertible with a con
tinuous inverse. Then there exist open neighbourhoods U2 C U of yo,
U\ of f(yo) = : XQ SO that f maps U2 bijective onto U\, and the inverse
l
 U2 is differentiable with
1
ZtyOro) = (Dfiyo))' . (2.4.14)
Proof. We shall apply Theorem 2.4.1 to F(x y) := f(y) x, and find}
an open neighbourhood U\ of XQ and a differentiable function
if : Ui -+ V
with (f(U\) C U for a neighbourhood U of yo, w i t h (^(xo) = yo and

2 2
F(x,(p(x)) = 0, i.e. f(<p(x)) = x for x e U\.

1
As y>(/i) = f~ (U\) is open, we may redefine t/2 as </?(C/i), and y> then
yields a bijection between U\ and U . As f(<p(x)) = the chain rule
2
implies
= L e 2 4 1 4
Df(<p(x ))
0 Zty(a? )) 0 " ( - - )-
q.e.d.
The next topic concerns ordinary differential equations in Banach spaces.

I n Chapter 9, we shall use the Picard-Lindelof theorem in a Banach
space that we shall now derive.
We need the integral of a continuous function
x . I ^ V
from some interval / = [a, 6] C R into some Banach space V ,
/ x(t)dt.
Ja
This can be defined as a Riemann integral as in the case of real-valued

functions through approximation by step functions.
Given a continouous
$ : R x V -+ V,
we say that x(t) solves the ODE (ordinary differential equation) on / ,
4-x(t) = x(t) = x(t)) with x(a) = x 0 (2.4.15)

dt
if for silt e I
x(t)=x + 0 [ $(T,x(T))dT. (2.4.16)
T h e o r e m 2.4.2 ( P i c a r d - L i n d e l o f ) . Suppose that $ is uniformly Lip-

schitz continuous, i.e. suppose there exists some L < 00 with
||*(ti,xi) - *(t ,x )|| < L

2 2 - t \ 4- | | x i -
2 x \\)
2
for allt G I,xi,x 2 G V. (2.4.17)
Then for any XQ G V, there exists a unique solution of (2.4-15).

156 Banach spaces
Proof. We shall solve (2.4.16) w i t h the help of Lemma 2.4.1. For a con
tinuous y : I V , we define Gy G C ( J , V ) ,
(Gy)(t) :=x + / 0 *(r,(r))dT

./a
We note that C(7, V ) , the space of continuous functions from / with
values in V , is a Banach space w.r.t. the norm
\\y\\ :=
c0 sup||y(t)||.
(To verify this, one just needs to observe that any sequence (y )neN n C
C(I,V) with
lim \\y - y \\ o
n m C (= l i m sup \y (t) - y (t)\)n m = 0
n,moo \ n,m*oo f^j J
converges uniformly to some continuous function y : I V.)

We have
- Gy \\2 co = supl f (*(T,y!(T)) - *(T,H,(T)))dr
<\t-a\ L\\yi - y \\ 2 Co because of (2.4.17).
We choose e > 0 so small that

u<_\.
Lemma 2.4.1 with V replaced by C([a, a + e], V ) and with q = \ then

implies that there exists a unique y G C([a, a + c], V) with
Gt/() = #o 4- / $ ( T , y(r))dT for a < t < a + e.
Repeating the construction with a - f e in place of a and y(t + e) in place

of #o yields the solution on [a, a + 2c], and so on.
q.e.d.
Remark 2.^.1. I f / is an infinite or semi-infinite interval, e.g. / = [a, oo),

and if (2.4.17) holds on J, we obtain a solution of (2.4.15) on / , since
Theorem 2.4.2 yields a solution on every interval [a, 6] with 6 < oo.
C o r o l l a r y 2.4.2. Let the assumptions of Theorem 2.4-2 be satisfied on

the interval I = [0, oo), and suppose that $ does not depend explicitly
ont, i.e. $ : V V, $ = $(#). For x G V we thus consider the ODE0
x(t) = *(x(t)), x(0) = x . 0 (2.4.18)

Exercises 157
(x(0), the value at 'time' 0, is called initial value). We denote the solu
tion by X(XQ, t). Then for s, t > 0,
x(x$, t + s) = x(x(t), s) (semigroup property).

1
Thus, the solution with initial value XQ at 'time 1 + s is the same as
the solution with initial value x(t) computed at 'time's.
Proof. This follows from the uniqueness statement i n Theorem 2.4.2, as

both sides of (2.4.18) are solutions.
q.e.d.
Exercises
2.1 Let (V,\\'\\y) (W, \ \ - \ \ ) be normed linear spaces. For a linear
w
functional
/ : V - . W,
put
x
xev\{o} \\ \\v
Show that / is continuous iff | | / | | < oo. Let L(V,W) := {/ :
V -+ W linear w i t h | | / | | < oo}. Show that i f ( W ^ I H I ^ ) is a
Banach space then so is (L(V, W), ||-||).
2.2 Show that a normed space (V, ||-||) is uniformly convex i f the
following condition holds:
Whenever ( x ) , (y )nN C V satisfy
n n N n
limsup||ar || < 1 , limsup | | y | | < 1

n n
and
lim ||ar + y \ n n
nco
then
lim (x - y ) = 0. n n
71OO
2.3 A normed space (V, ||-||) is called strictly normed i f the following
condition holds: Whenever x , t / G ^ , x , | / / 0 satisfy
l k + y|| = I W I+ llll
then there exists a > 0 with
x = ay.
Banach spaces
Show that any uniformly convex normed space is strictly normed.

Does the Banach fixed point theorem (Lemma 2.4.1) continue
to hold if we replace (2.4.2) by the condition
\\Gyi - Gy \\ < \\yi - y \\ for all y ,y

2 2 x 2 e AI
3
U and Sobolev spaces
p
3.1 L spaces
In the sequel, instead of functions / : A R U { 0 0 } (A measurable),
we shall consider equivalence classes of functions, where / and g are
equivalent if f(x) = g(x) for almost all x G A. We shall be lax w i t h
the notation, however, not distinguishing between a function and its
equivalence class. The equivalence class of the zero function is called the
null class, and a function in that class is called a null function.
d
D e f i n i t i o n 3 . 1 . 1 . Let A C R be measurable, p G R \ { 0 } .
P
L (A) = {(equivalence classes of) measurable
functions f : A R U {=boc} with
l
\f(x)f C (A)}.
p
For f e L (A), we put
I I / I I P ^ I I / I I L P ( A ) : = ( / J / W I P
^ ) P
- (3-1.1)
The notation suggests that ||-|| is a norm, and we now proceed to

verify this for p > 1. First of all,
| | / | | = 0 & f is a null function.

p (3.1.2)
Thus, ||-|| is positive definite (on the set of equivalence classes). Next,
for c G R,
l|c/|| = |c|||/|| .
p p (3.1.3)
It remains to verify the triangle inequality. This is obvious for p = 1:
||/i + h\\ iL {A) < WIIWLHA) + ll/allt^) 3

( - L 4
)
159
p
160 L and Sobolev spaces
For p > 1, we need
L e m m a 3.1.1 (Holder's inequality). Letp,q > 1 satisfy 1 + ^ = 1,

l
fi L*(J4), / L(A). Then f f
2 L (A), and u 2
||/i/2lli<||/i|| ll/2ll . p f l (3.1.5)

Proof. By homogeneity, we may assume w.l.o.g.
H/iH = l . P I l / 2 | | , = 1. (3-1.6)
Recalling Young's inequality, namely

p q
a b 1 1
ab < + for a,6 > 0 , p,g > 1 , - + - = 1, (3.1.7)
p q p q
we have for x A
/ 1 W / 2 W ^ zp + ~g
hence by our normalization (3.1.6)
/ \fl(x)f2{x)\dx< + l = l = \\h\\ p \\J*\\q

JA P q
q.e.d.
We now obtain the triangle inequality:
L e m m a 3.1.2 (Minkowski's inequality). Let / i , / 2 U>(A), p > 1.

Then
| | / i + / 2 l | < | | / i | | + ll/a|| .
p p p (3.1.8)
Proof. The case p = 1 is given by (3.1.4). We now consider p > 1 and
p
put q ~ j.
- ^ T (so that \1 + 1I = 1). For ${x) ~ \h{x) + / ( a ; ) | \ we 2
have p-1 v q p
^ = l/i+/ | , 2
P
i.e. V Li (A). Since

p
\h(x) + f (x)\ 2 < \Mx)1>(x)\ + \f (x)1>(x)\,
2
we get
||/ +/ || <||/lV'll
1 2
P
1 + ll/2V'll 1
< l l / i l l p l M I , + l l / 2 | | p I I V ' l l ,
by Holder's inequality
= (ii/iii +ii/ |i )n/i+/ ||)

P 2 P 2
p
3.1 L spaces 161
Since p - | = 1, (3.1.8) follows.

q.e.d.
P
We have thus verified that ||-|| is a norm on L (A).
p I n fact, we have:
T h e o r e m 3.1.1 ( R i e s z - F i s c h e r ) . Let A be measurable, p > 1. Then

P
L (A) is a Banach space.
P
Proof. Let ( / ) e N C L (A) be a Cauchy sequence. For every v G N, we
n n
may then find n G N w i t h

v
ll/n - / n j | p < ^7 for all n > n .
This implies that the series

CO
3 L 9
H/mllp + E l l ^ - ^ l l p ( " )
converges. We claim that the series

CO
P
then converges in L (A). Since all elements of the series are nonnega-
+
tive, ( # m ) m N converges to some g : A R U {oo} pointwise in A,
P
and Corollary 1.2.1 implies that (g ) also converges to g in L ( A ) . I n
m
particular, g(x) < oo for almost all x G A. Thus, our original sequence
(3.1.10) is absolutely convergent for almost all x A, towards some /
p
w i t h \f\<g+ |/m|; in particular / G L ( 0 ) . We interrupt the proof to
record:
P
L e m m a 3.1.3. Let (f )neN converge to f in L (A).
n Then some subse
quence converges pointwise almost everywhere to f.
In order to complete the proofs of Lemma 3.1.3 and Theorem 3.1.1,

P
it remains to show that the series (3.1.10) converges to / in L (A).
P
(Then a subsequence of ( / ) converges to / in L (A).
n Since ( / ) was n
P
assumed to be a Cauchy sequence in L (A), the whole sequence has to
P
converge in L (A). I t is in general not true, however, that the whole
sequence also converges pointwise almost everywhere to / . )
This is easy:
CO
/n>(*) + E (/.+(*) " " /(*)
p
converges to 0 almost everywhere in A, and since
2 2
/m(s) + E O W * ) ~ / . ) - /(*) ^ + \fm(x)\,
i/=i
we may apply Lebesgue's Theorem 1.2.3 on dominated convergence to

conclude that we get convergence also w.r.t. ||-||
q.e.d.
2
C o r o l l a r y 3.1.1. L (A) is a Hilbert space with scalar product
(/i,/ ):= /2 fi(x)f (x)dx.

2
JA
Proof. I t follows from Holder's inequality (Lemma 3.1.1) that
|(/l,/2)|<|l/l|l ll/2ll2- 2
2 2
Thus (, ) is finite on L (A) x L (A). A l l the other properties are obvious
or follow from Theorem 3.1.1.
q.e.d.
d
D e f i n i t i o n 3.1.2. Let A C R be measurable, f : A -> R U { 0 0 }
measurable.
ess s u p / ( x ) : = inf {A G R | f(x) < X for almost all x A}

xA
(essential supremum), and
L(A) := {(equivalence classes of) measurable

functions f : A R U { 0 0 } with
ll/lloo : = == esssup | / ( x ) | < 00}
xeA
T h e o r e m 3.1.2. L(A) is a Banach space.
Proof. I f is straightforward to verify that I H I ^ is a norm. I t remains to

show completeness. Thus, let (f )neN be a Cauchy sequence in L. For
n
v G N , we find n G N such that for ra, n > n
||/n /m|loo < *
Thus
|x.4| |/(x)-/ (x)|>ij m

p
3.1 L spaces 163
is a null set for ra, n > n , and so then is
N:= |J | x A | |/(x)-/ (x)|>i;} m
m , n > n
as the countable union of null sets. Since
\fn(x) - f (x)\ m < ~
for ra, n > n and x A\N, f converges uniformly on A \ N towards

n
some / . We simply put f(x) = 0 for x G N. Then
ess sup | / ( x ) - / ( x ) | = ess sup | / ( x ) - / ( x ) | ,

n
xeA xeA\N
since the essential supremum is not

affected by null sets,
J_
and f n converges to / in L(A).

q.e.d.
We also note that Holder's inequality admits the following extension

to the case p = 1, q = oo:
l l
L e m m a 3.1.4. Let f x G L (A), f2 G L(A). Then f f x 2 G L (A), and
< ll/illxll/alloo- (3.1-11)

Proof.
\ \fi{x)f (x))\dx
2 < esssup|/ (x)| / 2 |/i(x)|dx
JA xeA JA
Hl/alUI/ill!-
q.e.d.
d
T h e o r e m 3.1.3. Let A C R be measurable. Let 1 < p < oo, q =
9 P
i.e. ~ + ^ = 1. T/ien L ( A ) is t/ie dual space of L (A). In particular,
P
L (A) is reflexive.
l
Remark 3.1.1. The dual space of L (A) is given by L(A) while the
l l
dual space of L(A) is larger than L (A). Therefore, neither L (A) nor
L(A) is reflexive.
In order to prepare the proof of Theorem 3.1.3, we first derive:

164 LP and Sobolev spaces
d
T h e o r e m 3.1.4 ( C l a r k s o n ) . Let A CR be measurable, 2 < q < oo.
q
Then L (A) is uniformly convex.
Remark 3.1.2. Clarkson's theorem holds more generally for 1 < q < oo.
The proof for 1 < q < 2 is a little more complicated than the one for
2 < q < oo.
The proof of Theorem 3.1.4 is based on:

q
L e m m a 3.1.5. Let 2 < q < oo, f,g <E L {A). Then
q q 1
11/ + g\\ +11/ - \\
q 9 q < 2 * - (H/H; + y i;). (3.1.12)
Proof. For x, 1/ > 0, we have

q q 2 2 a q q
{x + y ) < ( x + t / ) * <2 *r(x + y )i. (3.1.13)
(In order to verify the left inequality in (3.1.13), we may assume w.l.o.g.
2 2 q 2 q 2
x + y = 1. Then x < x , y < y since g < 2, and the desired
inequality easily follows. The right inequality follows for example from
Holder's inequality (Lemma 3.1.1) applied to the following functions
/i,/ :(-U)-R
2
fx = 1,
2
, . / a for - 1 < t < 0
h ( t } 2
~ \ 6 for 0 < t < 1. )
The left hand side of (3.1.13) implies
q 9 2 2
(|a + b\ + \a- 6 | ) ' < (|a + 6| + |a - 6 | ) *
2 2
<v^(a +6 )5 (3.1.14)
for a, 6 R, and by the right-hand-side of (3.1.13), we have

2 2 q 9
V2(a + b )i < 2 ^ (\a\ + | 6 | ) ' . (3.1.15)
Equations (3.1.14) and (3.1.15) imply

9 l 9
|/(:r) + gix)]* + \f(x) - ( r r ) | < 2"~ {\f{x)\" 5 + | (x)| ),
f f (3.1.16)
and (3.1.12) follows by integrating (3.1.16).

q.e.d.
Proof {Theorem 3.14). Let f,g e L"{A) with

= N I = i. 0
3.1 L " spaces 165
By (3.1.12),
q q
\\f+9\\ q + \\f-9\\ <y- q
Therefore, for e > 0, we may find 6 > 0 such that
\\f-9\\ <e g
whenever | | | ( / + g)\\ q > 1 6. This shows uniform convexity.

q.e.d.
Proof (Theorem 3.1.3). We consider the map

p q
i: L (A) -> L (A)
with
i(f)(9) := J f(x)g(x)dx.
By Holder's inequality (Lemma 3.1.1)

||i(/)||= sup |i(/)(?)|<||/|| . p (3.1.17)
Il9ll <l q
q
Thus i(f) is indeed an element of L (A)*. We claim that we have equality
q
in (3.1.17). This means that there exists some g G L w i t h
/ f(x)g(x)dx l l / I L I M L . (3.1.18)
JA
p q q
We put g(x) : = s i g n / ( x ) \f(x)\ 'K Then \g\ = hence g G L (A),
and
/ f(x)g(x)dx\ = / |/(x)^(a
(x)\ dx
p
= / j / ( * ) ii dx ?
p p
= ( j \f{x)\ d y
y A X (j \f(x)\ d y
A X
= H/llplMI,-
This verifies (3.1.18), hence equality in (3.1.17). Equality in (3.1.17)
implies that i is an isometry, in particular injective. I n order to complete
the proof we need to show that i is surjective. Suppose on the contrary
that
L"(A)* \ i{W{A)) / 0.
p
P p
Since L (A) is complete and i is continuous, i(L (A)) is complete, hence
closed. By the Hahn-Banach theorem (Corollary 2.1.1), there then exists
veL<*(A)**,v^0, with
v\i(LP(A)) = 0.
We now suppose for a moment that 1 < p < 2. Then 2 < q < oo, and
q
L (A) is reflexive by Theorems 3.1.4 and 2.2.3. We may therefore find a
q
g in L {A) with
q
F(g) = v(F) for all F G L (A)*.
P
We then have for any <p G L ( A )
0 = v(i(<p)) = i(<p){g) = ^ <p(x)g(x)dx,
hence # = 0 (by a reasoning as in the derivation of (3.1.18)), hence also

v = 0, a contradiction. We have shown that i furnishes an isomorphism
P q q q
between L (A) and L (A)*. Since L (A) is reflexive, so is L (A)* by
P P
Corollary 2.2.2, hence L (A). I n conclusion, L (A) has to be reflexive
q
for any 1 < p < oo, and its dual space is given by L (A).
q.e.d.
p
3.2 Approximation of L functions by smooth functions
(mollification)
p
In this section, we shall smooth out L functions by integrating them
against smooth kernels. As these kernels approach the Dirac distribution,
these regularizations will tend towards the original function. For that
d
purpose, we need some g G C o ( R ) f with
d
g(x) > 0 for all x G R (3.2.1)
Q{X) = 0 for |x| > 1 (3.2.2)
/ g(x)dx[= I g(x)dx)=l. (3.2.3)

d
JR \ JB{0,1) j
Such a g is called a Friedrichs mollifier. I n this , ft will always denote an

d d
open subset of R . Let / G L^ft). We extend / to all of R by putting
d
t For Q C R open, Cg(Q) is the space of all C functions <p on O for which the
closure of {x E Q \ ip(x) ^ 0 } , the support of ip (supp<p), is a compact subset of
O. Elements of C^(Q) are often called test functions.
3.2 Approximation of LP functions by smooth functions 167
d
f{x) = 0 for x R \ fi. Let h > 0.
fh is called the mollification of / w i t h parameter h. I n order to appre

ciate this definition, we first observe
d
supp Q C B(y, h) := {z R \ \z-y\< h}> (3.2.5)
where Q (^7^) is considered as a function of x, and
For these reasons, one expects that fh tends towards / as h tends to 0.

I t remains to clarify the type of convergence, however. The advantage
of approximating / by fh comes from:
f
L e m m a 3 . 2 . 1 . Let Q C C fif> h < d i s t ( f y , d f i ) . Then
fh C(fi').
Proof By Corollary 1.2.2, we may differentiate w.r.t. x under the inte

gral sign in (3.2.4), and since Q C so then is fh.
q.e.d.
We now start investigating the convergence of fh towards / .
L e m m a 3.2.2. If f C ( f i ) , then for each ft' C C ft, fh converges

uniformly to f onW as h 0. In symbols: fh^f on Q' as h 0.
Proof. We have
f{x) = f g(w)f(x)dw by (3.2.3) (3.2.7)

J\w\<l
and
fh(x) = j Q{w)f(x - hw)dw (3.2.8)

J\w\<l
by using the substitution w = in (3.2.4). For fi' C C ft and h <
f
f 'ft' CC H ' means that the closure of Q is compact and contained in fi. We say
that Q' is relatively compact in Q.
\ d i s t ( f i ' , d f i ) , we then have
sup - h(x)\ < sup / g(w)\f(x) - f(x - hw)\dw

xQ' xefl' J\ \<l
t\w\< w
< sup - f{x - hw)\

M 0 as h -+ 0,

xQ'
i.e. uniform convergence.

q.e.d.
P
T h e o r e m 3 . 2 . 1 . Let f G L (Q), 1 < p < oo. Then fh converges to f
p
in L (Q) as h -+ 0.
P
Proof. We have for g G L (Q)
p
/ \g (x)\ dx
h
= I I g(w)g(x hw)dwdx
JQ J\W\<1
p
< / I / g(w)dw J I / g(w) \g(x - hw)\ dw J .
JQ \J\W\<1 J \J\U)\<1 J
= / Q(W) / \g(x - p
hw)\ dxdw,
J\w\<i Jn
using (3.2.3) and Fubini's theorem,

p
= / Q(w) / \g{y)\ dydw
d
J\w\<l JR
p
= [ \g(y)\ dy,
J n
using (3.2.3) again.
Thus
3 2 9
H^llLp(n)<IMlLp(n)- ( - - )
3.2 Approximation of LP functions by smooth functions 169
D
Let e > 0. By Theorem 1.1.4, (6), we may find 0,
Ik -WIIILP(R*) ^ |- (3.2.11)
Applying (3.2.9) to / - </?, we obtain
ll/n - </?n|| p(Rd) < L 11/ - ÎILP(M^) ' (3.2.12)
(3.2.10)-(3.2.12) yield
11/ ~ M I L P ( Q ) < H / - /nllLP(R-) < ' E 3

( ' ' 2 1 3
)
g.e.d.
p
Corollary 3.2.1. For 1 0. We may then find Q' C C 0 w i t h
<
H/llLp(n\n') 2*
We put / ' : = / X L P ( O ' ) - T h e n
/ 3 2 1 4
H/-/ llLp n)<5- ( ( - - )
By Theorem 3.2.1, for sufficiently small h,
f f
\\f -f h\\L (n)<~- P (3.2.15)
By (3.2.13), (3.2.14)
11/ ~" / ^ I I L P ( Q ) <

2*
Since G C^(Q) for /i < dist(fi',dfi), the claim follows.

q.e.d.
p P
Corollary 3.2.2. L (fl) is separable for 1 < p < oo. Every f G L ( Q )
con be approximated by piecewise constant functions.
Proof. By Corollary 3.2.1, i t suffices to find a countable subset BQ of

P
L (Q) with the property that for every 0,
there exists some a G BQ w i t h
< 3 2 1 6
llv-alli-(n) ( - - )
d
Let B the set of all functions a on R of the following form: There exist
some fc, N G N and rational numbers c * i , . . . , a* and cubes Qi,..., Qk G
d
M w i t h corners having all their coordinates in -^Z and of edge length
fa such that
for x Qi
otherwise.
Clearly, B is countable. Since a continuous function (p w i t h compact
support is uniformly continuous, we may easily find some a G B w i t h
a A E
Il ~ ÎILP(Q) ^ H - <P\\LP(**) < - (3.2.17)
We put BQ := {axn | a G B } . B Q is likewise countable, and from

P
(3.2.15), (3.2.16), we conclude that BQ is dense in L (Q).
q.e.d.
p
Remark 3.2.1. The separability of L (ft) can also be seen by using Corol
lary 3.2.1 and the Weierstrass approximation theorem that allows the
approximation of continuous function w i t h compact support by polyno
mials w i t h rational coefficients.
The preceding results do not hold for L(fi). Namely, i f a sequence

of continuous functions converges w.r.t. I H I X ^ Q ) , then it converges uni
formly, and therefore, the limit is again continuous. Therefore, noncon-
oc
tinuous elements of L (ft) cannot be approximated by continuous func
tions in the L-norm. Also, L(fi) is not separable. To see this, let
(a )nN be any subsequence of { 0 , 1 } , i.e. a G { 0 , 1 } for all n . To ( a ) ,
n n n
we associate the function / ( ) on ( 0 , 1 ) defined by

t t n
for < x < 2F=r i f a = 1 k
/(an) { 00 for k G N.
for < x < 2FTT i f a k = 0
Then for any two different sequences ( a ) , ( 6 ) , n n
= L
||/(a ) -/(M|| ~((0,1))
w L
Since the set of subsequences of { 0 , 1 } is uncountable, this implies that

L ( ( 0 , 1 ) ) is not separable. Of course, a similar construction is possible
d
for f2 any open subset of R .
We finally note:
2
L e m m a 3.2.3. Let f G L (Q), and suppose that for all ip G C(Q)
j f(x)(p(x)dx = 0.
Then f = 0.
3.3 Sobolev spaces 1 1 1
2
Proof. Since Co(fi) is dense in L ( f i ) , and since
9*-> I f(x)g(x)dx
JQ
2
is a continuous linear functional on L ( f i ) , we obtain that
/IQ
JQ
f(x)g(x)dx = 0 for all g G L ( Q ) . 2
Putting g = / yields the result.

q.e.d.
3.3 Sobolev spaces

p
In this section, we wish to introduce certain extensions of the L spaces,
the so-called Sobolev spaces. They will play a fundamental role in subse
quent chapters because they constitute function spaces that are complete
w.r.t. norms naturally occurring in variational problems. I n this section,
d
Q will always denote an open subset of R . We shall use the following
notation: For a d-tuple a : = ( a i , . . . ,a^) of nonnegative integers,
|| : = | > ,A:=(^r) ( ^ )
1
Definition 3.3.1. Let u,v G L (Q). Then v is said to be the oc-th weak
derivative ofu, v := D u,a if
j (pvdx = (-l)W j uD ipdx

a (3.3.1)
for every <p G CQ*'.

We can now define, for k G N and 1 < p < oo, the Sobolev space
k P
W *(Q) := {u G L (Q) | D u exists and lies in a
p
L (Q) for all\a\ < k } ,
jQ
\\<*\<k
k k k,p
Finally, let H *(n) andH *(Q) be the closures of C(fi) f l W (l)
k k
and C$ n W *(Q), respectively in W *(Q).
p
We shall use the following abbreviations for u G 1 < i < d.

D{U is the weak derivative for the multiindex ( 0 , . . . , 0 , 1 , 0 , . . . , 0), 1 at
th
the 2 position, and Du is the vector ( D i u , . . . , Dû) of all first weak
derivatives.
The following result is obvious.
k
L e m m a 3 . 3 . 1 . Let u G C (Q), and suppose all derivatives ofu of order
p k,p
< k are in L ( f i ) . Then u G W (Q), and the weak derivatives are given
by the ordinary derivatives.
q.e.d.
k p
Thus, the W > spaces constitute a generalization of the spaces of
k,p
k times differentiable functions. The W norm is considerably weaker
fc kyP k
than the C -norm, and so the W spaces are larger than the C spaces.
Before investigating the properties of these spaces, it should be useful
to consider an example: Let fi = ( - 1 , 1 ) C E, u(x) : = We claim that
l p
u G W ' (Q) for 1 < p < oo. I n order to see this it suffices that the first
weak derivative of u is given by
for 0 < x < 1

for - 1 < x < 0.
Indeed, we have for (p G CQ(( i.i))
2,p
We claim, however, that u is not contained in W (ft). Namely i f w(x)
were the second weak derivative of u, i t would have to be the first weak
derivative of and consequently, we would have w(x) = 0 for x ^ 0.
The rule for integration by parts (3.3.1) would then require that for all
V>C3((-1,1))
0 =
= 2^(0)
which is not the case. Thus, v does not have a first weak derivative.
3.3 Sobolev spaces 173
Remark 3.3.1. Some readers may have encountered the notion of a dis
tributional derivative. I t is important to distinguish between weak and
1
distributional derivatives. Any L ( f i ) function possesses distributional
derivatives of any order, but as the preceding example shows, not nec
essarily weak derivatives. I n the example, of course, the second distri
butional derivative of u is 2<5o, where <5o is the Dirac delta distribution
at 0. u does not possess a second weak derivative because the delta
1
distribution cannot be represented by an L function.
kyP
T h e o r e m 3.3.1. The Sobolev spaces W (ft) are separable Banach
spaces w.r.t. ||-|liyfc,P(n)-
18 a
Proof. That |Hlw*.p(n) norm follows from the fact that I H I X ^ Q ) is
a norm (see section 3.1). Similarly, we shall now derive completeness of
k p
W *(ft) from the completeness of the L (ft) spaces (Theorem 3.1.1).
k,p
Thus, let ( v ) n N C W (ft)
n be a Cauchy sequence w.r.t. ||-||w*.p(n)-
This implies that (D u ) ^ is a Cauchy sequence w.r.t. I H I ^ Q ) for
a n ne
p
all |Q| < k. By Theorem 3.1.1, (D u ) therefore converges in L (ft) a n
a]
towards some v . For a<peCl (n)
(3.3.2)
p
Therefore, v is the a - t h weak derivative of t>o, the L - l i m i t of ( u ) N ,
a n n
k,p
and consequently vo G W (ft). The separability again follows from the
p
corresponding property for L (ft) (Corollary 3.2.2).
q.e.d.
k k
T h e o r e m 3.3.2. W *(fl) = H *(ft).
kyP
This result says that elements of W (ft) can be approximated by
n
C(Q) functions w.r.t. IHIivfc.p(Q)' I general, however, for k > 1 one
p k p k
has # o ' ( Q ) ^ W ' (ft) so that W *(ft) functions cannot be approxi
p
mated by C g ( f i ) functions, in contrast to L (ft) functions where this
is possible (Corollary 3.2.1). This is seen from the following simple ex
ample:
ft = (1,1) C K, u(x) = 1. I f (<p )neN C C^ (ft) converges to u in n
3
1
L ( f i ) , then after selection of a subsequence, i t converges pointwise al
most everywhere (Lemma 3.1.3), and therefore, for sufficiently large n ,
there exists x n G ( - 1 , 1 ) w i t h ip (x ) n n > \. Since (l)>
n n
this implies that

f p
Therefore, <p cannot converge to v! = 0 in L ( ( - 1 , 1 ) ) , and therefore
n
1,p
(p cannot converge to u in W ( ( 1 , 1 ) ) .
n
k p
Proof (Theorem 3.3.2). We have to show that any u E W < (fl) can
be approximated by C(fi) functions. As in 3.2, we extend u to be 0
outside f i and consider the mollifications UH C(fi). We compute
u d u s i n
D (u (x))
a h = Da Q (^J^jiX ' (y) V ( g Corollary 1.2.2)
where D , A X is the derivative w.r.t. x ,
by definition of D u a
= (D u) (x). a h (3.3.3)
Thus, the derivative of the mollification is the mollification of the deriva

tive. Since D u E L (ft)
a by Theorem 3.2.1, (D u)h
p
1 converges to D u a a
in L (Q) for h 0. By (3.3.4), we conclude that D (uh)

P
converges to a
D u in LP(Q)> for all | a | < fc, and this means that Uh converges to u in
a
k
w *(n).
q.e.d.
Theorem 3.3.3. W * (Q) k p

is reflexive for k e N , 1 < p < oo.
kyP
Proof. I t follows from Theorem 3.1.3 that the dual space of W (ft) is
k,q
given by W (fl), w i t h ~ + ~ = 1. This implies reflexivity.
g.e.d.
Theorem 3.3.4. HQ (Q) ,p

is closed under weak convergence in K,P
W (Q).
,p
Proof. This follows from Lemma 2.2.5, since HQ (Q) by its definition is
k,p
a closed subspace (w.r.t. strong convergence) of W (fl).
q.e.d.
k,p
T h e o r e m 3.3.5. For 1 < p < oo, k N , any sequence in W (Q) that
is bounded w.r.t. IHIjy*.p(n) contains a weakly convergent subsequence.
3.4 Rellich's theorem, Poincare and Sobolev inequalities 175
k p
Proof. By Theorems 3.3.1 and 3.3.3, W ' (ft) is separable and reflexive.
Therefore, the result follows from Corollary 2.2.1.
q.e.d.
3.4 Rellich's theorem and the Poincare and Sobolev

inequalities
The compactness theorem of Rellich is:
d
T h e o r e m 3 . 4 . 1 . Let ft C R be open and bounded. Let (w )nN C n
p
HQ' (Q) be bounded, i.e. \\u \\ i, ^ < c (independent of n). Then a
n W P
p
subsequence of (u )neN converges in L (ft).
n
Remark 3.4-1-
Rellich originally proved the theorem for p = 2. Kon-
drachev proved the stronger result that some subsequence converges in
L(fi) for 1 < q < ^ if p < d and for 1 < q < oo if p > d. Of
course, these exponents come from the Sobolev Embedding Theorem
(see (3.4.12)). See Corollary 3.4.1 below.
p
Proof. Since u n G HQ' (Q), for every n G N and e > 0, there exists some
vn G CQ (ft) w i t h
t;
r n ~ n||vî.p(n) < o (3.4.1)
Therefore
bn||wn,p(n) < ' c

( c
+ (3.4.2)
We consider the mollification
v x
n,h( ) = Jfi J^Q (~/~) v (y)dy
n
of v n and estimate
\v (x)
n - V (x)\
nyh
L \w\<\
g(w)(v (x) n
h w
v (x
n hw))dw by (3.2.7), (3.2.8)
r r\\
w . (3.4.3)
< / Q{w) / v (xn - n?) drdw with
J\w\<l JO
\dr M
This implies
p
/ \v (x) n - v , (x)\ dx
n h
JQ
( f
f f h M
\ d Y
< / I / g(w) / v (x-rd) n drdw] dx
JQ \J\W\<I JO \ur J
w
f ( f f i\ i f^\ \ I Q
= 1 6 w Vn X
drdw dx
Jn\J\ |<i V ^ ^ " " / ^^ J \dr ^
p p p
< ( j g{w)dw] ( [ g{w)h \w\ f \Dv (x)\ dxdw]n ,
VM<i / vM<i J j
using Holder's inequality, Fubini's theorem and the notation
Dv n
Vn
dxi "-"dx<* J' Vn
Since J ^ < 1 l g{w)dw = 1 (by (3.2.3)), we obtain
I K - nM\ P(Q)
v
L ^ H
11 D v
n 11 LP (Q)
< he' by (3.4.2)
< - if h is sufficiently small. (3.4.4)
Next,
X c
KM )\ ^ ^ o l K I I L i ( n )
with CQ : = sup^ g(z) by definition of v ,h n
- ^co(measn) p |KH p L ( n ) (3.4.5)
by Holder's inequality,
and similarly
c m e a s n lp
X
^ Jî i( y \K\\ (3.4.6)
dx' Vn,h( ) LP{Q)
with Ci : = sup 2 From (3.4.2), (3.4.5), (3.4.6), we see that for

fixed h > 0,
K , h | | i ( ) < constant
C Q
(3.4.7)
(where the constant depends on h). Therefore, (v ^)neN contains a uni n
formly convergent subsequence by the Arzela-Ascoli theorem. Since uni-

3.4 Rdlich's theorem, Poincare and Sobolev inequalities 111
p
form convergence implies L -convergence (e.g. by Theorem 1.2.3), the
p
closure of v h is compact in L ( f i ) . Since a compact subset of a metric
Uy
space (e.g. a Banach space) is totally bounded, there exist finitely many
p
wi,... ,U)N L ( f i ) such that for every n N there exists 1 < j < N
with
v 3 4 8
\\ n,h - W j | | LP( n ) < I' (--)
By (3.4.1), (3.4.4), (3.4.8), for every n N we find 1 < j < N w i t h
\\u n ^jH^p(Q) < c.
p
Thus, ( w ) n N is totally bounded in L ( f i ) . Therefore, the closure of
n
p
(u ) w
n n in L ( f i ) is compact (again, a general result for metric spaces),
p
and i t thus contains a convergent subsequence in L ( f i ) .
q.e.d.
We now come to the Poincare inequality:
d p
T h e o r e m 3.4.2. Let fi C R be open and bounded. For any u e H^ (ft)
i
E d
IHIz, n)<( ^) |l^llLp(n)
P( (3-4-9)
d
where u?d is the Lebesgue measure of the unit ball in R .
p
Proof. Since CQ(Q) is dense in iJQ' (fi), we may assume u e C(j(fi). We
d d
put u(x) = 0 for all x e E \ S l . For $ e R with = 1, we have
f d
ulx) = u(x 4- rd)dr.
Jo dr
Integration w.r.t. d yields
1 f t d
/ / u(x + rfodddr
duj Jo J d w = 1 dr
I f 1 , .
p
Therefore
(J\u{x)\>dxY
p-1
I D u i v ) r d y d y dx
k ( i (L J T ^ F ) [L ^ )
i /
p 1
|Du(y)| d) ( / - ~dx\ 1
(3.4.10)
du, \ J
d 1
W \x - V "
N
using Fubini's theorem to exchange the order of integration in the first

factor.
In order to control
L \ x -y\
y\ - d l d X
'
we choose R with
d
meas ft = meas B(y, R) = UdR
d
(B(y,R):={zeR \ \z\ < R}).
Since
J ^ ^ - J F * **\*-v\>R
^ T ^ - ^ ior\x-y\<R,
we have
j d-i^ x
- / d~i^ x
y\ JB(y,R) \x - y\
JQ \x - y\ JB(y R)
y \x - y\
= dw R d (3.4.11)
1 1
= dw d (meas ft) .
Equations (3.4.10) and (3.4.11) yield (3.4.9).

q.e.d.
3.4 Rellich's theorem, Poincare and Sobolev inequalities 179
We now come to somewhat stronger results that will however only be

needed in Chapter 9. Namely, we have the Sobolev inequalities.
P
T h e o r e m 3.4.3. Let u G H^ (Q).
d
(i) If p < d, then u G L ~^p(Q), and
\\u\\^<c\\Du\\ . p (3.4.12)
(ii) Ifp>d, then u G C(fi), and
sup|u| < c(measfi)^~p \\Du\\ (3.4.13)
with constants c depending only on p and d. (Actually, by a Theo

rem of Morrey, forp > d, u G HQ (Q) is even Holder continuous ,P
with exponent 1 ^ . )
We only prove (i) as (ii) will not be used in the present book:
Proof. We first assume u G CQ(Q). Since u has compact support, we

have
oo
fort = l,2,...,d.
/ -OO
Multiplying these inequalities for i; = 1 , . . . , d yields
d
a ( f
H y ) \ ^ < Ml / IAti(W
Using Holder's inequalityf, we compute
oo d
1
\u(x)\*=* dx
/ -oo
< ( r \DMV)\dy ) 1 7=1

r (n r d y i
) d x i
\J oo / J oo \ J oo J
1
1 d
1
< (y iDMy^dy ^ (jlf |A()|dy'dx^ .
f More precisely, one uses Exercise (2) below with pi = = Pd-i = d 1.

p
2 d
Iteratively also integrating w.r.t d x , . . . , dx finally yields
i
d
Hi A L ( f i ) ^(n/ i^)i^ n
1
jlPILi(n)- (3-4-14)
This is (3.4.12) for p = 1. The case of general p may now be obtained by

applying (3.4.14) to for suitable fi > 1 and using Holder's inequality.
Namely, from (3.4.14) for in place of u
M
H I I <gjf Nx)rMiM*)i<fc
lA
<^||H-
by Holder's inequality.
p
For p < d, we may take = ^7}^ and obtain
which yields (3.4.12), since =

q.e.d.
As a consequence, we obtain the theorem of Kondrachev:

d
C o r o l l a r y 3 . 4 . 1 . Let ft G R be open and bounded. Let (u ) n C n ne
P
HQ (ft) be bounded for some 1 p as otherwise the result is an easy
consequence of Holder's inequality since ft is bounded. We denote this
converging subsequence again by (u ). n
From Holder's inequality, we obtain

u w u u u
\\ n ~ m||,a(Q\ ^ \\ n ~ ^rnWîtQ) \\ n ~ m\\ dp
if fi satisfies - = - f (1 /z) ( -
q \p d
M
< c\\u n - tz ||i
m ( n ) \ \D(u n - Wm)|lip n)
( (3.4.15)
by Theorem 3.4.3 (i).
Exercises 181
P
Since Du is bounded in L (Q) by assumption, and (u ) is a Cauchy
n n
p
sequence in L ( f i ) , hence also in (3.4.15) then implies the Cauchy
q
property i n L (ft).
q.e.d.
Exercises
3.1 Let
d
Ai :== {x G R | > 1} , A : = {x E 2
<i}.
and consider
x
f{x) = \\x\\ for A E R .
P P
For which values of d,p, A is / G L ( A ) , or / G ( A ) ? X 2
d =
3.2 Let A C R be measurable. Let pi,... ,p/t > 1, ]Ci=i jr 1>
X
/< G LP* (A) for * = 1 , . . . , k. Show / i ... f k G L ( A ) , with
^nii/'iu-
i <=i
d
3.3 Let A C R be measurable, meas A < oo, 1 < p < g < oc. Then
q P
L (A) C L ( A ) , and for / G L(i4)
" 7 7 T II./ I I L P ( A ) - < , -vi II-/ I I L ( A ) '

(meas A) p (meas A)
(Hint: Apply Holder's inequality w i t h / x = 1, / 2 = /)

3.4 d
Let A C R be measurable, 1 < p < ? < r, 5 = + i i
7 i
,
p r
/ G L ( A ) f l L ( A ) . Then / G L(A), and
l-Q
L(A) ^ LP(A) L"(A) '
d
3.5 Let A C R be measurable, meas A < oo, / : A - R U { 0 0 }
measurable. Then
lim T I I / I I LP(A) Loo (A)

p->oo ( m e a s ^4) p
(where we allow these quantities to be infinite).

P
d P
3.6 Let A C R be measurable, ( / ) N C L ( A ) w i t h

ll/nllp < constant.
Suppose f n converges pointwise almost everywhere on A to some

P
/ . Is / G L (A), and do we necessarily get
||/-/|| ^0 p asnôo?
3.7 Let A\,A f2l be as in exercise 1). For which d,fc,p, A is / in

k k
W *(Ai) or in W *(A )l 2
2
3.8 Consider the sequence ( s i n ( n x ) ) N in L ( ( 0 , 1 ) ) . Does i t con
ne
2
verge in the L -norm? Does i t converge weakly? I f so, what is
the limit?
4
The direct methods in the calculus of
variations
4.1 Description of the problem a n d its solution

The typical problem of the calculus of variations is to minimize an inte
gral of the form
d
where fi is some open subset of R (in most cases, fi is bounded), among
functions
u : fi -+ R
belonging to some suitable class of functions and satisfying a boundary

condition, for example a Dirichlet boundary condition
u(y) = g(y) for ye on

for some given g : d f i R. Thus, the problem is
F(u) min for u G C ,
where C is some space of functions. The strategy of the direct method

is very simple: Take a minimizing sequence (u ) n C C, i.e.
n ne
lim F(u )n = inf F(u)

n*oo uC
and show that some subsequence of (u ) converges to a minimizer u G C.

n
To make this strategy be successful, several conditions should be met:
(1) Some compactness condition has to hold so that a minimizing

sequence contains a convergent subsequence. This requires the
careful selection of a suitable topology on C.
183
184 Direct methods
(2) The limit u of such a subsequence should be contained in C. This

is a closedness condition on C.
In particular, for (1) and (2) to hold, C should not be too restric
tive. I n other words, one should not specify too many properties
for a solution u in advance.
(3) Some lower semicontinuity condition of the form
F(u) < l i m i n f F(u ) n if un converges to u

n+oo
has to hold, in order to ensure that the limit of a minimizing
sequence is indeed a minimizer for F.
The lower semicontinuity condition becomes easier i f the topology of

C is more restrictive, because the stronger the convergence of u to u n
is, the easier that condition is satisfied. That is at variance, however,

with the requirement of (1) since for too strong a topology, sequences
do not always contain convergent subsequences. Therefore, we expect
that the topology for C has to be carefully chosen so as to balance these
various requirements. I n order to gain some insights into this aspect, i t
is useful to approach the problem from an abstract point of view. Thus,
we shall return to the concrete integral variational problem raised in the
beginning only later.
4.2 L o w e r s e m i c o n t i n u i t y
We say that a topological space X satisfies the first axiom of countability,
if the neighbourhood system of each point x X has a countable base,
i.e. there exists a sequence (t/ ) eN f open subsets of X w i t h x G U
I/ 1/ v
with the property that for every open set U C X with x G U there exists
n G N with
V cV.n
X satisfies the second axiom of countability i f its topology has a count

able base, i.e. there exists a family {U ) ^n of open subsets of X w i t h
u y
the property that for every open subset V of X , there exists n G N with
U CV.n
We note that separable metric spaces X satisfy the second axiom of

countability. I n fact, let be a dense subset of X, and let ( r ^ ) ^ ^
+
be dense in 1R . Then
{ 7 ( x , r ) := {x G X : d{x,x )
M v < r^}
4-2 Lower semicontinuity 185
) the distance function of X) forms a countable base for the topol

ogy.
If the first countability axiom is satisfied, topological notions usually
admit sequential characterizations. For example, i f ( # ) n N C X is a se
n
quence in a topological space X satisfying the first axiom of countability,

then any accumulation point of (x ) (i.e. any x X w i t h the property
n
that for every neighbourhood U of x and any m G N , there exists n > m

w i t h x G U) can be obtained as the limit of some subsequence of (x ).
n n
Although we shall often employ weak topologies which typically do not

satisfy the first axiom of countability, for our purposes it will usually be
sufficient to use sequential versions of topological properties. For that
reason, we shall define our topological notions in sequential terms, with
out adding the word 'sequentially'.
Definition 4.2.1. Let X be a topological space. A function F : X >

R : = R U { 0 0 } is called lower semicontinuous (Isc) at x if
F(x) < liminf F(x )n
n00
for any sequence ( x ) n N C X converging to x. F is called lower semi-

n
continuous if it is Isc at every x G X.
The following properties are immediate:
L e m m a 4.2.1.
(i) If F :X -^Ris Isc, X > 0, then XF is Isc.

(ii) / / F, G : X > R are Isc, and if their sum F - f G is well defined
(i.e. there is no x G X for which one of the values F(x),G(x) is
-hoo and the other one is 00), then F + G is also Isc.
(iii) For F, G : X -+ R Isc, inf (F, G) is also Isc.
(iv) / / (Fi)ii is a family of Isc functions, then s u p Fi is also Isc.
i /
Examples.
(1) Any continuous function is lower semicontinuous.

(2) I f X satisfies the first axiom of countability, then A C X is open
if and only if its characteristic function \ A is Isc.
Definition 4.2.2.
(i) Let X be a normed space, with norm ||-||. F : X R is weakly

proper, if for every sequence ( x ) N C X with \\x \\ 00 we
n n n
have F(x ) oo for n oo.

n
186 Direct methods
(ii) Let X be a topological space. F : X R is coercive if every

sequence (x ) C X with F(x ) < constant (independent of n)
n n
has an accumulation point.
We now formulate the following general existence theorem for mini

mizers:
T h e o r e m 4 . 2 . 1 . Let X be a separable reflexive Banach space, F : X

R weakly proper and lower semicontinuous w.r.t. weak convergence. Then
there exists a minimizer Xo for F, i.e.
F(x ) 0 = inf F(x) (> -oo).
Proof. Let ( x ) N be a minimizing sequence for F , i.e.

n n
lim F(x )n = inf F(x).

noo xX
Since F is weakly proper, | | x | | is bounded. Since X is reflexive, after

n
selection of a subsequence, x converges weakly to some x G X by

n 0
Corollary 2.2.1. By lower semicontinuity of F ,
F ( x ) < l i m F ( x ) = inf F ( x ) ,
0 n
noo xX
and since xo G X , we must have in fact equality. Also, since F assumes

only finite values by assumption, this implies that
inf F(x) > - o o .

v y
xex
q.e.d.
Remark 4-2.1. The argument of the preceding proof also shows that in a
separable reflexive Banach space, a weakly proper functional is coercive
w.r.t. the weak topology.
Lower semicontinuity w.r.t. weak convergence is a rather strong prop

erty, in fact much stronger than lower semicontinuity w.r.t. to the Ba
nach space topology of X. Fortunately, there exists a general class of
functionals, namely the convex ones for which the latter property im
plies the former.
D e f i n i t i o n 4.2.3. Let V be a convex subset of a vector space; F : V >

R is called convex if for any x, y EV, 0 < < 1 ,
F{tx + (1 - t)y) < tF{x) + (1 - t)F(y)
(convexity of V means that tx + (1 - t)y G V whenever x,y G V, 0 <

t < 1).
4-3 Existence ofminimizers 187
L e m m a 4.2.2. Let V be a convex subset of a separable reflexive Banach

space, F : V R convex and lower semicontinuous. Then F is also
lower semicontinuous w.r.t. weak convergence.
Proof. Let ( x ) n N C V converge weakly to x G V. We may assume that

n
F(x )
n converges to some K G R. By Theorem 2.2.4, for every m G N and
every e > 0, we may find a convex combination
N N
Vm '= ^ ^ X x n n (X > 0, ^ ^ A = 1)
n n
nm n=m
with
\\Vm ~ X\\ <
Since F is convex,
N
X F
F(y ) m < n (*n)- (4.2.1)
n=m
Given e > 0, we choose m = m(e) G N so large that for all n > m,
F(x )n < K+ e.
Letting e tend to 0, we get from (4.2.1)
l i m s u p F ( ? / ) < K. m
raoo
Since F is lower semicontinuous
F(x) < l i m i n f F(y ) m < limsupF(i/ ) < K = limF(x ). m n
moo m-ôo
This shows weak lower semicontinuity of F .

q.e.d.
4.3 T h e existence of minimizers for convex variational

problems
We return to the concrete variational problem discussed in Section 4.1
and begin with:
d d
L e m m a 4.3.1. Let Q C R be open, f : fl x R R, with f(-,v)
d
measurable for all v G M. , /(#, ) continuous for all x eft, and
p
f(x,v) > -a(x) + b\v\
188 Direct methods
d 1
for almost all x G ft, and all v G R , with a G L (ft), b G R, p > 1.
$(v) := / f(x,v(x))dx
JQ
p p
25 o /ower semicontinuous functional on L (ft), $ : L ( f i ) > R U {oo}.
Proof. Since / is continuous in v, f(x,v(x)) is a measurable function,

p
and so $ is well-defined on L ( f i ) , by Theorem 1.1.2. Suppose (v )nN n
p
converges to v in L (ft). Then a subsequence converges pointwise almost
everywhere to v by Lemma 3.1.3. We shall denote this subsequence again
by (t; ), noting that the subsequent arguments may also be applied to
n
any remaining subsequence. Since / is continuous in v (actually, it would

suffice to have / lower semicontinuous in v), we have
p p
f(x,v(x)) - b\v(x)\ < liminf ( / ( x , v ( x ) ) n -b\v (x)\ ).
n
noo
Because of the lower bound
p
f(x,v (x))
n -b\v (x)\
n > -a(x)
x
with a G L (ft), we may apply the Theorem 1.2.2 of Fatou to conclude
p
(f(x,v(x)) b\v(x)\ )dx < l i m i n f / {f(x,v (x)) n - p
b\v (x)\ )dx.
n
p
Since v converges to v in L ( f i ) ,
n
p p
f b\v(x)\ dx lim / b\v (x)\ dx,
n
JQ JQ
and we conclude lower semicontinuity, namely
f(x,v(x))dx < liminf / f(x,v (x))dx.

n
q.e.d.
L e m m a 4.3.2. Under the assumptions of Lemma 4-3.1, assume that

d
/(#,-) is a convex function on R for every x G ft. Then $(v) :=
p
J / ( # , v(x))dx defines a convex functional on
Q L (ft).
4-3 Existence of minimizers 189
Proof Let v,w G LP (ft), 0 < t < 1. Then
*(*v + ( l = / /(x,*v(x) + ( l - * ) ^ ( x ) ) d x
< / {*/(x,t;(x)) + ( l - t ) / ( ^ ^ ) ) } d x
by the convexity of /
= **(v) + ( l - * ) * ( w ) .
g.e.d.
We may now obtain a general existence result for the minimizer of a

convex variational problem.
d d
T h e o r e m 4 . 3 . 1 . Let fl C R be open, and suppose f : fl x R E
d
(i) / ( , v) fcs measurable for all v G E .
(ii) / ( # , ) 25 convex for all x Efl.
p d
(iii) f(x,v) > a(x) 4- 6|t;| / o r almost all x fl, all v R , with
x
aL (fl), 6 > 0 , p > 1.
1>p P
Le* # G i f (f2), and /e* A : = ^ + #o' (fi)-
F(u) := / f(x,Du(x))dx
Jn
assumes its infimum on A, i.e. there exists uo G A with
F(u )
0 = inf F(u).
liP
Proof. By Lemma 4.3.1, F is lower semicontinuous w.r.t. H (fl) con-
vergencef, and by Lemma 4.2.2, F then is also lower semicontinuous
1,p 1,p
w.r.t. weak H (fl) convergence, since H (fl) is separable and reflex
ive for p > 1 (see Theorems 3.3.1 and 3.3.3). Let ( u ) n e N be a minimizing n
sequence in A, i.e.
lim F(u ) n = inf F(u).
noo uA
Since
p
/ \Du \ n < \F(u )+ n \ [ a(x)dx,
Jn o b J Q
p p
(Du ) n
n ne is bounded in L ( f i ) , hence ( w ) n N C g+H^ (fl) is bounded n
1 p
in H *^) by the Poincare inequality (see Theorem 3.4.2). Since H^ (fl)
d
f Note that convex functions on R are continuous.
190 Direct methods
is a separable reflexive Banach space, by Theorem 3.3.5, after selec

tion of a subsequence, (w )nN converges weakly to some UQ A (A
n
is closed under weak convergence, Theorem 3.3.4). Since F is convex

by Lemma 4.3.2 and lower semicontinuous by Lemma 4.3.1, i t is also
liP
lower semicontinuous w.r.t. weak H (Q) convergence by Lemma 4.4.2.
Therefore
F(u ) 0 < l i m F(u ) n = inf F ( u ) ,

noo uA
and since UQ A, we must have equality.
q.e.d.
p ,p
Remark 4.3.1. The condition u g^H^ (ft), i.e. u-g HQ (Q), is a
(generalized) Dirichlet boundary condition. I t means that u = g on dQ
in the sense of Sobolev spaces.
4.4 Convex functionals on H i l b e r t spaces and M o r e a u - Y o s i d a

approximation
In this section, we develop a more abstract method for showing the ex
istence of minimizers of variational problems. I t has the advantages that
it does not need the concept of weak convergence and that it provides a
constructive approach for finding the minimizer. I n order to concentrate
on the essential aspects, we shall only treat a special situation.
Definition 4.4.1. Let X be a metric space with metric d(-,-)> and let
F : X > R U {oo} be a functional. For X > 0, we define the Moreau-
x
Yosida approximation F of F as
x 2
F (x) := mi(\F(y) + d (x, y)) (4.4.1)
yex
for x X.
Remark 4-4-1- This is different from the definition in Section 5.1 where
2 a
we shall take d(x,y) instead of d (x,y). Here, one might take d (x,y)
for any exponent a > 1. For our present purposes, it is most convenient
to work with a = 2.
We now let i f be a Hilbert space with scalar product (, ) and norm

||-|| and induced metric d(x,y) = | | : r - 2 / | | . Let D(F) C i f , and let
F : D(F) - > l b e a functional. We say that F is densely defined i f D(F)
4-4 Convex functionals 191
is dense in i f . For x D ( F ) , we put F(x) = oo. We say that F is convex

if whenever 7 : [0,1] > i f is a straight line segment, then for 0 < t < 1
F ( ( * ) ) < t F ( ( 0 ) ) + (1 - i ) F ( ( l ) ) .
7 7 7 (4.4.2)
In particular, i f 7(0), 7(1) G D ( F ) , then also 7(f) G D ( F ) for 0 < t < 1.
L e m m a 4 . 4 . 1 . Le F : i f R U { 0 0 } 6e convex, bounded from below,

and lower semicontinuous. Then for every x G i f and X > 0, there exists
a unique
y =:
x
J\x)
with
x x 2 x
F (x) = \F(y ) + d (x,y ) (4.4.3)
Proof. We have to show that the infimum in (4.4.1) is realized by a

x
unique y .
x
Uniqueness: Let y ,y be solutions of (4.4.3), and let
2
Vo = \(Vi +02)
be their mean value. By convexity of F
x x
F(yZ)<\(F(y )+F(y )), (4.4.4)
x
and by Euclidean geometry, i f y ^ y , we have
2
A 2
lk-J/o H <^(|k-îr + ||x-^|| ), 2
(4.4.5)
hence
x A 2
XF(y ) + | | x - j/o || < mvi) + Ik - 2/2 If
A A 2
= AF(j ) + | | z - /
/2 2 2 || ,
x
contradicting the minimizing property of y and 1/2 Thus, we must have
2/i = 2/2 > proving uniqueness.
Existence: (4.4.5) may be refined as follows: For 1/1,1/2 i f and
2/o : = ^(2/1+1/2)
we have for any x G i f
2 2 2 2
l b ~ 2/o|| = 5 ( | | * ~ yi\\ + \\x - 1/2II ) - \ lift - 2/2II (4.4.6)
192 Direct methods
We now let (y )neN

n be a minimizing sequence, i.e.
2 2
XF(y )
n + \\x - y \\ n - inf (xF(y) + ||x - y | | ) = : . A (4.4.7)
yH \ /
We claim that (y ) is a Cauchy sequence. For /, k G N , we put

n
Vi,k '= +
Using the convexity of F as in (4.4.4) and (4.4.6), we obtain

2
><F(yk,i) + \\x-yk,i\\
2 2
< \ (\F(y )k + \\x - y \\ ) k + \ (\F(y ) t + \\x - yi\\ ) - \\\yk yi\\ >2
(4.4.8)
By definition of K\ (see (4.4.7)), the left hand side of (4.4.8) cannot be
smaller than K\, and so we conclude that
2
llifc-wll -o
as fc, I > oo, establishing the Cauchy property. Since the norm is con
x
tinuous and F is assumed to be lower semicontinuous, the limit y of
(Vn)neN then solves (4.4.3). q.e.d.
x x
L e m m a 4.4.2. Let F and y = J (x) be as in Lemma 4-4-1- Let x be
in the closure of D(F). Then
x
x = l i m J (x). (4.4.9)
A+0
Proof. Since x is in the closure of D(F), for every 6 > 0, we may find
x B(x,6)
6 := {yeH :\\x~y\\<6}
with
F(xs) < oo.
Then
lim (\F(x ) 6 + \\x - 2
,511 ) < 62
and therefore
lim sup K\ < 0 (4.4.10)

A-+0
(see (4.4.7) for the definition of K,\).
Let us now assume that there exists a sequence A 0 for n oo w i t h n
Xn 2
\\x-y \\ >a>0 for all n. (4.4.11)
4-4 Convex Junctionals 193
Then from (4.4.10)
x Xn 2
]im8up(\ F(y )
n + \\x-y \\ ) < 0, (4.4.12)
hence
Xn
F (y ) -> - o o asnôo. (4.4.13)
(4.4.12) and (4.4.13) imply
1 x 2 Xn Xn
Fiy ) + | | x - y \\ < F(y ) + \\x - y \f -+ - o o as n ^ oo
which is impossible. Thus, (4.4.11) cannot hold, and (4.4.9) follows.

q.e.d.
T h e o r e m 4 . 4 . 1 . Let F : H E U { o o } be convex, bounded from below,

x x
and lower semicontinuous, and F ^ oo. For x M, we let y = J (x)
Xn
as in Lemma 4-4*1- If (y ) eN n is bounded for some sequence X oo, n
x
then (y )\>o converges to a minimizer of F as X oo.
Xn
Proof. Since (y ) eN n is bounded and since y Xn
minimizes
2
F(j/) + i - | | a ; - j , | | ,
we obtain
x
F(y )^ inf F(y)
yen
Xn
so that (y )nen is a minimizing sequence for F. We now claim that
A
n*- ir
is monotonically increasing in A. Indeed, let 0 < fi\ < fa- Then by
1
definition of y^
2 2 l 2
W ) + \\x - y^\\ > W ) + ||x - y^\\ ,
Mi Mi
hence
2 2
W 2
) + - | | x - y^\\ > F(y^) + \\x - y ||
M2 M2
+ l 2 2 2
( r - r ) ( l ^ - ^ l l - l l x - ^ i i ) .
\Mi M2/ V /
2
This is compatible w i t h the minimizing property of y^ only i f
2 2 2
| | * - J H I > I I * - 2 / ' I I
194 Direct methods
and monotonicity follows. This monotonicity then implies that
a
\\*-ir\\
is bounded independently of A since it is assumed to be bounded for the
sequence A > oo. We next claim that
n
x
F(y )
monotonically decreases towards
inf F(y).
yH
A
Indeed, from the definition of t / ,
x
F(y ) = inf F(t/),
{y.\\x-y\\<\\x-y>\\}
x x
and therefore y has to decrease since | \x y 1 1 increases. The limit has
Xn
to be i n f # F(y) since this is so for the subsequence {y ) eN-
y We now n
a
claim that (?/ )A>O satisfies the Cauchy property, i.e. for every e > 0,
there exists Ao > 0 such that for all A,// > A 0
x 2
\\y -y\\ <e.
For that purpose, we choose AQ SO large that for A, \i > AQ
< \ (4-4.14)
which is possible by the preceding monotonicity and boundedness re

sults. We may also assume
i V ) > F(yn- (4-4.15)
We let
Then from the convexity of F , (4.4.15), and (4.4.6),
F(y^) + {\\x-y^\f
x x 2 x 2
< * V ) + \{\\\ -y \\ + \ I I * - w " l l a
-\\\y - y"\\ )
>M .. 1
(ll~ ..AM , 2 6 1
|L,A
<F (y*) + ^\\ -y*\\+- -- \\y*-y\
x 4 4
by (4.4.14).
4-5 Euler-Lagrange equations 195
This, however, is compatible with the minimizing property of y only i f

x
y -y'
x
Thus (y )\>o satisfies the Cauchy property for A oo, and it therefore
x
converges to some y H. y then minimizes F , because F(y ) decreases
towards i n f / f F(y) for A ^ oo, and F is lower semicontinuous.
y
q.e.d.
The preceding reasoning is adapted from J. Jost, Convex functionals

and generalized harmonic maps between metric spaces. Comment. Math.
Helv. 70 (1995), 659-673.
For a more general construction, see J. Jost, Nonpositive Curvature:
Geometric and Analytic Aspects, Birkhauser, Basel, 1997, pp. 61-4. I n
particular, the method also works in uniformly convex Banach spaces.
General references for Moreau-Yosida approximation are the books of
Attouch and dal Maso quoted in Chapter 6.
Theorem 4.4.1 yields an alternative proof of Theorem 4.3.1 in case
p = 2. Namely, Lemma 4.3.1 implies the lower semicontinuity, Lemma
4.3.2 the convexity of the functional, and the Poincare inequality the
boundedness of any minimizing sequence, as described in the proof of
Theorem 4.3.1. The present proof, however, does not need the concept
of weak convergence. As mentioned, the method extends to uniformly
convex Banach spaces, and thus can handle also arbitrary values of p > 1
(see Remark 3.1.2).
4.5 T h e E u l e r - L a g r a n g e equations and regularity questions

In this section, we return to the variational problems considered in Sec
tions 4.1 and 4.3; we consider variational integrals of the form
d
on a bounded, open subset fl of R , and we make the following assump
( i
tions o n / : f i x E x E ^ E = E U { o o } :
d
(i) / ( , u, v) is measurable for all u E , v R .
(ii) f(x, , ) is differentiable for almost all x ft.
p p
(iii) \f(x,u,v)\ < c - f ci \u\ - f c \v\ , c , c i , c constants, for almost
0 2 0 2
all x fl, and all u E, v R . d
1,p
Condition (iii) implies that # ( u ) is finite for u H (fl), since fl is
bounded. (If fl is unbounded, this still holds provided c = 0.) I n the 0
196 Direct methods
preceding section, we have obtained some results on the existence of a

,p l p
minimizer for # in the class g + i ? o ( f i ) , for given g G H ' (fl). I n the
present section, we wish to characterize such minimizers by necessary
conditions. These conditions will assume the form of differential equa
tions. In fact, these differential equations will hold for arbitrary critical
points of # (as specified in the assumptions of our subsequent results),
and not only for minimizers.
T h e o r e m 4 . 5 . 1 . Let f satisfy (in addition to (i)-(iii))
(iv)
df p p
(x, u, v) r(x,W, v) < c 3 + c \u\ + c \v\ ,
4 5
du dv
2=1
d
C3, C4, C5 constants, for almost all x G fl, and all u G R , V G M. .
p
Let u be a minimizer for # m he c/ass g + HQ* (l) (g G i f ^ f i ) given).
We then have for all )= / f(x,u(x) + tip(x),Du(x) + tDip(x))dx.

Jn
By (ii), (iii), (iv), we may apply Corollary 1.2.2 to conclude that $(u+t(p)
is differentiable w.r.t. </?, and
$ ( u 4- tip)
= J | ^ ( x , u ( x ) +t<p(x),Du(x) +tDip(x))ip(x)
d</?(x)'
x d x 4 5 3
+J2^~i( ^ ^+^fr)+^( ))
x u
^?} - (--)
t=l
Furthermore, (4.5.2) implies
j $(u
t + t<p)\t= = 0. 0 (4.5.4)
Equations (4.5.3) and (4.5.4) imply (4.5.1).

q.e.d.
Remark 4-5.1. From the preceding proof, it is clear that we do not need
to assume that u is a minimizer for I f suffices that u is a critical point
for # in the sense that
~ $ ( u + tip)\ t=0 for all if G C (ft). 0 (4.5.5)

at
Corollary 4.5.1. Suppose that f satisfies (i)-(iv), and in addition, f G
2 2 }P
C . If u G C ( Q ) minimizes $ in the class g + HQ (Q) (or, more gen
erally, satisfies (4-5.5)), then
du
ij=l i=l
d
a : w x I ) u a :
+E^( ' ( )' ( )) - ^(x,u(x),Du(x))=0.
(4.5.6)
Definition 4.5.1. Equation (4-5.6) is called the Euler-Lagrange equa

tion for
Proof (Corollary 4-5.1). By the differentiability assumptions made, we

may integrate (4.5.1) by parts to obtain
,u{ ),Du{ ))-^-{ )

u{x X X X
1=1
{ x u { x ) D u { x ) ) ] i p { x ) d x ( 4 5 7 )
- Y . S L ' ' - -
1=1
From Lemma 3.2.3 (applied to supp</? C C fi so that the term in { }
2
is in L ) , we then obtain (4.5.6).
q.e.d.
198 Direct methods
Equations (4.5.6) constitutes a quasilinear partial differential equa

tion of second order for u. Many such partial differential equations arise
as Euler-Lagrange equations of variational problems. Therefore, i f one
wants to solve such an equation, one might t r y to find a minimizer of
the associated variational problem. However, the existence theory for
minimizers as described in Section 4.3 naturally yields an element u
p
of the Sobolev space H^ (fl), whereas in Corollary 4.5.1 it is required
2
that u be of class C (ft). Thus, there exists a gap, since in general
p 2
elements of Hl' (ft) are not of class C . I t is the task of regularity
theory to bridge this gap, i.e. to show that under suitable assumptions
2
on / , any minimizer of # is smooth, and specifically here of class C .
The theory of partial differential equations indicates that such a result
does not hold without additional assumptions on / , like an ellipticity
tJ
assumption, meaning that the matrix (a '(x))tj=i,...,d w i t h coefficients
a^(x) = yiQ j (x, u(x), Du(x)) is positive definite. Indeed, examples
d v
show that without such an assumption, in general one does not get
smoothness of minimizers. On the positive side, however, we do have de
Giorgi's and Nash's:
d d
T h e o r e m 4.5.2. Let ft be open and bounded inR , f : ft x R * E be
of class C, with
(i)
2 2
A M < / ( x , t ; ) < A ( l + |t;| )
and
d
for all x G fi, u E E , G E , with constants X > 0, A < oo,
(")
^-(x v) < M(l + \v\) for a constant M < oo.
2
Let u G g + HQ (Q) be a bounded minimizer of F(u) :=
1,p
f(x,Du(x))dx (g G H (ft) given). Then u is smooth in ft (u G
/ n
C(ft)).
The proof of the theorem of de Giorgi and Nash is too long to be pre
sented here. We refer to M . Giaquinta, Introduction to Regularity Theory
for Nonlinear Elliptic Systems, Birkhauser, Basel, 1993, pp. 76-99 and
J. Jost, Partielle Differentialgleichungen, Springer, Berlin, 1998 where a

detailed proof is given. Of course, there also exist extensions of this result
to more general integrands of the form / ( x , u, v). We refer the interested
reader to O. Ladyzhenskaya, N . Ural'tseva, Linear and Quasilinear El
liptic Equations, Academic Press, New York, 1968 (translated from the
Russian), Chapters I V - V I .
One remark is in order here: Since Sobolev functions are only equiva
lence classes of functions (in the sense specified at the beginning of Sec
tion 3.1), a more precise version of Theorem 4.5.2 is: Under the stated
assumptions, the equivalence class of u contains a function of class C.
This point, however, usually is assumed to be implicitly understood in
statements of regularity theorems.
In order to display at least one regularity result, however, we consider
a particular example:
d c 1 2
For a bounded, open ft C E , g G J f ' ( n ) , we wish to minimize
Dirichlet's integral
2
D(u) := I \Du(x)\ dx (4.5.8)
JQ
yP
in the class g + HQ (ft). By Theorem 4.3.1, a minimizer u exists, and
by Theorem 4.5.1, i t satisfies
/ Du(x) Dip(x)dx =0 for all </? G C ( n ) (4.5.9)

JQ
(here Du(x) D<p(x) := Yli=i Diu(x)Di<p(x)). I f u can be shown to be

2
of class C , i t would satisfy
(A is called Laplace operator.) by Corollary 4.5.1, i.e. i t is harmonic.

This is the famous D i r i c h l e t p r i n c i p l e : obtain a harmonic function u
in ft with boundary values g by minimizing the Dirichlet integral among
all functions with those boundary values.
In order to justify Dirichlet's principle i t thus remains to show that any
2
solution of (4.5.9) is of class C . Actually, one can show more, namely,
u G C (in fact, u is even real analytic in ft but this will not be demon
strated here), and at the same time weaken the assumption. Namely, we
have:
200 Direct methods
1
T h e o r e m 4.5.3 (Weyl's lemma). Let u L ( f i ) satisfy
[ u(x)Aip(x) =0 for allv GC (O). 0 (4.5.10)

JQ
ThenuC(Q).
Remark 4-5.2.
(1) Clearly, (4.5.9) implies (4.5.10) by definition of Du.

(2) The remark made after Theorem 4.5.2 again applies.
Proof (Theorem 4-5.3). We consider the mollifications with a rotation-

ally symmetric p (and we express this by writing p as a function of \x\)
Uh{x) = u{y)dy
hjjici^)
as in Section 3.2. Given (p Co(Q), we restrict h to be smaller than
dist(supp</?, dft). We obtain
u(y)dyA(p(x)dx
= / u(x)A<ph{x)dx, (4.5.11)
JQ
using Fubini's theorem.

q.e.d.
Remark 4-5.3. We have also used the fact that A commutes with mol
lification, i.e.
(A<p) = A(<p ).
h h (4.5.12)
For this, one needs that g is a function of |x| only, i.e. rotationally
symmetric. Also, this point needs the rotational invariance of the Laplace
operator A . Therefore, the present proof does not generalize to other
variational problems.
After this interruption, we return to (4.5.11) and conclude that
/ u (x)Aip(x)dx
h =0 (4.5.13)
JQ
by applying (4.5.10) to tph CQ(Q) (by our choice of h). Since UH is

smooth, we obtain e.g. from Corollary 4.5.1
Au h = 0
in Slh ' {x Q | d i s t ( i , 9 Q ) > h}. Also
/ K ( y ) l \dy
<
< / M * ) l \dx
< (4.5.14)
Jn
X
by Fubini's theorem, using ~ J ^ fe ^ dy = 1 by (3.2.3)
1
< oo since u L ( f i ) .
Therefore, the functions Uh are uniformly bounded in L . We now need 1
2
L e m m a 4 . 5 . 1 . Let f C (Q) be harmonic, i.e.
Af(x) = 0 into.
r
Then f satisfies the mean value property, i.e. for every ball B(xo, ) C 0,,
f(xo) = "^3 / /(*)<** = ^ - ^ r r / f(x)da(x) (4.5.15)

yB(a:o,r) JdB(x ,r)
0
d
where uJd is the volume of the unit ball in R.
Proof. For 0 < g < r
0 = / Af(x)dx
JB{X ,Q)
0
I j(x)da(x),
JdB(x ,Q)0
where v denotes the exterior normal of S(xo, g)

d l
(y + gu)g ~ duj
JdB(0X i) dg
in polar coordinates UJ
Q
Q JdB(0,\)
1
Q -* f f(x)da(x))
dg
1
(dWd^" J DL
202 Direct methods
Thus,
is constant in g, and since its limit for g 0 is / ( s o ) as / is continuous,

it has to coincide with f(xo) for all 0 < g < r. Since
d l
-i-y / f(x)dx = 4 M T - V T / f(x)da(x)) g ~d Q)
the first inequality in (4.5.15) also follows.

q.e.d.
We return to the proof of Theorem 4.5.3: Since UH is harmonic, it

satisfies the mean value properties of Lemma 4.5.1. Since the family Uh
1
is bounded in L ,
Uh(x ) 0 = ~ j u (x)dx
h
is bounded for fixed r w i t h B(xo,r) C fi^. Therefore, the Uh are uni

formly bounded in flh for 0 < / i < ^ . Furthermore, from (4.5.15)
0
\u (xi)
h - u (x )\
h 2 < ~ (-) [ \u {x)\dx
h
dWd XT 1 / B(x ,r)\B(8 r)
1 2 )
X /
UB(i r)\B(*i,r)
2 l
< c(r)\xi -x \2 (4.5.16)
for some constant depending on r, i f B(x\, r ) , B(x ,r) C . Therefore, 2
the gradient of Uh is also uniformly bounded on fi^ . Likewise, deriva 0
tives of Uh of all orders can be uniformly bounded on lh (0 < h < ^ ) , 0
either by repeating the same procedure, or by observing that together

with Uh, also all derivatives of UH are harmonic so that (4.5.16) can
be iteratively applied to all derivatives in order to convert a bound on
some derivative into a bound for a higher one. Therefore, a subsequence
of un converges towards some smooth function v, together with all its
derivatives, as h 0. Since all the Uh satisfy Auh = 0 so then does v:
Av = 0 in Q.
Since on the other hand Uh converges to u in L ( 0 ) by Theorem 3.2.1, x
the two limits have to coincide (e.g. by Lemma 3.1.3). Therefore u = v,

and consequently u is smooth and harmonic.
q.e.d.
As an application, we consider the following
Exercises 203
Example 4-5.1. Let a : R R be Lipschitz continuous w i t h
0 < A < a(y) < A < oo for all y G R.

d
Let fi C R be open. We want to minimize
d
r
F(u):= / V]a(u(x))AM(a:)Aw(a:)da; (4.5.17)
p 1 p
in the class A : = # + # d ' ( f i ) , with given # G i f (f2). By the Picard-
Lindelof theorem, the ordinary differential equation
P = (4-5.18)
1 1
admits a solution u(v) of class C ' . We then have
. . du du ,, .
( . ) - - = !. (4.5.19)
Since ^ > A~* > 0, the inverse function v(u) exists and is of class C '
as well, and we have by (4.5.19) and a chain rule for Sobolev functions
that easily follows from the chain rule for differentiable functions by an
approximation argument that
d d
a(u)DiuDiU ^2 F>ivDiV.
i=l i=l
Therefore, (4.5.17) is transformed into Dirichlet's integral
F(u) = D(v).
Since the latter admits a smooth minimizer, the original problem (4.5.17)
1 1
then admits a minimizer that is of class C ' in fi.
Exercises
4.1 Weaken the growth assumption required for | in (iv) of The
orem 4.5.1. Hint: Use the Sobolev Embedding Theorem.
4.2 Compute the Euler-Lagrange equations for the variational in
tegral
2
A(u) := / y/l + \Du{x) \dx.
Jn
(A(u) represents the volume of the graph of u over fi. Critical
points are minimal hypersurfaces that can be represented as
graphs over fl.)
204 Direct methods
4.3 Compute the Euler-Lagrange equations for
lj
E(u) := / g (x)Diu(x)Dju(x) (detgij(x))^ dx,
Jn
u
where (<7 ())i,j=i,.,,,d is the inverse matrix of {gij{x))ij=i,...,d.
Assume that (0ij(x))i,j:=i,...,d is positive definite for all x G ft.
x 2
Show that for given g G H ' (ft), there exists a unique minimizer
1 , 2 1 , 2
of J? among all u G i ? ( f i ) with u g G i f ( f i ) . (Minimizers
for J? are harmonic functions w.r.t. the metric gij(x).)
5
Nonconvex functionals. Relaxation
5.1 Nonlower semicontinuous functionals and relaxation

From Section 4.3, we recall the following
d d
T h e o r e m 5.1.1. Let ft C R be open, l E
measurable and suppose:
d
(i) For almost all x eft, / ( # , ) is convex on R
1
(ii) There exist a G L (ft), b G E with
p
f{x,v) > -a(x) + b\v\
d
for almost all x G ft and all v G E .
Then
1,p
is Isc and convex on H (ft) equipped with its weak topology and assumes
lyP yP
its infimum in the class of all f G H (ft) with f g G H (ft) for 0
lyP
some given g G H (ft).
Here, (ii) is just a coercivity condition ensuring that a minimizing

lyP
sequence stays bounded w.r.t. the H -norm (w.l.o.g. F ^ oo) (i)
x p
implies that F is lsc, w.r.t. the norm topology of H ' , and the convexity
l,p
then implies that F is also lsc w.r.t. the weak H topology. Since
1,p
bounded sequences in H have weakly convergent subsequences, any
minimizing sequence has a convergent subsequence, and a limit of such
a subsequence then minimizes F by lower semicontinuity.
Not all functionals that one wishes to consider in the calculus of varia
tions are convex, however. As a motivation for what follows, we consider
205
206 Nonconvex functionals. Relaxation
the following example of Bolza:
fi = ( 0 , l ) c R , u:(0,l)->R, u(0) = 0 = u{l)
2 2 2
F(u) = (u (x) + (u'(x) - l ) ) dx.
We claim that
4
i n f { F ( u ) : u G H ^ ( ( 0 , 1 ) ) } = 0. (5.1.1)
For the proof, we consider 'sawtooth'-functions: Let n G N ,

i r 2i 2i + 1
x for < x <
n 2n 2n
u (x)
n := {
, i+ 1 f 2i: + 1 ^ ^ 2i + 2
x H for < x <
n 2n 2n
(2 = 0 , l , . . . , n - l ) .
M
u is contained in ^ ^ ^ ( ( 0 , 1 ) ) C # ( ( 0 , 1 ) ) and satisfies:
n
For all x G (0,1) 0 < u (x) < n (5.1.2)

2n
Mn (0)=o = ti (l),
n (5.1.3)
for almost all x G (0,1) K ( x ) | = 1. (5.1.4)

Consequently
lim F(u ) n = 0.
Since F(u) is nonnegative for every u, (5.1.1) follows. The infimum of F

4
therefore cannot be realized by any H Q ' function, because if we had
F(u) = 0,
;
then u(x) = 0 for almost all x G (0,1) and | w ( x ) | = 1 for almost all
x G (0,1), and these two conditions are not compatible. (In fact, since
4
d = 1 here, any u G # Q ' ( ( 0 , 1)) is absolutely continuous, and so u = 0 i f
u(x) = 0 a.e., hence u is differentiable and u' = 0. (More generally, any
Sobolev function that is constant on some set A has a representative u
whose derivative Du vanishes on A.)
We have thus shown that the problem
M
F(u) -+ min in # 0 (fi)
does not have a solution.

We observe that our minimizing sequence (u ) n converges to zero

4
weakly in # d ' , by (5.1.2) and
/ u' {x)<p(x)dx = - f u (x)<p'(x)dx

n n 0 for all (p G C^((0,1)).
Jo Jo
However,
F(0) = 1 > 0 = l i m F{u ).
n
noo
1
Therefore, F is not lsc w.r.t. weak H ^-convergence although the inte
grand is continuous in u'. As we shall see this results from the lack of
convexity of the integrand. We also observe that any sequence of saw
tooth functions u , i.e. satisfying
n
\u' \ = 1 a.e.
n
2
that converges to 0 in L is a minimizing sequence for F .
Remark 5.1.1. Functionals of the type of our example often arise in op

timal control theory as described in Section 5.2 of Part I . For example,
one considers problems of the following type
[ f(t, u(t),a(t))dt -+ min (5.1.5)

Jo
under the side conditions
u(0) = u , 0 u{T) = u T (5.1.6)
u'(t)=g(t,u(t),a(t)) (5.1.7)
w i t h given functions / and g. u is called a state variable, a a control

variable. This means that one assumes that u describes the state of
some system evolving in time t whose derivative or rate of change can
be controlled through a parameter a. The aim then is to choose a in
such a manner that the functional, often considered as 'cost function',
is minimized.
Thus, one needs to find some equation
a(t)=<p(t,u(t))
for an optimal control a at time t assuming a given state u(t) of the sys
tem. I f one knows the optimal control, one can reconstruct the evolution
u(t) of the state of the system from (5.1.6) and (5.1.7) under appropriate
assumptions. The simplest control equation (5.1.7) is
u'(t)=a(t),
and this leads to minimizing functionals of the type
/ f(t,u(t),u'(t))dt.
Jo
f 2 2
Expressions of the type (u (t) l ) can occur in many technical exam
ples, like boats sailing against the wind.
Faced w i t h a problem that one cannot solve, one may contemplate

several options:
One could t r y to modify the problem, or one might generalize the
concept of a solution, or both.
We shall discuss several such strategies. We first modify the problem
via relaxation. This is an important method in the calculus of variations,
and we therefore discuss it in some generality.
D e f i n i t i o n 5 . 1 . 1 . Let X be a topological space, F : X E . We define

the lower semicontinuous envelope or relaxed function sc~F of F as
follows:
(sc~F)(x) : = sup { $ ( # ) : $ : X E is lower semicontinuous

with < F(y) for ally e X}
L e m m a 5 . 1 . 1 . sc~F is the largest Isc function on X that is < F

everywhere.
In particular, F is lower semicontinuous if and only if F = sc~F.
Proof. sc~F is Isc as a supremum of Isc functions, see Lemma 4.2.1 (iv).
Obviously, sc~F < F , and for all Isc $ w i t h # < F , we have
$ < sc~F
by definition of sc~F.
q.e.d.
T h e o r e m 5.1.2. Let X be a topological space, F : X E a function.

Then every accumulation point of a minimizing sequence for F is a
minimum point for sc~F. Consequently, if F is coercive, then sc~F
assumes its minimum, and
min sc~ F = inf F.

x x
Proof Let (# )neN C X be a minimizing sequence for F w i t h accumu

n
lation point #0- Then
(sc~F)(x )
0 < l i m i n f ( s c ~ F ) ( x ) by lower semicon-
n
noo , . _ .
tinuity of sc F (see
Lemma 5.1.1)
< lim inf F(x )n since sc~ F < F
n>oo
= inf F(y) since (x ) is a min-
n (5.1.8)
lmizing sequence tor b .
On the other hand, the constant function
* ( * ) = inf
yX
is Isc and < F , hence by Lemma 5.1.1 for every x G X
inf F(t/) < {sc~F)(x). (5.1.9)
From (5.1.8) and (5.1.9) we conclude
{sc~F)(x ) 0 = inf F(y) = m i n ( 5 c " F ) ( x ) . (5.1.10)

yX xX
This implies the first claim. I f F is coercive, then every minimizing

sequence has an accumulation point, and the second claim also follows.
q.e.d.
What does Theorem 5.1.2 tell us for our example?

It simply says that i f we cannot minimize our original functional F due
to its lack of lower semicontinuity, we then minimize another functional
instead, one that is lower semicontinuous and as close as possible to F .
Theorem 5.1.2 then says that limits (or more generally, accumulation
points) of minimizing sequences for F do not minimize F , but the re
laxed functional sc~F. Since sc~~F is the largest Isc functional < F by
Lemma 5.1.1 that is the best one can hope for.
I t then remains the task to determine the relaxed functional of some
given F . Before proceeding to do so for our example, let us relax ourselves
a little and derive some easy consequences of the definition of the relaxed
functional and consider some easier examples first.
L e m m a 5.1.2. Let X satisfy the first axiom of countability. Then sc~F

is the relaxed function for F : X R iff the following two conditions
are satisfied:
(i) whenever x n x
(sc~F)(x) < l i m i n f F ( x ) n
noo
(ii) for every x G X, there exists a sequence x n x with
(sc~F)(x) > lirn F ( x ) n

nco
Proof. We claim that, since X satisfies the first axiom of countability,
(sc~F)(x) = inf { l i m inf F(x ) n :x n x in X}. (5.1.11)
We denote the right hand side of (5.1.11) by F~(x). Then F ~ is lsc. I n

order to verify this, we have to check
:
l i m i n f (inf { l i m i n f F (y^.n) Vv.n Vv\) > inf { l i m inf F (x
n j Xn X j
v>oo
(5.1.12)
whenever y > x. Indeed, otherwise, for some 5 > 0, we would find some
v
diagonal sequence y ,n ^ as v oo w i t h
u u
^ (2/^,nJ < inf { l i m inf F(x )

n : x n x} - 6
which is impossible. Thus, F~ is sequentially lsc, hence lsc, because X

is assumed to satisfy the first axiom of countability. Also, F~ < F, and
for every lsc $ < F , we have for x x n
$(x) < liminf $ ( x ) < l i m i n f F ( x ) ,n n
noo n+co
and hence
< F"(x).
Thus, F"* is the largest lsc functional < F , and (5.1.11) follows from
Lemma 5.1.1. I t is then easy to see (and left as an exercise) that F ~ ( x )
satisfies and is characterized by the properties (i) and (ii).
q.e.d.
Example 5.1.1. Let X be a topological space, A C X a subset. The

indicator function %A is defined by
i i x A
t A ( x ) . = l *
{oo if X iA.
We then have
SC~%A i A,
where A is the closure of A in X.

The characteristic function \ A is defined by
G A
XA{X)1
={1 tf* iA.
Then
sc \A = XA
where A is the complement of X \ A.
d
Example 5.1.2. Let fl C R be open, 1 R
defined by
/( ) = I L
U :
Du
\ \" d x
+ fa M " d x i f
C ^) 1
i oo otherwise.
1
(Note that 7(u) may also be infinite for some u G C (f2).) We claim
u d x d i f u 1 , P n
(*r/)() = ( /n l ^ l " + /n M " * # ( )
1 oo otherwise.
In order to show this, we shall verify the conditions of Lemma 5.1.2:
p
(i) (sc~I) is lower semicontinuous on L which yields condition (i).
The lower semicontinuity is seen as follows:
p
Suppose u u in L (fl).
n For the purpose of lower semicontinu
ity, we may select a subsequence (w)eN C (w )nN w i t h n
lim (sc~I)(w ) l/ = liminf(sc~J)(u ), n
v>oo noo
and we may also assume that this limit is finite. ( W ) N then
c 1 p
is bounded in J f ' (f2). A subsequence of (w ) then converges v
p
weakly in H^ (fl) (Theorem 3.3.5), and by the Rellich-Kondra-
chev compactness Theorem 3.4.1, i t also converges strongly in
p
L (fl). The limit has to be u, because the original sequence (u ) n
liP
was assumed to converge to this limit. Since the H -norm is Isc
l,p
w.r.t. weak H convergence (Lemma 2.2.7), we have
(sc~I)(u) < l i m (sc~J)(uv)

VKX)
= liminf(sc~/)(u ). n
(ii) Let u E H^(n). Since C (f2) fl H (Q) 1

is dense in / f ^ f i ) , we llP
1 , p
may find a sequence ( w ) N C C ^ f i ) f l ff ( f i ) w i t h
n n
p p p P
lim f/ |>u | + / K |
n ) = / |Du| + H ,
N
^ \JQ JQ ) JQ
i.e.
lim I(u ) n (sc~I)(u).
1
HugH *^), then
7(u) = (sc~/)(ti) = oo.
This verifies condition (ii).
Example 5.1.3. Similarly, for
o K 1
'' \oo if w L P ( n ) \ c ( n ) , 0
the relaxed functional is

D
(sc-l )(u)
= U \ 0 < i f t c e ^ n )
I oo otherwise.
Remark 5.1.2. We may also define the above functionals 7, J on L f ( f i ) 0 oc
p
instead of L ( f i ) . The relaxed functionals will be given by the same
formulae.
Remark 5.1.3. For p = 1, the relaxations of I and I are not given 0
c 1 1
anymore by the J f ' -norm, but by the BV-norm which is defined in
Chapter 7.
In metric spaces, there is an alternative useful characterization of the

relaxation of a given functional which we now want to describe.
D e f i n i t i o n 5.1.2. Let X be a metric space with distance function d(-,

F :X RU {oo} be bounded from below, F ^ oo. For X > 0, we define
the Moreau- Yosida transform of F as
F (x)
x : = inf (F(y) + \d(x,y)). (5.1.13)
T h e o r e m 5.1.3. The functionals F\ satisfy
\F ( )
x Xl - F (x )\ x 2 < Ad(xi,x ) 2 (5.1.14)
for every X > 0, # i , x 2 X. In particular, they are Lipschitz continuous.

For any x G X
(sc~F)(x) = lim F (x). x (5.1.15)
5.2 Representation of relaxed functionals via convex envelopes 213
Proof. For # i , x , t / G X , A > 0, we obtain from the triangle inequality

2
F(y) + Xd{x y) u < F(y) + \d(x ,y) 2 + Xd(x x ). u 2
The definition of Fx (#2) implies then
inf (F(y) + Xd(x y)) u < F (x )

x 2 + Xd(x x ),
u 2
yex
hence
F {xi)
x < F ( x ) + Ad(x ,x ).
A 2 1 2
Interchanging the roles of # i and x , we conclude 2
\F (xi)-Fx(x2)\<\d(x x ).
x u 2
Since we have now shown that F\ is Lipschitz continuous, and since

F\ < F , we obtain
FA < 5C"F,
hence for all x G X
s u p F ( x ) < (sc'F)(x).
A (5.1.16)
A>0
For any A > 0, we find x\ X with
F ( x ) + A d ( x , x ) <F (x)
A A x +j .
Therefore
lim x A = x
A+00
and
(sc~F)(x) < l i m i n f F ( x ) < l i m i n f F ( x ) .

A A (5.1.17)
A+00 A+00
Equations (5.1.16) and (5.1.17) imply (5.1.15).

q.e.d.
5.2 Representation of relaxed functionals v i a convex

envelopes
d d
T h e o r e m 5.2.1. Let Q C R be open, 1 <p < 00, f :R -+R contin
uous with
p
Co \v\ < f(v) < C\ | v | + c p
2 for some constants CQ, CI, c . 2
hp 1 , p
Let F :{u H {Q) :u - u e # 0 0 ( f t ) } -+ R 6e given by
hp
F(u) := / f(Du(x))dx , (u G H (Q)0 given).
JQ
lyP
Then the relaxed function of F w.r.t. the weak H topology is given by
(sc~F)(u)= / (cvx~ f) (Du(x)) dx

JQ
where
(cvx~f)(v) := sup{g(v) : g < f,g convex}
is the largest convex function < f.
For the proof, we shall need the following:

c R d b e
L e m m a 5 . 2 . 1 . Let W = UUi^M
an open rectangle,
P d
1 < p < oo. We let f G L (W) and extend f periodically to R , i.e.
1 d 1 d
f (x + m i (0i - ai),... ,x + m d (0 - a ))
d d = / (x , ... , x )
1 d
for m i , . . . , md E Z, x = ( x , . . . , x ) G W, and put
fn(x) : = f{nx) for n G N .
ITien we /ie lyeo/c convergence

l x dx i n L P W r
fn-f = -TTr I f( ) ( ) f n-ôo. (5.2.1)
meas W J w
Proof. First
p p p p
/ \f (x)\ dx
n = I \f(nx)\ dx = ^ f \f(y)\ dy= f \f(x)\ dx
n
JW JW JnW JW
by the periodicity of / . Thus
\\fn\\ LP{w) = ll/H L P { W ) - (5-2-2)
In the same manner,
/ f {x)dx=
n [ f(x)dx= [ fdx. (5.2.3)
Jw Jw Jw
Let now WQ be a subrectangle of W, written in the form
d
WQ = J J (ai + 6;0!i, a* + 6;$) ,

2= 1
or more compactly
W0 = a + bW (a = ( a , . . . , a ) , 6 = (&i, ...,&<*))
x d
Then
/ (/(*) ~f)dx= f (f(nx) - f) dx

JWo Ja+bW
= - d l (f(y)~f)dy
n
Jna+nbW
= A / ( / ( ) - f) V d
7 1
Jna+[nb]W
d
+ - d I (/() - f) y
n
Jna+(nb-\nb\)W
+ A / {f(y)~f)dy
n i+(nb-[nb})W
Jna+(nb-[nb})W
by periodicity of / .
The first term in the right-hand side vanishes by (5.2.3), and thus, again
using the periodicity of / ,
1/ {fn{x)-f)dx\<- d [ \f(y)~f\dy.
\JWo I 7 1
JW
Letting n oo, we obtain for every subrectangle WQ of W

(5.2.4)
lim / (f (x)-f)dx
n = 0.
q
Let now g L (W), w i t h + = 1. We have to show
lim / f (x)g(x)dx
n = / fg(x)dx. (5.2.5)
n
-*Jw Jw
Given e > 0, we then find subrectangles W\,..., Wk {k k(e)) and
\ e R (i = l , . . . , f c ) w i t h
{
< e (5.2.6)
L*(W)
q d
(The possibility of approximating L (fl) functions g (Q open in R ) in
such a manner by step functions can easily be seen as follows:
9
Since Cg(ft) is dense in L ( f i ) , there exist y> C(fi) w i t h
110 "~ ^ c l l / ^ n ) < I t is then easy to construct a step function A^x^t

(Xi R, Wi disjoint rectangles contained in supp</? ) w i t h
e
sup 2 meas supp (p
supp l | / {fn(x)-f)dx +e||/ -/||
n
by (5.2.6) and Holder's inequality (Lemma 3.1.1).
The first term tends to zero as n oo by (5.2.4), whereas the second one
is bounded by 2e | | / | | , P ( W ) by (5.2.2) and can hence be made arbitrarily
small. Therefore, (5.2.5) holds.
q.e.d.
The proof of Theorem 5.2.1 will be broken up into several steps:
(1) We put
(q-f)(v):=ud{^J f(v u + D (x))dx:

V <p G H**(U),
d
U bounded domain in R | ,
(5.2.7)
and we claim:
L e m m a 5.2.2.
(sc-F)(u) < [ (q~f)(Du(x))dx.
JQ
Proof. Replacing F(u) by G(v) := F(v + UQ) forv = u UQ,

we may assume u = 0, i.e. u G HQ' (Q). 0 Since the piecewise P
affine functions, i.e. those u for which Du is constant on disjoint

rectangles W{ C fi, with \ (J W\ arbitrarily small, are dense in
LYP
H (for the same reason that the functions that are piecewise
p
constant on disjoint rectangles W{ are dense in L ) , and since F
is continuous under strong H^-convergence, it suffices to treat

the case where
Du = VQ = constant
on some rectangle W. We next observe that for a given constant

vector v, (q~ f)(v) is independent of the choice of U in (5.2.7).
First, the value of the inf on the right hand side of (5.2.7) does not
change under translations or homotheties of U. The general case
of U\ and XJi then is handled by approximating U\ by disjoint
homothetical translations of U2 and vice versa. We may therefore
take U = W in (5.2.7). We now choose a sequence (y> )nN C n
P
H*' (W) with
(q~f)(vo) + - > ^ 7 / f(v 0 + D (q-f)(v ).0
n meas W J w
(5.2.8)
d
We extend (p periodically from W to E
n and put
u(x) := vox (then Du = v ) 0
and
u (x)
n := u(x) + - ( / ? ( n x ) . n
n
lyP
By Lemma 5.2.1, u n converges to u weakly in H . Then u n = u
on dW by periodicity of (p and ^ n n ( a v v = 0. We have
/ f(Du (x))dx
n = j f(v 0 + -D(p (nx))dx n
n
Jw Jw
= v D(
I f( o + Pn(y))dy
n d
JnW
/ W
f(v 0 + D<p (y))dy n (5.2.9)
since (p is periodic. n

lim F(u ) n = lim / f(Du (x))dx
n
n-KX) J w
= f (q-f)(v ) 0 = (q-f)(v )measW. 0
Jw
The claim then follows from the characterization of (sc~F), see
e.g. Lemma 5.1.2(i).
q.e.d.
(2) We observe
(q~f)(v)<f(v) (put p = 0 in (5.2.7)). (5.2.10)
With
(q~F)(u) := / (q-f)(Du(x))dx,
Jn
we obtain from Lemma 5.2.2 and (5.2.10)
s c - F = 5C-(g~F), (5.2.11)
and upon iteration

n
sc-F = sc-((q~) F), (5.2.12)
n
where (q~~) means performing the construction q~ iteratively
n times. From the growth conditions on / assumed in The
orem 5.2.1, we conclude that
n
(Q- f)(v)
is monotonically decreasing and bounded from below in n, hence

converges to some limit
(Qf)(v).
From B . Levi's Theorem 1.2.1, we conclude
n n
l i m (q~ F)(u) = lim f (q~ f)(Du(x))dx
noo n>oo J
= f (Qf)(Du(x))dx =: (QF)(u). (5.2.13)
Since by (5.2.12)
n
(sc~F)(u) < (q~ F)(u) for all n ,
we conclude from (5.2.13)
(sc-F)(u) < (QF)(u).
From the definition of Q / , we also conclude
Q/(t;) = i n f { 1 / (Qf)(v + D<p(x))dx,

meas U J v
P d
 R is called quasiconvex if for all
d P d
v G R , <p C H^ (U), U C R bounded and open
l
g{v) < r I g{v + D<p{x))dx. T (5.2.15)
meas U J v
Equation (5.2.14) then implies that Qf is quasiconvex.

d
(3) L e m m a 5.2.3. / : R E is convex if and only if it is quasi-
convex.
Proof. '=>':
Jensen's inequality says that i f / is convex, for every ip G
1 d d
L {R ,R )
f (J^dx^ <jf(i>(x))dx (5.2.16)
(see Theorem 1.1.6). Since, as observed above, in Definition 5.2.1

it suffices to consider one fixed domain U, we may assume
meas U = 1
and put
ip(x) = v 4- D<p(x).
p
Since <p G H^ , f ip(x) = v meas U = v, and (5.2.16) therefore
implies that / is quasiconvex.
We assume that / is quasiconvex, i.e.
f{Vo) = f{vo)dx
^bjl
P
for all ip e HQ' {U).
d
We have to show that for all v i , v R , 0 < t < 1 2
f(tVl + (1 - t)v ) 2 < tf( )

Vl + (1 - t)f(v ). 2 (5.2.18)
f(tvi + (1 - t)v ) < !

2 f f(t Vl + (1 - t)v 2 + D<p{y)) dy
meas U Ju
(5.2.19)
for all U and all (p P

HQ (U). After a rotation, we may assume
that v\ V2 is a positive multiple of the first basis vector of our

standard basis of R , i.e. V\ v points in the ^ - d i r e c t i o n . We
d
2
d d
shall take a cube W : = (a, b) C R as our set U and construct a
family of functions
(v>) N C 6 H^{W)
with
on a set W C W w i t h meas
{l-t)( -V2) Vl d 1
W? = t(b-a)(b-a-) -
V(p (x)
n =
on a set Wg C W w i t h meas
-t(v x - v) 2
d _ 1
WT = (i-*)(fc-o)(fc-o-s)
and
H V ^ n l l ^ o o ^ ) < Co for some fixed constant Co that does not

depend on n.
Using these (p in (5.2.19) yields

n
f(tvx + (1 - t)v ) 2 < tf(V!) + (1 - t ) / ( v ) 4- Pn

2
with p 0 as n oo, hence (5.2.18).

n
n 1
It remains to construct (p . We divide the interval (a, 6) into 2 +
n
subintervals as follows:
h = (a,a+^(b-a))
l2 = (a+^(b-a),a+^(b-a))
h = (a+~(b~a),a+~(b-a) + ^(b-a))
i.e. the intervals hv-i have length ~(b a), and they alternate
with the intervals I 2v of length ^ ^ ( 6 - a). We then put
d 1
Wr'=((jl2,-i)x(a+- b--) - )
X
i/=i
n n
^-^ 2 , , ) x ( a + - n., & - - )n. ~ -

n d 1
W 2 :=(M
2 d
We then put 
8x \ -t\vi-V2\ forxeW?,
*$>=0*i = 2,...,d.
(Remember that we assume that V\ v 2 points in the positive

1
x -direction.)
2 d
We then have <p (b, x , ... x )
n y = 0. We also put
(p = 0 on dW,
n
and on W \ (W U W ) we choose an interpolation that is affine

2
2
linear i n x , Since
, /., t(l t),, ., , Ci
sup \ipn(x)\ < (6 - a)|vi - v | r *
2
n w
\- 2
n
we get
xdip (x)\n Ci
sup < n.
Thus, for large enough n ,

sup |Vy>(x)| = sup 1 " 1 < |vi - v \ = : c .
2 0
xW xEWfUW? ux
This completes the construction of (p and the proof of Lemma n
5.2.3.
q.e.d.
(4) We may now complete the proof of Theorem 5.2.1 From (2), we
know
(sc'F)(u) < QF(u) = jQf{Du(x))dx.

By Lemma 5.2.3, Qf is convex. By Lemma 4.3.1, Qf{u) therefore
1,p
is lsc w.r.t. weak H convergence. Since QF < F (see (5.2.10)
and the definition of Q F ) , we must also have from the definition
of sc~F that
QF(u) < (sc~F){u).
Hence equality. Thus
(sc-F)(u) = J (Qf)(Du(x))dz
Moreover, for every convex function g < /,
G(u) := j g(Du(x))dx
ltP
is a weakly H Isc functional < F . Therefore, from the defini
tion of sc~F, the convex function Qf must in fact be the largest
convex function < / . This completes the proof.
q.e.d.
Corollary 5.2.1. F as in Theorem 5.2.1 is weakly lower semicontinu

1,p
ous in H if and only if f is convex.
Proof. Lemma 4.3.1 says that convex functionals are weakly lower semi-
continuous. I f / is not convex, then by Theorem 5.2.1 sc~~F ^ F , hence
F is not weakly Isc by Lemma 5.1.1.
q.e.d.
Remark 5.2.1. One may also consider variational problems for vector
d n
valued functions u : fl C R -+ E ,
F(u) := [ f(Du(x))dx.
Jn
d
Again, / is called quasiconvex i f for all open and bounded U cR and
p n nd
sl\ ip e H^ (U;R ), v eR
/ ( / (
^ n ^ / ^ ^
In this case, however, while convex functions are still quasiconvex, the
converse is no longer true. Theorem 5.2.1 continues to hold but w i t h con
vexity replaced by quasiconvexity. Also, one may consider more general
problems of the form
F(u) = J f(x,u(x),Du(x))dx
with similar results and conceptually similar, but technically more in

volved proofs.
Remark 5.2.2. The notation of quasiconvexity and many of the basic

corresponding lower semicontinuity results are due to C. Morrey. In
fact, the quasiconvex functionals are precisely the weakly lower semicon-
tinuous ones. For detailed references to the work of Morrey and other
researchers, see the book of Dacorogna quoted at the end of this chap
ter.
Remark 5.2.3. Theorem 5.2.1 can be considered as a representation the

orem for relaxed functionals. I n particular, i t says that a functional on
obtained by integrating an integrand f(Du(x)) (with certain tech
nical assumptions on / ) has a relaxed functional of the same type, i.e.
again representable by integration w.r.t. to some integrand g(Du(x)) of
the same type. Furthermore, g may be computed explicitly from / .
We now return to our initial example
2 2 2
F(u) = j f * | u ( x ) + (u'(x) - l ) } dx
M
for u # ( ( 0 , l ) ) . F(u) is the sum of a functional which is continuous
o
2 1 4
w.r.t. strong L -convergence, hence also w.r.t. weak H * convergence,
and another one to which Theorem 5.2.1 applies. We conclude that
1 2
(sc~F)(u)= f {u (x) + Q(u'(x))}dx,
Jo
with
( v 2 1 ) 2 i f H
Q(v) = l - ^
\ 0 otherwise,
2 2
the largest convex function < (v l ) .
References
For the definition of relaxation and its general properties:
G. dal Maso, An Introduction to F-Convergence, Birkhauser, Boston 1993,
pp. 28-37.
G. Buttazzo, Semicontinuity, Relaxation and Integral Representation in the
Calculus of Variations, Pitman Research Notes in Math. 207, Longman
Scientific, Harlow, Essex, 1989, pp. 7-28.
For Theorem 5.2.1 and generalizations thereof:
B. Dacorogna, Direct Methods in the Calculus of Variations, Springer,
Berlin, 1989, pp. 197-249.
Nonconvex functionals. Relaxation
Exercises
Determine sc~F and discuss the relaxation for
2 2 1 4
F(u) = J (l-u'(x)) u(x) dx foruGtf '
with = 0 , u ( l ) = 1,
, 2 2 1A
F(u) = J {2x-u {x)) u(x) dx for ueH
with = 0 , u ( l ) = 1,
and
1 2 ; 2 1 4
F(u) = J ((u(x) - a) + ( u ( x ) - 1)) dx for u G / J '
with a G i
p d
Determine sc"I for / : L ( f i ) R ( f i G R open and bounded),
1
J(ii) : = { / n W <** + Jn cfa i f ti C ^ )
I oo otherwise.
Why does the proof of Lemma 5.3.3 not work for vector-valued
d n d n d n
mappings R -+ M with n > 1, i.e. # : M -+ R, v G R , y> G
p n
# o ' ( R , R ) as in Remark 5.2.1?
6
r~convergence
6.1 T h e definition of T-convergence

In this chapter, we treat the important concept of T-convergence, intro
duced and developed by de Giorgi and his school.
Definition 6.1.1. Let X be a topological space satisfying the first ax

iom of countability, F : X > R functions (n G N). We say that F
n n
T-converges to F,
F = T- l i m Fn r
if
(i) for every sequence (x ) ^ n n converging to some x G l ,
F(x) < liminf F (x )

n n
and
(ii) for every x X, there exists a sequence x converging to x with n
F(x) = lim F (x ).
n n
Example 6.1.1. F : R -+ R
n
1 for x >
n
F (x)
n nx for < x <
n n
-1 for x < .
n
Then
for x > 0
for x < 0
225
226 r'-convergence
while the pointwise limit is 0 for x = 0, 1(1) for x > 0 ( < 0).
Example 6.1.2. F n : R R
nx for o < x <

n
F {x)
n := 1 2
2 nx for < x <
n n
0 otherwise.
Then
(r-UmF )(x) = 0 n
which is the same as the pointwise limit.
nx
for 0 < x <
n
F (x)
n := I 1 2
nx 2 for < x <
n n
0 otherwise.
Then
-1 for x = 0
(r-limF )(x) = {
n
0 otherwise.
whereas the pointwise limit is again identically 0. Note that the F of n
6.1.3 is the negative of the F of 6.1.2. Thus, in general

n
(r-limF )^r-lim(-F ).
n n
Example 6.1.4- F : n
nx for 0 < x <

n
F (x)
n := { nx 2 for < x < for odd n
n n
0 otherwise
0 for even n.
F n then converges pointwise to 0, but does not T-converge at x = 0.
F ( x ) = sinnx.
n
Then
(r-limF )(x) = - l ,
n
whereas F n does not converge pointwise.

6.1 The definition ofT-convergence 227
From Examples 6.1.4 and 6.1.5, we see that among the two notions of
pointwise convergence and T-convergence, neither one implies the other.
Example 6.1.6. F : X E converges continuously to F : X E if for

n
every x G X and every neighbourhood V of F(x) in E (i.e. V = {y G

E : | F ( x ) - i / | < e} for some e > 0 in case F(x) G E, V = {y G E :
y > i f } U { o o } for some K G E in case F(x) = oo, and analogously for
F(x) = oo), there exist no G N and a neighbourhood U of x w i t h
F (i/) G V
n
for all n > no, y U.

F converges continuously if and only if both F and F converge
n n n
to F and F, respectively. Continuous convergence implies pointwise

convergence, and we conclude from Examples 6.1.2 and 6.1.3 that T-
convergence is weaker than continuous convergence.
Example 6.1.7. Let X satisfy the first axiom of countability,
F n EE F : X -> E
a constant sequence. Then
r-limF = n (sc~F)
is the relaxed function of F . Thus, we have the remarkable phenomenon

that a constant sequence may converge to a limit different from the
constant sequence element.
Remark 6.1.1. Without changing the content of the definition of T-

convergence, condition (ii) may be replaced by the following condition
which is weaker and therefore easier to verify:
(ii') for every x G X , there exists a sequence x n converging to x w i t h
l i m s u p F ( x ) < F(x).
n n
noo
The following result is useful in approximation arguments:
L e m m a 6.1.1. Let X satisfy the first axiom of countability. Suppose

x
( m)meN converges to x in X, and
l i m s u p F ( x ) < F{x).
m
moo
Suppose that (ii') is satisfied for every x (i.e. for every m, there exists
m
a sequence {x ,n)neN
m converging to x with m
lim sup F ( x , ) <

n m n F(x )).
m
noo
228 T-convergence
Then (ii') also holds for x.
Proof. Since X satisfies the first axiom of countability, we may take a

neighbourhood system (U ) ^ of x and renumber i t and take intersec
u u
tions so that
xm G U m for all m G N,
and that every sequence (y^^n with y G U^ ) for all v and some v u
sequence ^(y) oo as v oo converges to x. For n G N , we let
ran : = max jra G N : x , G U m n m , F (x n m ) n ) < -f F ( x ) j .

m
Then
lim m = oo. n
n+oo
Namely, otherwise, we would find fco G N with
Fn u O&fc.uJ > ~ +F(x ) k
or
%k,n Uk u for all k > ko and some
sequence n oo for
z/ > oo.
To see that this is impossible we simply observe that since x ko G Uk0
and since Xk converges to Xk as n oo we have

0>n 0
r a
^fc ,n 4
0 0 f ^ sufficiently large n,
and likewise since we assume
limsupF (x n f c o > n ) < F(x ), fco
71 OO
we have
Fn{%k n) < F(xk )

0 0 + 7- for all sufficiently large n .
We then have
%m n ,n G f/m n
1
F (x
n m n , ) < ÔmJ +
n
Therefore y n := x m n y n converges to x as n oo, and
l i m s u p F ( t / ) < lim sup ( F(x )

n n mn + J < F(x)
n+oo n+oo \ rn n J
6.1 The definition ofT-convergence 229
by assumption and since m n oo as n > oo. Thus, (i/n)neN is the

desired sequence.
q.e.d.
Let F : X -> R U { 0 0 } satisfy
inf F(t/) > - 0 0 .
Given e > 0, we say that x G X is an e-minimizer of X i f
F ( x ) < inf +
Note that x is a minimizer of F i f i t is an e-minimizer for every e > 0.

In contrast to minimizers, e-minimizers always exist for any e > 0.
The following result is a trivial consequence of the definition of T-
convergence, but quite important.
T h e o r e m 6 . 1 . 1 . (Let X satisfy the first axiom of countability). Let the

sequence of functions F : X > R F-converge to F : X > R.
n
Let i n f x F (y) > 00 for every n G N .

y n
Let x be an e -minimizer for F .

n n n
Assume e > 0 and x > x for some x G X. Then x is a minimizer

n n
for Fj and
F(x) = l i m F (x ).
n n (6.1.1)
n-+oo
Proof. I f x were not a minimizer for F , there would exist x' G X w i t h
F{x') < F(x). (6.1.2)
Since F n T-converges to F , there exists a sequence (x' ) C X with n
lim x' = x'

n
,
\imF (x )=F(x').
n n
We put 6 := \(F(x) - F(x')). We may choose n so large that
e <6
n (6.1.3)
F (x' )<F(x')
n n +6 (6.1.4)
^n(^n) > F(x) - 6 (by property (i) of Definition 6.1.1). (6.1.5)

230 T-convergence
Since x n is an e -minimizer of F ,
n n
F )
n > F (x )n n - e n (6.1.6)
>F (x )-S
n n by (6.1.3)
>F(x)-2 by (6.1.5).
From (6.1.4) and (6.1.6), we get
F(x) < F(x') + 36
contradicting (6.1.2) by definition of 8. Thus, x is a minimizer for F . I f

(6.1.1) did not hold, then after selection of a subsequence,
F(x) < limF (x ) n n
whereas by property (ii) of Definition 6.1.1, there would exist a sequence

(x' ) converging to x with
n
F(x) = l i m F ( x ' ) , n n
and we would again contradict the e -minimizing property of x . n n
q.e.d.
Corollary 6.1.1. (Let X satisfy the first axiom of countability.) Let

F : X R r-converge to F : X R. Let x be a minimizer for F . If
n n n
x x, then x minimizes F, and

n
F(x) = lim inf F (x ). n n
The following result is similarly both trivial and important.
T h e o r e m 6.1.2. (Let X satisfy the first axiom of countability.) Let F n
Y-converge to F. Then F is lower semicontinuous.
Proof. Otherwise, there exist some x G l and some sequence ( x ) ^ N m m
with
lim Xm = x
m*oo
lim F{x ) m < F{x). (6.1.7)

moo
By T-convergence, for every m , there exists a sequence ( x , ) m n n G N C X

with
lim x m n x m
n>oo
lim F ( x m > n ) = F(x ). m (6.1.8)

We assume oo < l i m F ( x ) , F(x) < oo simply to avoid case distinc

m
tions. We let
6 := ] (F(x) - l i m F(x j)
m >0 by (6.1.7).
4 \ ra+00 /
For every ra G N , we may find n m G N with
Fn (x n )-F(x )<6
m mt m m (6.1.9)
lim x , n m m = x , lim n m = oo.

771 oo raoo
Then by T-convergence
F(x)< lim i n f F n m (x ,
m n m ). (6.1.10)
ra>oo
We may then choose ra so large that
F(x ) m <F(x)-36 (6.1.11)
and
F (x , J
nm m n > F(x) - 6. (6.1.12)
Equations (6.1.9), (6.1.11) and (6.1.12) are not compatible, and the re
sulting contradiction proves the lower semicontinuity.
q.e.d.
Remark. As a consequence of Corollary 3.2.2 and Theorems 3.1.3, 3.3.1,

p
and 3.3.3, in combination w i t h Lemma 2.2.4, the weak topology of L (ft)
kyP
and W (Q) for 1 < p < oo satisfies the first axiom of countability so
that the preceding notions are applicable.
The reference for this section is

G. dal Maso, An Introduction to V-Convergence, Birkhauser, Boston,
1993
6.2 Homogenization
In this section and the next one, we describe two important examples
of T-convergence. They are taken from H . Attouch, Variational Conver
gence for Functions and Operators, Pitman, Boston, 1984.
In the discussion of these two examples, we shall be more sketchy
about some technical details than in the rest of the book, because
the main point of these examples is to show how the concept of T-
convergence can be usefully applied to concrete problems that arise in
various applications of the calculus of variations.
232 T-convergence
d d
Let M be a smooth subset of the open unit cube (0, l) of R . M is
considered as a hole. Let
M := c ( J e ( M + ra)
d
mGZ
(e(M + 777.) : = { x = y + em with ^ G M}) be a periodic lattice of 'holes'
of scale .
d
Let ft C R , fi : = ft \ ( M f l ft), i.e. a domain with many small
c c
holes. Such domains occur in many physical problems like crushed ice,
porous media etc. Often, the physical value of e is so small that it is
useful to perform the mathematical analysis for e 0. This is called
homogenization. Let
d
, x / x f0 forxGM \Mi
^ oo for x G M i
d
be the indicator function of R \ M\. a ( ~ ) then is the indicator function
d
of R \M . eWe consider the functional
2 2 2
F (u):=\e
e [ | D u ( x ) | d x + / a (-) u (x)dx (6.2.1)
2 V e
Jn Jn '
2
for u G # o ' ( n ) . A minimizer of the functional
F (u)e / f(x)u(x)dx
Jn
2
(for given / G ( f i ) ) satisfies
A u = - ~ in fi and u = 0 on dft . c e (6.2.2)
Here d f i = d f i U (dM f)ft).

c The boundary condition on dft comes from
e
2
the requirement that u G jfiFo' (fi), while the boundary condition on dM e
is forced by the functional.

2
T h e o r e m 6.2.1. With respect to weak L (ft) convergence
r - l i m F = F, c (6.2.3)
with
where
2
fi(M):= j \Drj{x)\ dx= [ r/(x)dx,
d d
J(o,i) \M J(o,i)
and 77 is the solution of

d
AT? = - 1 m(0,l) \M
77 = 0 inM (6.2.4)
d d d
77 25 Z ~periodic (i.e. rj(x + ra) = 77(0;) /or x G (0, l ) , ra G Z ) .
Proof. We put 77 (x) : = 77(f). By Lemma 5.2.1, 77 converges weakly in

e e
2 2
L ( f i ) to fJ>(M) as e 0. Let now u G L (ft). By approximation, we
/1 2
may assume that u is smooth, e.g. contained in W ' ( f i ) f l C ( Q ) . We
put
2
Then u converges weakly in L (Q)
c to u, and
u 0
e on M . c
Moreover
^ e K ) = ~ / l ^ c l 2
(6.2.5)
2
^2
l^jf 2
2
(u \Drif + 2ur Du-Dri ]e e + 2 2
r \Du\ )
1
2 /x(M)
2
I f C/ C fl is open, because of the periodicity, J V |-DT7 | asymptotically
c
behaves like
meas U f .^ ,2 meas U
d d
J(0,e) \M e J(0,l) \M
This means that

2 2 2
lim e / \Drj \ e = messU [ \Dn\ = measU fi(M) (6.2.6)
c 0 d
~> JU J(0,l) \M
hence, approximating u by step functions, we also get
2 2 2 2
lime / u \Drj \ e = fi(M) [ u (6.2.7)
(note that we assume u to be continuous). Moreover, since % is bounded

independently of e,
2 2 2
lime / 77 |>u| = 0, (6.2.8)
and from (6.2.6), (6.2.7) and the Schwarz inequality, also

2
lim e / un Du >77 = 0.
e c (6.2.9)
C
-+ JQ
234 r-convergence
Equations (6.2.5)-(6.2.9) imply
lim F (u ) e e = F(u). (6.2.10)
In order to complete the proof of T-convergence, we need to verify that

2
whenever functions v that vanish on M converge weakly in L (Q) to
e
tx, then
l i m i n f F (v ) e e > F(u). (6.2.11)
By an approximation argument, we may assume u G Co(fi). We put
as before. We have
2
F (v )
e e + F (u )
e >e [ Dv Du e e
f (m
2
e !
V(M) w*Dve'Du + uDv -Dri ).
(6.2.12)
Using (6.2.10), we obtain from (6.2.12) in the limit e -+ 0
2
UminfF (t; ) + I r j r
e e f u > lim inf - 1 / uDv Dr] e e (6.2.13)
since the other term on the right hand side of (6.2.12) goes to 0 by a
similar reasoning as above. Equation (6.2.4) implies
2
e Ar) e = -1 in fi . c (6.2.14)
Moreover
6 2
Duv D e Vt <e 2
(^j \Du\ v )j ' (^J 2 2
\D \ ^Ve
2
' - 0, (6.2.15)
2
since v as a weakly converging sequence is bounded in L , \Du\ is
e
bounded by our approximation assumption that u is smooth enough,

and since we may use (6.2.6).
Integrating the right-hand side of (6.2.13) by parts, and using (6.2.14)
and (6.2.15), we obtain
u 2
=
M M ) i '
6.3 Thin insulating layers 235
2
since v converges weakly in L to u. This implies (6.2.11) and concludes
e
the proof.
q.e.d.
6.3 T h i n insulating layers

We consider an insulating layer of width 2e and conductivity A, and we
want to analyse the limit where e and A tend to 0.
3
Let fi C R be bounded and open, S a smooth complete surface i n
3
R , e.g. a plane, S : = fl f l 5,
3
S : = {x R : dist(rr, 5 ) < e}
e
S c : = fl f l S e
fl c :=fl\S . c
Conductivity coefficient
_ f 1 on fl c
a c A :
' ~\A onE c (A>0).
Variational problem:
C A 1 , 2
J ' :# 0 ( f l ) -> R
2 2
r*(u):=\f \Du\ dx+~f \Du\ dx (6.3.1)
2 2
JQ / c
c A 2
J ' (u) - / fu -> min (/ L (fl) given).
The Euler-Lagrange equations are
Au eA + / = 0 on flc (6.3.2)
AAu C ) A + / = 0 on S c (6.3.3)
W ,A|Q
C C = W ,A|E
C c on d f l f l d c c (6.3.4)
A on d f l f l d
c c (6.3.5)
dux*
(where nu denotes the exterior normal of a set U)
u iX = 0 on d f l . (6.3.6)
236 r -convergence
T h e o r e m 6 . 3 . 1 . We let e -> 0, \ -+ 0. If j -> a with 0 < a < oo, then

2
Ue,x * u weakly in L ( f i ) u \ =t u uniformly on every fl C C fi \ E,
et 0
w/iere w solves
Aw + / = 0 on fl \ E
w|an = 0
^p
and
= â n | = H E
2
w/iere is the jump ofu across E, and and ore the exterior
normal derivatives for the two components of ft \ E. (In case a = oo, u
is continuous across E, and! A u = / in fi.) Furthermore
2 2
\ f lê,A| + ^ / |ê,A|
2
/ | D u | + ^ / [u]|dE.
e , A 2
J r-converges w.r.t. the weak L -topology to I(u):
e 2
0 < a < oo : I(u) = ( J / n \ l ^ f + f # " o ' ( " \ E)
I oo otherwise
2
a = oo:I(u)=iUn\Du\ ifuZH^{Sl)
I oo otherwise
(in this case,
the result holds
2 2
a = Q : 1 { u ) = / \ J n V E \Du\ if u e H^ (Q \ E) for the strong
L
loo otherwise -topology
in place of
the weak one)
Thus, in case a = 0, we obtain a perfect insulation in the limit,
whereas for a = oo, the limiting layer does not insulate at all.
3
We assume for simplicity S = {x = 0 } .
L e m m a 6 . 3 . 1 . There exists a constant c\ (depending on / , fl, 5, but not

on , A) such that for all sufficiently small e, A
Proof.
2 2
[ \Du ,\\
e + \ f \Bu , \ e x
n
Jn e JdQ n Jz
+ A / W C , A ^ ~ -
= / / *W A C ) because of the Euler-Lagrange equations
2
^ l/L (Q)' M
l .Alx,2( ) N
By the Poincare inequality (Theorem 3.4.2),
2
/ < A <C2 [ \Du , \ e x
2
[ < A <C3 / | ^ C , A | .
By a change of scale
3 3 1 2 2
y = ex (y = x\y = x ),
2
( < A <c e 3 [ \Du \ yX
2
(we only get e instead of e , because the area of the portion of # E on C
which U A vanishes, namely Oil f l 5 , is proportional to e). Altogether

C ) C
2 2
< A < C 4 ( l + ^) (jf \Du , \ e X + \J^ |> W e ,A| )
and the estimates follow.

q.e.d.
Proof (Theorem 6.3.1). We only consider the case 0 < a < oo (the
other cases follow from a limiting argument). We first observe
c A 2 2
T- lim J ' ( u ) = oo i f u e L ( f i ) \ H^ (Q \ E).
We assume for simplicity

c c A ( c
A = A() = a , J : = J ' ).
1,2
Let u e H (fi \ E). We first check property (ii) of T-convergence:
238 T-convergence
x 2 c
We need to find u e u weakly in L ( f i ) w i t h l i m / ( u ) = I(u). c We
define
1 2 3 3
[ ^(a: ,^ ,^ ) if | z | > e
1 2 1 2
u (ar,ar,ar) := < i{u(x ,x ,e)-fw(a: ,a: ,-e)}

3
3
if rc <.
x
1 2 1 2
+{u(x ,x ,e)-u(x ,x ,-e)}
2 2
Then u # o ' ( f i \ E), w w weakly in L ( f i ) for c -> 0,
e
J(ti ) e
2 2 2
= 5 / \Du\ + ^ [ \\D(u(z\x ,e)+u(x\x ,-e))
2
* JQ J\x*\<e \*
/ 3 \ ^
+D ( y (u(x\x ,e) 2
- u(x\x ,-e))J 2
2 2 2 2
- I I \Du\ + ?L [ \D(x*(u(x\x ,e)-u(x\x ,-e)))\
D u 2 2 2 2
~\l \ \ + rf Hx\x ,e)-u(z\x ,-e)\
z 3
-f terms that contain x and go to zero as e > 0 ( | # | < e).
If u is smooth (which we may assume by an approximation argument),
therefore for e > 0
2
i2 / |D| + f /[]
JQ\E 4
/E
We now check property (i) of T-convergence:

2
Let v * u weakly in L ( f i ) . We need to show
e
c
liminf/ (t; )>/(u). c
e>0
For u as above,
e
2 2
~ f \Dv \ e + ^ [ \Du \ >aef e Du e Dv (
vE e 'Eg </ E e
> f / e (!>{u(.,6)+u(.,-6)}
+ ) { ( - , e) - ( - , - e ) }
+^(u(., )-u(.,- )) e e
3
where 3 of course is the unit vector in the x direction.
We may assume u smooth (otherwise, we use an approximation argu

ment). Then as above
3
> lim inf / (D e) - -c))) D f z e
+ l i cm inf ^ / e 3 D?J c e) - -c)).

-^ 2 J E E
c
Without loss of generality l i m i n f _ ^ J (ê) < oo. Then e 0
2
supe / \Dv \e < oo. (6.3.7)
Consequently,
c a e Dv
limae J Ql)w(-,c) + ^ > w ( - , - ) ^ >v < (^J \ *
cae
-+ 0 for e -+ 0.
Similarly,
3
lim sup a / (D (w(-, e) - it(-, - e ) ) ) Dt; x c -* 0
3
since | # | < e. Thus
> lim inf - / e >v

3 c) - -c)).
3
Since u(-, c) and u(-, e) do not depend on x , we obtain by integration
^ / e -/?v (w(-,c)-w(-,-c))
3
= 77 / v
e ) - W(., - ) ) - ^ / V ( u ( . , C) -
C -c)) ,
3
where here of course dUf = 0 0 {x = e}. Since we may assume
c ly2
l i m i n f _ o ^ (^c) < oo, v is bounded in H (Q ).
c e Therefore, we may e
240 r-convergence
-
assume that the traces of v on d converge*) . Since u is assumed smooth
e c
2
and v converges to u weakly in L , we may assume
e

^,(0E ) e - w(',0 ) weakly in L ( d ) . 2
We then get
h
e ^2 / ^Dv (u(-,e)-u(-,-e))
{ = - / []..
J Eg J E
Altogether
, i m < iff/ _| v,|' E D + ^ / c M | > 2 / M | .
Therefore
2
liminfrK)>i / \Du\ + ^ f [u)l.
q.e.d.
Exercises
6.1 Determine the T-limits of the following sequences of functions
F (x)
n := n(smnx + 1)
for a: = 0
n z 2
for 0 < x < -
F () : - | ^
n Q f^ I
^ ^
^ n2
2
2n - n x for < x < ~
n n
F (x) : = sinnrr + cosnrr.
n
6.2 Show the following result: Let X be a topological space satisfy

ing the first axiom of countability, F , G : X R. Suppose n n
that F T-converges to F , G T-converges to G, F - f G T-

n n n n
converges to H (assume that the sums F + G , F + G are n n
always well defined; for example, there must not exist x X

w i t h F(x) = oo, G(x) = - o o or vice versa). Then
F + G < H.
Does one get equality instead of ' < ' here? (Hint: Consider
F (x) sinnrr, G (x) = - s i n n r r . )
n n
f For this technical point, see e.g. W . Ziemer, Weakly Differentiable Functions,
Springer, G T M 120, New York, 1989, pp. 189ff.
7
BV-functionals and T-convergence:
the example of Modica and Mortola
7.1 T h e space BV(Q)

d d
Let Co(M ) be the space of continuous functions on R with compact
support. For each Radon measure fi and each //-measurable function
d
v : R R with = 1 //-almost everywhere, we can form a linear
functional
d
L : C$(R ) -* R
L(f) = f fudpi.
d
JR
Conversely, we have the Riesz representation theorem, given here with
out proof (see e.g. N . Dunford, J. Schwartz, Linear Operators, Vol. I ,
Interscience, New York, 1958, p. 265).
d
T h e o r e m 7.1.1. Let L : Co(M ) R be a linear functional with
d
\\L\\K : = s u p { L ( / ) : / C ( E ) , | / | < l , s u p p / C K] < oo
0 (7.1.1)
d
for each compact K C R . Then there exist a Radon measure ii on
d d
R and a ji-measurable function v : R R with \u\ = 1 ji-almost
everywhere with
d
L(f)= f fisdfi forallfC(R ). (7.1.2)
If L is nonnegative, i.e. L(f) > 0 whenever f > 0 everywhere, then

v = 1, i.e.
L(f) = / /rf/i. (7.1.3)
241
242 Modica-Mortola example
d
Thus, the Radon measures on R are precisely the nonnegative linear
d
functionals on C o ( R ) . (Note that (7.1.1) automatically holds i f L is
nonnegative; namely
\\L\\ K = L( ) XK
in that case where XK is the characteristic function of K.)

d
The same result more generally holds for C o ( R , # ) where H is a
finite dimensional Hilbert space with scalar product ( , ) . Then linear
d
functionals L : C g ( M , # ) R satisfying (7.1.1) are represented as
L(f)= I (f,")dn, (7.1.4)

d
JR
d
where // again is a Radon measure and v : R H is //-measurable with
|z/| = 1 //-almost everywhere. Also, in the situation of Theorem 7.1.1,
one has
/x(n) = s u p { L ( / ) : / G C (fi), | / | < 1} 0
d
for any open ft C E .
The expression z/d// in (7.1.4) = 1 //-almost everywhere) is called
a vector-valued signed measure. (// is supposed to be a Radon measure
and v a //-measurable function with values in H.)
d
D e f i n i t i o n 7 . 1 . 1 . Let ft R be open. The space BV(ft) consists of
1
all functions u L (ft) for which there exists a vector-valued signed
measure z/// with //(fi) < oo and
udivg = j gvdfi (7.1.5)

L
d
for all g C o ( f i , R ) . In this case, we write Du = z///, DiU = z/^//
(y = ( z / . . . , i / ) , i = 1 , . . . , d ) . For u BV(ft),
l5 rf we put
\\Du\\(n) :=// =
d G d
sup { / udivgdx
n :g = (g\...,g )C% ( f i , R ), \g(x)\ < 1
for all x fi}
< oo
and
n
W \\ v(n) :=IHlLt(n) +
u
B IPll()-
7.1 The space BV{SI) 243
For u B V ( f i ) , \\Du\\ is a Radon measure on f l :
| | D u | | (fio) = sup I j f udivgdx : g G C (fto,M ), M < 1 J

0
d
for fto open in fi. We write
||Z>||(no)=: / ll^ll,
/Oo
Jftn
and also
||Du||(/)=: / /||Du|| for a nonnegative Borel

n
measurable function /
on ft.
We have for / Cg(fi), / > 0:
\\Du\\(f)
= su yûdiyg:gC (n,R ),\g(x)\<f(x)

P 0
x d
V x fij . (7.1.6)
L e m m a 7.1.1. If u Wîft), then u BV(fl), and
dfi = dx where Du is the weak

derivative ofu anddx
is d-dimensional Lebesgue
measure,
and
D U { X )
ifDu(x)^0
u(x) = { \Du(x)
0 otherwise.
The proof is obvious. q.e.d.
d
On a compact hypersurface S CR oi class C , we have an induced
metric and in particular a volume form dS. The (d l)-dimensional
volume of 5 then is
|Sld-i = / dS.
Js
d
L e m m a 7.1.2. Let E be a bounded open set in R with a boundary dE
of class C. Then
1 ^ 1 ^ = 1 1 ^ 1 1 ^ ) , (7.1.7)
where \ E is the characteristic function of E.

Proof. We have to show
\dE\ _d x = sup | ^ d i v 5 : 5 C ( R , R ) , \g\ < 1 J .

0
d d
By the Gauss theorem
[ dlvg= [ g(x)n(x)d(dE)
JE JdE
where n(x) is the exterior normal. Therefore
\dE\ _d t >sa y P divg:geCZ>(R ,R ),\g\ d d

< l j .
For the converse inequality, we use a partition of unity to extend n to a

d d
C-vector field V on R w i t h | V ( x ) | < 1 for all x R . For sup|y ipd(dE) :y>CS(R ).M < l j d
This completes the proof.

q.e.d.
The same conclusion holds if J? C C for some bounded open set;

namely
IdElî = \\D \\(0)

XE = sup [Je^9 9 C 0 ( n , R ) , |<?| < 1 J
d
in that case.
d
D e f i n i t i o n 7.1.2. A Borel set c R has finite perimeter in an open
set fi if X E \ BV(Q).
N The perimeter of E in fi m that case is
P(E,Q) := \\DxE\m
d
( = sup|^div f f : C (n,R ),| | < l | ) .
5 0 5 (7.1.8)
d
E is a set of finite perimeter if XE S V ( R ) .
7.1 The space BV(Q) 245
The following lower semicontinuity result is easy to prove and very

useful.
d
T h e o r e m 7.1.2. Let ft C R be open, ( u ) e N C BV(ft), n n and suppose
l
u n u in L (ft).
Then for every open U C ft
\\Du\\(U) < lim inf \\Du \\(U). n (7.1.9)

noo
// in addition
sup {\\Du \\(fi) n : n N} < oo, (7.1.10)
then
u BV(Q).
d
Proof. Let g C${U,R ) w i t h \g\ < 1. Then
/ udivg = l i m / u div n g < l i m i n f | | D u | | (U).

n
JU n-ôc JJJ n-ôo
Taking the supremum over all such g, we obtain (7.1.9). I f (f Co(Q),

then for i = 1 , . . . d
lim / (fDiU n = - l i m / u Di(p = - n uDup

n + n
~ i n -* i n in
and hence
uDiip < sup\<p\ l i m i n f | | D w | | (ft) < oo n
i/ nn
in case (7.1.10) holds. Since C%{ft) is dense in C(f2), for i = 1 , . . . , d
Diu((p) : = - / uDup,
in
then is a bounded linear functional on CQ(Q), and thus u BV(ft).

q.e.d.
We next discuss the approximation of BV-functionals by smooth ones

d
through mollification. As usually, we let p C o ( R ) by a mollifier
w i t h p > 0, suppp C S(0,1), f p(x)dx = 1, and we also impose the
Rd
symmetry condition
p{x) = p{-x). (7.1.11)

We then put as in Section 3.2
d
p {x)
h := h- p{j^
l l d
and for u L (fl), we extend u to L (R ) by defining u(x) = 0 for
d
x R \ fl and put
u (x)
h : = p^ * u(x) : = / p^(x - y)u(y)dy C(Q).
T h e o r e m 7.1.3. If u BV(ft), then u h ~+u in L ^ f i ) and \\Du \\ h ->

\\Du\\ in the sense of Radon measures as h * 0, i.e. for every f C (ft)
lim / f\\Du \\^ h f f\\Du\\. (7.1.12)
In particular,
Um \\Du \\ h (SI) = \\Du\\ (SI). (7.1.13)

h()
Proof. Uh u in L (ft) l
by Theorem 3.2.1. I t suffices to consider the
case / > 0. Prom (7.1.3) i t follows as in the proof of Theorem (7.1.2)
that for every / C (fi) with / > 0
/ / | | D u | | < l i m i n f / f\\Du \\. h (7.1.14)

H
JQ ~* JQ
I t thus remains to prove that for such /
limsup / f\\Du \\< h f f\\Du\\. (7.1.15)

h~+o JQ JQ
For that purpose, we first obtain from (7.1.6)
/ f\\Du \\ h =
Ja
s u p j y g(x)Du (x)dx
h : g Cf(Sl,R),\g(x)\ < f(x) V x fij .
(7.1.16)
Here, Dun = ( g f r w ^ , . . . , ^ J U A ) is the gradient of u/,, since UH is smooth.
7.1 The space BV(ft) 247
For any such g as in (7.1.16)
/ g(x)Duh(x)dx = / Uh(x) divg(x)dx

Jn J
= - J J p (x - y)u(y)dy
h divg(x)dx
= -j J p (y - x) div g(x)dx u(y)dy

h by (7.1.11)
= - j u(y)div(g )(y)dy. h (7.1.17)
Since we assume \g\ < / , we have

\9h\ < \g\ < h A ,
and since / is continuous, fh =3 / uniformly as h 0 (see Lemma

3.2.2), i.e. \f (x) - f(x)\ < r) for all x fi, with \\m ^r]
h h = 0. By h h
definition of the right hand side of (7.1.17) therefore is bounded

b y / ( / + fr)l|0ll-
n
Thus, for every such g
lim / g(x)Duh{x)dx < / /||Du||,

hhJn Jn
and (7.1.15) follows (cf. (7.1.16)).
q.e.d.
d
Corollary 7.1.1. Let ft be a bounded, open subset of R . Then any
sequence (u ) ^ C BV(ft)
n ne with
or s o m e
U
\\ TI\\BV - K f K
1
contains a subsequence that converges in L (ft) to some u BV(ft) with
\ H B V < K -
Proof. By Theorem 7.1.3, there exist functions v n C(fi) with
\\Un ~ ^ n | | i ( n ) <L n
\\Dv \\(Q)<K
n + l.
lyl
Therefore ( f ) n N is bounded in W (ft).
n By the Rellich-Kondrachev
compactness theorem 3.4.1, after selection of a subsequence, ( f ) n G N n
1 1
converges in L (ft) to some u L (ft). (u ) has to converge to u as well
n
x
(in L ( f i ) ) . By Theorem 7.1.1, u BV{Q), and
\H\ <K.
BV
q.e.d.
A reference for the BV theory is W . Ziemer, Weakly Differentiable

Functions, Springer, G T M 120, New York, 1989, Chapter 5.
7.2 T h e e x a m p l e o f M o d i c a - M o r t o l a
We now come to the theorem of Modica-Mortola:
T h e o r e m 7.2.1. Let
v oo otherwise,
] D U l = l | Z ? U | 1 U B V ( R d )
F{u) := { t L *
\ oo otherwise.
1 d
Then w.r.t. to L (M ) convergence
F = T- l i m F . n (7.2.1)
Proof.
(i) We first want to show
F(u) < lim inf F (u ) n n (7.2.2)

n*oo
whenever
l d
u ~+u
n \nL (R ).
For that purpose, we put

nt
1 f
h (t) := - /
n |sin(7rr)| dr.
n Jo
We note that
\h (s)
n - h {t)\
n <\s-t\ for all n G N , 5, t G .
Therefore
H^n Un~- h o u\\ n L1 < \\u - u\\

n L1 0 as n oo. (7.2.3)
Also
lim h (t) n = (7.2.4)

We now obtain
, 2 , 2
< ||/i ou -/i ou|| -f h ou u
nn ou n u n n n L 1
n
1
L 1
L
7T 7T
0 as n oo (7.2.5)
by (7.2.3), (7.2.4), and Lebesgue's Theorem 1.2.3 on dominated

convergence. We may assume
h2 d
un G H {R ) for every n G N , (7.2.6)
because otherwise F (u ) n n = oo, and (7.2.2) is trivial. Then
l i m i n f F (u )
n n > 2 lim inf / \Du \ |sin(7rnw )|
n n
n+oo n>oo J^d
= 21iminf / \D(h ou )\ n n
noo J
> - f \Du\ by (7.2.5) and Theorem 7.1.2
= F(u).
This shows (7.2.2).
x d
(ii) We want to show that for every u G L ( E ) , there exists a se
1 d x d
quence (tXn)nGN C L (R ) converging to u in L ( R ) w i t h
l i m s u p F n K ) < F(u), (7.2.7)

noo
thereby completing the proof of T-convergence. This inequality

will be much harder to show than (7.2.2), however. We shall pro
ceed in several steps:
d
(1) We may assume u G C o ( E ) . By a slight extension of the
d
reasoning of Theorem 7.1.3, we may find Uh G C(R )
(take a smooth (fh w i t h (fh = 1 on B(0, ^ ) , ip(h) = 0 on
d
R \ JB(0, + 1), \D(f \ < 2 and multiply the mollification h
of u with parameter h by tp^) with
lim / \uh(x) - u(x)\ dx = 0
lim F(u ) h = F(u).

hO
d
Applying Lemma 6.1.1, we may indeed assume u G Co(R ).
(2) We now want to show that i t suffices to verify the claim
for certain step functions.
Modica-Mortola example
d
By (1), we assume u e <7(R ). By Sard's theorem, for
almost all t G l ,
l
u~ (t) = {x : u(x) = t}
is a hypersurface of class C . For every v G Z, n G N , we

may then choose t , w i t h this property, w i t h
v n
v v+l
n n
and satisfying
The coarea formula (Theorem A . l ) then implies
/ \Du(x)\dx^ I [u^^l^dt
d
Ju Ju
oo rl
* E / " l ^ w L v
J
l / = - 0 0 n
OO j
u _ 1
* E ^l (^)L-i
v oo
oo 1
=
2 D
^ l l X { t i > t , , } | | by Lemma 7.1.2.
n
t>= oo
We choose iV(n) G N with iV(n) > (nmax \u\ + 1) and put
N(n) 1 ^
t/=-AT(n)
The preceding inequality implies
limsupF(u ) < F(u). n
n>oo
If t ,v n < u(x) < i / , + i , then u ( x ) = . Therefore

n n
n
s u p p u C suppu, n
and for all x

2
\u(x) - u (x)\ n < .
n
Since u is assumed to have compact support, therefore
lim / \u (x) n u(x)\dx = 0.
Lemma 6.1.1 then implies that i t suffices to prove the claim

for the functions u . n
(3) I n (2), we have reduced the claim to step functions

N
2=1
where the fti are disjoint bounded open sets with bound
ary dfti of class C. Since the general case is completely
analogous, for simplicity, we only consider the case N = 1,
i.e.
u = an X (7.2.8)
with ft bounded and dft of class C . Thus
F(u) = |an| d - 1 (cf. Lemma 7.1.1). (7.2.9)
We let 0 < p < o, where eo is given in Lemma B . l . Thus,

the signed distance function d(x) as defined i n Appendix
B is smooth on {x G R : dist(x, dft) < p). We need the
d
following auxiliary result:
L e m m a 7.2.1. Let n G N , let a n G R, with lim _>oo a =

n n
Q G R , na n G Z ,
2
4>n{x) : = fjy ^ + nsin (7mx())} *
6e /ie one-dimensional analogue of F . Then there exist n
Lipschitz functions \n R R
XnW = 0 /orKO
XnW = n for t> -~=

\/n
0< n(t)<a
X n / o r O < K - f ,
and
4
l i m s u p 0 ( x n ) < -ex.
n (7.2.10)
n+oo fl"
Modica-Mortola example
We postpone the proof of Lemma 7.2.1 and proceed with

the proof of the theorem. We choose a sequence
with
lim a n = a,
noo
and na n Z as in Lemma 7.2.1. We put
fin := jxefi:d(x)<--^=J
and
w (a:) : = Xn(d(x))
n w i t h %n as in Lemma 7.2.1.
(7.2.11)
Then
d
u (x)
n = 0 for xeR \fi
(x) = a
n n for X G fi \ fi n
0 < u (x) n <a n for x e fi - n
We also note
lim I f i J . = 0. (7.2.12)
Thus (cf. (7.2.8))
lim / |u(x) - u ( x ) | dx = 0,
n (7.2.13)
1
and u n converges to u in L . We also let (as in Appendix
B)
d
E : = {x R : d(x) = t}.
t
We note
d
Du (x)
n = 0 , sin(n7n/ (z)) = 0 n for x e R \ fi ,
n
(7.2.14)
and
|Dd(x)| = 1 for x fi n by Lemma B . l . (7.2.15)

Then
lim sup F ( u )
n n
l D U n { x ) l 2
= limsup / ( + nsm (n7ru (x)) n \ \Dd{x)\dx
l i m s u p ^ ^Xn(*)l 2
+ n s i n 2( n 7 r X w ( )) j
t
>
rft
by Corollary B I (coarea formula)
< limsup sup 0 (Xn)*|Et|d_i J

n
n->oo y)<t<^= y
4
< -a by Lemma 7.2.1 and Lemma B . l
= F(u) (cf. (7.2.9)).
This is (7.2.7).
(4) I t only remains to prove Lemma 7.2.1:
The idea is of course to minimize 4> {x) under the given n
side conditions on x- The Euler-Lagrange equations for <fi n
are
=
^x" 7rnsin(7rnx) cos(7rnx),
and these are implied by
2 2
^ X ' = s i n ( 7 m x ) + ci. (7.2.16)
We now construct a solution of (7.2.16) with the desired

properties: w.l.o.g. a > 0 (the case a < 0 is analogous).
We choose ci = in (7.2.16). We put
i
2
*n(0 : = / ~ I 1 ^ r I
2
ds
7o n ^ +sm (n7T5) y
Vn : = </>n(c*n).
Then
0 < rjn < A=a . n
We let
Xn : [0, Vn] -+ [0, a ]
1
be the inverse of i/j . Then x-n is of class C
n and
-Xn (0 = ^ - + sin>7r (*))J

Xn . (7.2.17)
We extend Xn to R as a Lipschitz function by putting
Xn(t) = 0 for t < 0
Xn(0 = n for t > rj . n
Then
<MXn)
+ nsin^(7rnxn(0) I ^
2
^ v ' (f> / 1
X b 1 2
' - + n ( - +sin (7rn nW) X
n \n
i
2
n + sin (7rn n(0))' X n b y
X (7.2.17)
2
= 2 / f h sin (7rns) 1 ds
io V /
and Lemma 7.2.1 follows.
q.e.d.
References
L. Modica and St. Mortola, Un esempio di r~-convergenza, Boll. U.M.I. (5),
14-B (1977), 285-99.
L. Modica, The gradient theory of phase transitions and the minimal
interface criterion, Arch. Rat. Mech. Anal. 98 (1987), 123-42
Let us also quote without proof the following result of L . Modica,

loc. cit., which plays an important role in the theory of phase transitions:
d +
Let fl C R be open and bounded with Lipschitz boundary, W : R R
be continuous with precisely two zeroes a,/3 (which then are absolute
minima, because W is nonnegative)
a
F {u)
n := { In ( IP(*)H + nW(u(x))) dx for u e H{Q)
v oo otherwise
and
u
Fo(u) = { h 2 c
H^ ll for u e
BV(Q) and for almost all x G E
I oo otherwise
with
co= W2(s)ds.
Ja
1
Then F is the T-limit of F w.r.t L -convergence.
0 n
The proof is similar to the one of Theorem 7.2.1, except that we cannot
apply Sard's lemma anymore, because even for a smooth function u, a
and (3 need not be regular values. Thus, one has to consider nonsmooth
level sets as well and appeal to some general results about BV-functions
and sets of finite perimeter.
The interpretation of Modica's theorem is the following:
Consider first the problem
W(u(x)) dx min
L
under the constraint
1_ /
meas S 2
u(x) = 7,
with a < 7 < (3 (w.l.o.g. assume a < (3). A minimizer then is of the
form
f a for A\ C ft / p ? n l o ,
( ? 2 l 8 )
= for^Cfi - -
such that Ai U A<i = ft,
a meas A\ + /3 meas A 2 = 7 meas ft. (7.2.19)
u thus jumps from the value a to the value (3 along dAiilft = dA2C\fl =:
y
T. However, apart from the preceding relations (7.2.19), A\ and A2 and

hence also T are completely arbitrary. I n particular, Y may be very
irregular. I n order to gain some control over the transition hypersurface
2
T, one adds the the regularizing term f \ \Du(x)\\Q to the functional, al
beit with an arbitrarily small weight, and in fact one passes to the limit
where this weight vanishes so that one preserves (7.2.18), (7.2.19). A l
though this regularizing term disappears in the limit it still has the effect
of regularizing the hypersurface T along which the transition from a to
(3 occurs. Namely, the hypersurface of discontinuity of the minimizer u
now is constrained by the requirement that the B V norm of u, J \ \Du\\, n
be minimized. This means that T is a so-called minimal hypersurface.

The existence and regularity theory for such minimal hypersurfaces may
be found for example in E. Giusti, Minimal Surfaces and Functions of
Bounded Variation, Birkhauser, Boston 1984, pp. 3-134.
Exercises
d
7.1 Try to construct bounded sets in R that do not have a finite
perimeter.
7.2 Prove the preceding theorem of L . Modica for d = 1.
Appendix A
The coarea formula
T h e o r e m A . l (coarea formula for smooth functions).

d
Let u G C # ( R ) . Then by Sard's theorem,
d
C u := {t G R : 3x G R : Du(x) = 0, u(x) = t}
has one-dimensional Lebesgue measure zero, and thus, for almost all
1
t G R, w~" (0 is a smooth hypersurface by the implicit function theorem.
d
We then have for every open ft C R
/ \Du{x)\dx= f 1^(0 rifled*. (A.l)

Jn J-oo
Proof.
(1) We first show the result for a linear map

d
/ : R > R
d
(w.l.o.g. / ^ 0). Let 7r : R R be the projection onto the first
coordinate. We may find A G G/(1,R), R G 0 ( d , R ) f w i t h
I = A O 7T o R.
d
For every measurable subset E of R , we have by Fubini's theorem
\E\d = ^ {Enir-Wl^dt,
J OO
where
\E\ d = [ XE
t Gl(d,R) := {d x d-matrices A with real entries and det A / 0 } , 0(d,R) := {A

1 1
Gl(d,R) | A = A~ } (orthogonal group).
257
258 Appendix A
is the Lebesgue measure of E. Since R is orthogonal, we likewise

have
oo
lEniTWÎêft.
/ -oo
We then change variables via s = At and obtain
oo
lEnR-'on-'oA'^s^ds
/ -oo
{Enr'is^ds. (A.2)
Since \A\ = |dZ|, and / is linear, this is the coarea formula for
linear maps.
(2) Let
d
S u = {x G R : Du(x) = 0}
d
U := {x G R
t : u(x) > t} for t G R.
We put
f Xt/ t at > 0
U t
" 1 -Xn*\c/ t i f * < 0.
Then
u(x) = / u (x)dt.
t
JR
d
Let G Q ( R \ S ), u \(p\ < 1. Then
/ tt(x) div (/?(#) dx = / / u (x) div (p(x)dtdx

t
D
JR JR<* JR
u (x) div ip(x)dxdt t
- I RS JR d
JR byJR Fubini's theorem. (A.3)
By definition of S and the implicit function theorem,

u 0
d d
R \ S is a hypersurface of class C . Since we assume supp< C
u
d
R \ S , we may apply the divergence theorem to obtain
u
/ div(p(x)dx= / (f(x)n(x)d(dU )(x)

t
d
Jut J(dU )nR \s
t u
and
- / div(p(x)dx= / (p(x)n(x)d(dU )(x),

t
jR*\U t J (dUt)r\R<*\S u
The coarea formula 259
where n(x) is the exterior normal of U . We use this in (3) (recall t
the definition of Ut) to obtain
/ Du(x)ip(x)dx
d
JR
= / u{x) div (p(x)dx

d
JR
= [ [ <p(x)n(x)d(dU )(x)dt t
d
Jm Jdu nm \s t u
1 d
< / \ - (t)nR \S \ _ dt
u u d 1 since we assume \(p < 1|
JR
JR
Taking the supremum over all such (p, we obtain
/ \Du{x)\dx= [ \Du{x)\dx< [ I ^ W L i ^ . (A.4)

d d
JR JR \S u JR
d
(3) We now prove the reverse inequality. We let l n : R > E be
piecewise linear maps w i t h
lim / \l - u\ = 0
n (A.5)
n d
-^ jR
lim / \Dl \= n [ \Du\. (A.6)

n d
-* jR JRd
Let
d
U? := {x e R : l {x)
n > t}.
By (A.5), there exists a countable set T\ C R with the property

that for all t T i
lim / |Xt - X?l = 0, (A.7)

n > 0
~ JRd
where \t is the characteristic function of {u(x) > t}, and Xt the

one of {l (x) > t}. As noted above, by Sard's theorem and the
n
implicit function theorem, there exists a null set T2 C R such

- 1 d
that for all t T 2 , u ( ) is a smooth hypersurface of class C .
We put
T := Ti U T . 2
Appendix A
Let t e R \ T , e > 0. By Lemma 7.1.2, there exists g CQ*

with \g\ < 1 and
div g(x)dx + - . (A.8)
We let M : = J Rd \div g(x)\dx. We choose no so large that for

n > n 0
e
/ \Xt-Xt\ (A.9)
Then for n > UQ
/ div g(x)dx j div g(x)da

J{u(x)>t} J{ln(x)>t}
<M [ | X t - ?|ds<5.
x (A.10)
(A.8) and (A. 10) imply
div g{x)dx + -
2
(*)>*}
<
J{1
// div g(x)dx + e
= j g{x)n(x)d{d{l {x) > t}) + e, n
./0{M*)>t}
n(x) denoting the exterior normal of {l (x) n > t}
1
<K (t)\ _ d 1 +^
Thus, for t < T,
From Fatou's lemma (Theorem 1.2.2), (A.2) and (A.6), we obtain
/ R d I t i - V * ) ! ^ ! * < l i m t a f jf I ^ W I ^ dt
< lim inf / \Dl (x)\dx

n
n->oo JRd
= / |Du(x)|dx. (A.ll)
(A.4) and ( A . l l ) easily imply the claim.

q.e.d.
The coarea formula 261
d d
C o r o l l a r y A . l . Let u e C$(R ), # : R - R integrable, Q C R open.
Then
[ g(u(x))\Du{x)\dx= f g(t)\u~\t) DQl^dt. (A.12)

in J-oo
Proof. (A.12) follows from Theorem A . l i f g is the characteristic func

tion of an open set and similarly i f g is the characteristic function of a
measurable set. By considering
g+(t) : = m a x ( 0 , (?(*))
g~(t) : = m a x ( 0 , -g(t))
separately, it suffices to consider the case where g > 0, since always

+
g(t) = g+(t) g~(t). We thus assume g > 0. Let now (p )neN C R n
with
lim p n = 0
n*oo
oo
X) Pn = OO,
n=l
and put inductively
A n := I x R : g(x) > p + ] T ^ x ^ (x)

n
d
Then for all x R
oo
0(x) = ] T p X A ( z ) .n n (A.13)
n=l
Since we observed that (A.12) holds for \ A in place of g, the repre n
sentation in (A.13) in conjunction with Beppo Levi's Theorem 1.2.1 on

monotone convergence then implies (A.12) for g.
q.e.d.
Remark A.l. The coarea formula is due to Federer. I t holds more gen
d
erally for Lipschitz functions u : R > R. See H . Federer, Geometric
Measure Theory, Springer, New York, 1969, pp. 241-760, 268-71.
Appendix B
The distance function from smooth
hypersurfaces
We also need some elementary results about the (signed) distance func
d
tion from a smooth hypersurface. Let ft C E be open with nonempty
boundary dft. We put
M \ _ / dist(x,dft) if x G ft
d : d
W~ \_dist(x,0ft) ifxGE \a
d is Lipschitz continuous with Lipschitz constant 1. Namely, for x,y G

d
R , we find 7r G dft with d(t/) = | j / 7r |, hence
y y
< \x - 7r | < |x - y\ 4- |y - ?r | = |x - y| 4- % ) ,
y y
and interchanging the roles of x and t/ yields
\d(x)-d(y)\ <\x-y\.
2
We now assume that dft is of class C . Let #o dft. Let n ( x ) be the 0
outer normal vector of ft at #o, and let be the tangent plane of dft
d
at xo. We rotate the coordinates of W* so that the x coordinate axis
1
is pointing in the direction of - n ( # o ) . I * some neighbourhood U(xo) of
#o, # 0 can then be represented as
(B.l)
1 1 d 1 2
with x = ( x , . . . , x - ) , where / G C ( T i D C/(x )), Df(x' ) = 0. The 0 0 Q
2
Hessian D f(xo) is symmetric, and therefore, after a further rotation of
coordinates, it becomes diagonalized,
( Kl 0 \
2
D f(x ) 0 = (B.2)
262
The distance function from smooth hypersurfaces 263
2
K I , . . . , K d - i are the eigenvalues of D f(xo), and they do not depend on
the special position of our coordinates. They are invariants of dfi, and
are called the principal curvatures of dQ at x$. The mean curvature of
dft at #o is
d _ 1
1 1
H ( x o ) = = A
E ^ rfn /(ô). (B.3)
2 =1
The outer normal vector n(x) at x G dfl f l ^ ( X Q ) has components
TZ(X')
n'(x) = ^ ' ^ ' , , i= l,...,d-l (B.4)
d
n (a;) = (B.5)
( l + |Z?/(x')|)'
1 d _ 1
(#' = ( x , . > x ) ) . I n particular
n'(x ) = 0 for t , j = 1 , . . . , d - 1. (B.6)
d
L e m m a B . l . Suppose is open in R and that dfl is bounded and of
k
class C with k > 2. For rj e R, pu
d
:= { x R : d(x) = r/}.
TTiere exists eo > 0 (depending on dQ) with the property that for
M < c, 0
k
is a hypersurface of class C . vMso,
lim 12,1 = ^ 1 ^ . (B.7)
2
Proof Since d f i is compact and of class C , there exists e > 0 with the
following property: Whenever \rj\ < e for each x$ d f i , there exist two
d
unique open balls B B with B C fi, JB C R \ f i ,
u 2 x 2
Bi n an = x = B n an 0 2
2
of radius |r/|. The eigenvalues of the Hessian D f(xo) of a normalized
representation / of d Q at #o as above then have to lie between - and
i,i.e.
\i*\<\ (B.8)
for the principal curvatures K i , . . . , Kd_i. I f a; is a centre of such a ball,

264 Appendix B
then #o K . Also, by uniqueness, these balls depend continuously on

X
#o dft. Thus, i f \rj\ < e, each x is the centre of such a ball, and
nx = x 4- n(x)d(x) with n(x) : = n(n )

x (B.9)
is the unique point in dfl with |x 7r | = We once again employ x
the coordinates used for the definition of / and rewrite (B.9) as
x = F(x', d) = {x\ f(x')) - n(x', f{x'))d. (B.10)

f c _ 1 d
Then F G C (((T Xo n C/(x )) x R) , R ) and at the point
0 (x' ,d(x))
0
( 1 - Kid(x) 0 \
DF by (6) . (B.ll)
K -\d(x)
d
V 0 1/
By (B.8) and since \q\ < e,
det D F ^ 0.
k x
By the inverse function theorem, x' and d therefore locally are C -
functions of x (cf. (B.9)). Since
d(x) = d(x 0 ~ rjn(xo)) = 77,
we have
Dd(x) - U(XQ) = 1.
Since d is Lipschitz w i t h Lipschitz constant 1, we conclude
\Dd(x)\ = 1
and
1
Dd(x) =-n(x ) 0 C*- .
k k
Thus d e C locally, and the level hypersurfaces Y> are of class C . For v
(B.7), we may w.l.o.g. take n > 0 as the case n < 0 succumbs to the
same reasoning. We consider the vector field
V(x) = Dd(x).
The Gauss theorem yields
/ divV(x)= / K(x)n(x)d{E }(x)+/ 0 K(a?)n (x)d{E }(ar)

t| t| >
J{0<d(x)<rj} JHo
The distance function from smooth hypersurfaces 265
where is the normal vector of pointing in the direction opposite

to n. Since the measure of { 0 < d(x) < 77} goes to zero w i t h 77 and
V{x) = -n(x) for x G E = dQ

0
V(x) = n^x) for x G 7 ? ,
(B.7) easily follows.

q.e.d.
References
D. Giibarg, N. Trudinger, Elliptic Partial Differential Equations, Springer,
Berlin, 2nd edition, 1983, pp. 354-6.
8
Bifurcation theory
8.1 Bifurcation problems in the calculus of variations

We wish to consider a variational problem depending on a parameter A,
and to investigate how the space of solutions depends on this parameter.
We thus consider
1
A is supposed to vary in some open set A C E . Often, one has / = 1.
We assume that
F : [a, 6] x R x R x A R
d d
is sufficiently often differentiable so that all derivatives taken in the

sequel exist. For that purpose, one may simply assume that F is of class
C in all its arguments although that is a little stronger than needed
in the sequel.
Remark 8.1.1. One may also impose boundary conditions depending on

A, i.e.
u(a) = ui(X)
u(b) = u (\),
2
and finally, one may vary the boundary points themselves,
a = a(A)
6 = 6(A).
This latter variation, however, can formally be incorporated in the vari

ation of F , by transforming the integral.
266
Let
T(-,A):[a(Ao),&(Ao)]-[a(A),6(A)]
be a bijective linear map, for some fixed Ao- Then

,6(A)
/ F(T,tl(T),u(T))dT
Ja(X)
dr(t, A)
dt
rb(X )
0
yields a parameter-dependent variational integral for v w i t h fixed bound

ary points a(Ao),6(Ao).
As established in Theorem 1.1.1 of Part I , a critical point u of / ( , A) of

2
class C satisfies the Euler-Lagrange equations
F (t,
pp u(t), u(t), X)u(t) + F (t, pu u(t), ii(t), X)u(t) (8.1.1)
+ F ( * , u(t), ii(t), A) - F (t, u(t), ii(t), A) = 0.
pt u
We abbreviate (8.1.1) as
L\u = 0. (8.1.2)
In the light of Theorems 1.2.2 and 1.2.4 and Lemma 1.3.1 of Part I , we
shall assume
det F (t,u(t),u(t),X)
pp ^0 (8.1.3)
for all functions u occurring in the sequel. Equation (8.1.3) implies that
(8.1.1) can be solved for u in terms of u and i.e.
1
ii = -Fpp(t, ti(t), ii(t), A ) " {F (t, pu u(t), u(t), X)u(t)
+F (t,
pt u(t), u(t), A) - F (t, u{t), u(t), A)}
u (8.1.4)
=:f(t,u(t),u(t),X),
268 Bifurcation theory
see (1.2.10) of Part I . (8.1.2) thus is equivalent to
u(t) ~ / ( t , u { t ) , A) = 0. (8.1.5)
The topic of bifurcation theory then is to study the space of solutions of

(8.1.5) in its dependence on the parameter A. Before approaching this
problem from a general point of view i n the next section, we should
briefly comment on the relations w i t h the Jacobi theory introduced in
d
Section 1.3 of Part I . For a critical point u of / ( , A) and rj G Dl(I,R ),
we had established the expansion
2 2
I(u 4 sr), A) = J(u, A) 4 ~s 6 I(u, ly, A) + o(s) for s - 0, (8.1.6)
with
2 b
d f
2
6 I(u, r/, A) = Q (n)
x := ^ J F(t, u(t) 4 sn(t), u{t) + A)cft, a==0
= / {F i j{t,u,u,X)rji7jj
p p + 2F t (t,u,u,\)rjirjj
p UJ (8.1.7)
4- F i (t,u,u,\)r)ir)j}dt,
u uj
abbreviated as
b
{ F , r ) r ) 4 2F p 'nV + F uVV} dt.
f A pp Xy U XyU
Critical points of Q satisfy the Jacobi equations
Jx(u)n:= ^(F (t,u,u,\)r)

pp + F (t,u,u, pu \)rj) (8.1.8)
-F (t, pu u, ii, A)r) - F (t, uu u, u, X)rj = 0.
J\(u) is called the Jacobi operator associated w i t h the critical point u

of / ( , A). We also observe that
d_
J\(u)n = L (u 4 srj) . x ls=0 (8.1.9)
ds'
Of course, this is not surprising since L \ represents the first variation of
/ ( , A) and J\ the second one. From the expansion (8.1.6) we see that
2
I(u 4 si/, A) < A) if 6 I{u y n A) < 0.
y (8.1.10)
No such conclusions can be achieved, however, if

2
6 /(u,77,A)=0. (8.1.11)
Now by Lemma 1.3.2 of Part I , for a Jacobi field 77 that vanishes at

the boundary points a and 6, (8.1.11) holds. This indicates that Jacobi
fields play a decisive role for deciding about the minimizing property of
a critical point u of / ( , A). Jacobi fields satisfy
Jx(u)rj = 0, (8.1.12)
i.e. are solutions of the linearization of the equation L\u = 0 satisfied

by u. This also indicates that Jacobi fields will play a decisive role in
analysing the bifurcation behaviour of L\u = 0 as A varies. Namely,
in finite dimensional problems, the presence of a nontrivial solution of
the linearization of a parameter-dependent equation L\u = 0 at some
parameter value Ao either results from a nontrivial family u(r) of solu
tions of L\ U(T)
0 = 0 by differentiating the equation w.r.t the parameter
r , or it indicates a nontrivial bifurcation as A varies in the vicinity of
Ao- I n the next section, we shall see that under appropriate assump
tions, the same also holds in the present infinite dimensional context.
In fact, the bifurcation problem will be reduced to a finite dimensional
one via Lyapunov-Schmid reduction. The reason why this is possible in
our variational context is that under our assumption (8.1.3), the space of
Jacobi fields is always finite dimensional. Namely, analogously to (8.1.4),
(8.1.5), the assumption (8.1.3) implies that (8.1.8) can be solved w.r.t
77, i.e
77 <>(, u, u, 77,77, A) = 0. (8.1.13)
(Although this is not indicated by the notation, (8.1.13) is a linear equa

tion for 77, and so the space of solutions is a linear space.)
Now suppose that we have a sequence (rj ) ^ of solutions of (8.1.13)
n ne
(for fixed A) that are bounded in some appropriate function space

2 2,2 2
like C (I) or W (I). For concreteness, let us consider C ( 7 ) , i.e. for
example
2
H^Hc (7) 1-
By the Arzela-Ascoli theorem, after selection of a subsequence, (f] )neN n
l
then converges in C (I) to some limit denoted by 770. (8.1.13) then i m
plies that (77) N converges in C(I) (as it follows from our assumptions
n
on the differentiability of F that ip is smooth, in particular continu

ous). Thus (since the uniform limit of derivatives is the derivative of the
2
limit), ( 7 7 ) N converges in C (I) to 770, and consequently 770 also solves
n n e
(8.1.13). From this compactness result, one easily deduces that the space
of solutions of (8.1.13) has finite dimension.
8.2 T h e functional analytic approach to bifurcation theory

We consider the following general situation. We have Banach spaces
V, W, and a parameter space A . We assume that A is an open subset
of some Banach space. We consider a parameter dependent family of
equations
Lu
x = 0, (8.2.1)
with
V x A-> W
(u, A) H-+ L\u.
We assume that L\U is sufficiently often differentiable w.r.t. to u and

A so that all subsequent expansions are valid. The aim of bifurcation
theory is to study the set of solutions u of (8.2.1) as A varies, to identify
the bifurcation values of A, i.e. those values of A where the structure
of the solution set changes, and to investigate that structure at such
bifurcation points. I n order to arrive at concrete results, we need an
additional assumption. We consider the derivative of L\u w.r.t. u,
J {u)v
x : = (D L {u))v
u x : = ^ L ( u + tv)^
A 0 (8.2.2)
for v V. We assume that J\ is a Predholm operator of index 0, i.e. that

ker JA and coker J\ are of finite and equal dimension, and furthermore
that there exists a canonical isomorphism
ker J A ^ coker J .A (8.2.3)
We first consider the case where
Lu
Xo 0 =0 (8.2.4)
ker JAO(Ô) = { 0 } tor some Ao G A , uo G V. (8.2.5)
We shall see that in this case, no bifurcation can occur at Ao- Namely,
we have:
T h e o r e m 8.2.1. Let L\ uo 0 for some Ao A , UQ V, ker J A ( ^ O ) =

0 0
{ 0 } . Then there exist neighbourhoods U(\Q) of Ao in A and V(UQ) of UQ

in V such that for all A G U(\Q), there exists a unique u V(UQ) with
Lu
x = 0.
Proof. Since J\ 0 is assumed to be a Fredholm operator of index 0, (8.2.5)

implies that
JA 0 : V - W
is an isomorphism. Thus the derivative w.r.t. the variable u of the map
VxA->W
(u, A) i L\u
is an isomorphism at (tto,Ao), and the implicit function Theorem 2.4.1

implies that the equation
L\u = 0
can be locally resolved w.r.t. u, i.e. there exist neighbourhoods

J7(Ao), V(uo) and a map
U(X ) 0 - V(u ) 0
A i-+ u(X)
such that
Lu x = 0
precisely i f
u = u(X).
q.e.d.
We next consider the case where
Lu Xo 0 = 0 (8.2.6)
K := ker J \ ( U Q ) is one-dimensional.
0 (8.2.7)
The assumption that this kernel is one-dimensional may look restrictive,

but it is typically satisfied in variational problems, and in this situation,
we can already see the typical phenomena of bifurcation while avoid
ing additional technical complications that arise for higher dimensional
kernels. I n the sequel, we shall assume for simplicity
u = 0
Q
(which can always be achieved by changing the dependent variables in

our equation by a translation). I n the sequel, we shall also usually write
J\ in place of J\ (uo) = J\ (0).
0 0 We may write
o
V = KV U (8.2.8)
and in view of (8.2.3), we may also write
W = KW U (8.2.9)
with
W x = Jx (V) 0 = JAW). (8.2.10)
We denote by
:V-+K
the projection onto K according to (8.2.8), and we consider n(V) as a

subspace of W, according to (8.2.9). Thus, i f
u = -f w
with K, w V i , then
n{u) = .
I n particular,
TT(0) = 0.
We consider the map
AXo :V^W
u LA W +
0 n(u).
L e m m a 8 . 2 . 1 . A\
0 is a local diffeomorphism, i.e. the derivative
DA\ 0 = DA\ (0) : V W is an isomorphism.
0
Proof. The derivative is computed as
DA v
Xo = J A o f + 7r(v) for v G V. (8.2.11)
The Fredholm operator J yields a bijective continuous linear map be

A o
tween V\ and W\ because of the decompositions (8.2.8), (8.2.9), (8.2.10),

and its inverse is likewise continuous (by Definition 2.3.1). From the
definition of K and TT and (8.2.3) we then conclude that DA\ is an 0
isomorphism.
q.e.d.
We now consider the map
A: V x A - > W
(tx, A) >-+ A$u) := L\{u) + 7r(tx).
By Lemma 2.3.4, there exists a neighbourhood V(Ao) of Ao in A such

that for all A V ( A ) , A\(0) is a local diffeomorphism. We may therefore
0
apply the implicit function Theorem 2.4.1. Consequently, as
i4(0,A ) = 0,
0 (8.2.12)
there exist neighbourhoods U(0) of u = 0 in V, Ui(0) of 0 in W such

0
that for all A V ( A ) and G t / i ( 0 ) , there exists a unique u e U(0)

0
with
A(u,$ = , (8.2.13)
i.e.
L n + T T ( U ) = .
A (8.2.14)
We write
tx = tx(f,A)
for the solution tx of (8.2.13). We have in particular
u ( 0 , A ) = 0,
0 (8.2.15)
since L\ 0
o = 0, 7r(0) = 0 (remember txo = 0). In this notation, (8.2.13)
is
A{u{t,\)A)) = Z-
The aim now is to find with
T(tf,A))=, (8.2.16)
because (8.2.14) will then give
L tx(,A) = 0,
A (8.2.17)
which is the equation that we wish to solve. Since the image of IT is

assumed to be one-dimensional (and in any case finite dimensional as J A
is supposed to be a Predholm operator), we have reduced our bifurcation

problem to a finite dimensional problem. I n the sequel, we shall thus let
vary only in K, the image of TX. Thus, we may consider as a scalar
quantity, = ao, with a G l , where o is a generator of K. We denote
the derivative of u(a:o> A) w.r.t. a and A, respectively, at a 0, A Ao

by
d u and d\u,
a respectively.
(Note that A in general is not a scalar quantity, as we do not assume

that A is one-dimensional.) Differentiating (8.2.14) w.r.t. a yields
J duXo a + 7T(d u)a = ^ (8.2.18)
Since G K, also
0
</A o + 7r(o) = o.
0 (8.2.19)
Lemma 8.2.1 then implies
du = .
a 0 (8.2.20)
We are now ready for the essential point, namely the asymptotic expan
sion of the equation (8.2.16), i.e.
7r(u(,A)) = (8.2.21)
near 0, A = Ao-
2
We let d u, d\u be the second derivatives of u(ao> A) w.r.t. a and A,
respectively, at a = 0, A = Ao, and likewise d\ u be the mixed second x
derivative w.r.t. a and A. Higher derivatives will be denoted similarly by

corresponding symbols. The Taylor expansion of (8.2.16) then is
= 7r(u(, A)) = 7r(0) 4- 7r(d u)a a 4- n(d\u)/i

1 1
-f terms of higher order in a and / i . (8.2.22)
Since 7r(0) 0 and since, by (8.2.20), d u = > hence n(d u)a a 0 a ao =

, we may write (8.2.22) as
2 2
0 = ir{d\u)fji - f ^n(d u)a 4- higher order terms in a only (8.2.23)
-f 7r(<9^ u)a/i
x - f higher order terms that also involve / i .
Remark 8.2.1. I n order to interprete the terms in this expansion, we

differentiate (8.2.14), i.e
L u(tx A) 4- 7r(n(, A)) = a 0 (8.2.24)
twice w.r.t. a. One differentiation yields
Jx(u)d u a + 7T(d u)a = Zo, (8.2.25)

and differentiating once more gives

2 2
DJ {d uf
x a + Jd u
x a + 7T{d u) = 0.
a (8.2.26)
We put A Ao and project onto K in the decomposition (8.2.9). We

may also denote that projection by 7r, and we then have TT O J \ 0. 0
Also, from (8.2.20), d u = o, and so we get

a
n(DJ e ) = -*{d u).

x 0
2
a (8.2.27)
Thus, the first term in the expansion of Q in (8.2.24) can be expressed via
DJ\. I n a variational context, J\ represents the second variation, and so
DJ\ represents the third variation of the variational integral. Likewise,
2
if d u vanishes, i.e. if the third variation vanishes on the Jacobi field o>
then 7r(dû) can be expressed by the fourth variation, and so on.
We now discuss the simplest case of a bifurcation, namely where

2
7T{d ii) a ^ 0. (8.2.28)
2 2
We put a : = tr, ji = : /i, ao : = 7v(d\u)p, a\ : = ~7r(d u), y and (8.2.23)
becomes
2 2 2 2
0 = at 0 + ait r + t E ( t , r, p) (8.2.29)
with
(, r, ft) = 0(t) for fixed r, ft for t 0. (8.2.30)
For ^ 0, (8.2.29) is equivalent to

2
0 = a + air 0 + (*, r, /2). (8.2.31)
We shall now see by a simple application of the implicit function theorem

that the bifurcation behaviour of equation (8.2.31) is equivalent to the
one of
2
0 = a -fair . 0 (8.2.32)
We assume ao 0; as will be discussed below (see Lemma 8.2.2), this

can be derived from a suitable assumption about the variation of L \ as a
function of A. (8.2.28) of course means that a\ ^ 0. I f ^ > 0, then there
is no solution r of (8.2.32), whereas for ^ < 0, we have two solutions
T i , T 2 . We keep / i fixed for the moment and write (8.2.31) as
2
0 = a + a i r - f (, r, p) = : #(, r ) .
0
(8.2.33)
We consider the case ^ < 0 w i t h the solutions T I , T 2 of (8.2.32). As

E(0,,/2) = 0, we have
# ( 0 , T i ) = 0 f o r i = 1,2, (8.2.34)
whereas
_ $ ( 0 , T J ) ^ 0, because of a , a i ^ 0 and (8.2.30).

0 (8.2.35)
The implicit function theorem then implies the existence of (locally

unique) functions
n{t) : ->R
for i = 1,2, 0 < |t| < , for some > 0 that satisfy
$(t,r<(t)) = 0. (8.2.36)
We have thus found two solutions T I ( ) , T ( ) of (8.2.33), hence (8.2.22),

2
hence (8.2.16), hence (8.2.17), i.e. (8.2.1) for t ^ 0, for the parameters
2
A = A + t Jl.
t 0 (8.2.37)
In the other case, ^ > 0, (8.2.30) implies that for sufficiently small
\t\ ^ 0, there is no solution of (8.2.33), i.e. of (8.2.1). Thus, as promised,
the bifurcation behaviour in case 7r(9^n) ^ 0 ( (8.2.28)) is completely
described by the simple quadratic equation (8.2.32). Of course, replacing
ft by ft changes the sign of ao and thus interchanges the cases ^ > 0
and < 0.
ai
We summarize our result in:
T h e o r e m 8.2.2. We consider a parameter dependent family of equa
tions
Lu
x = 0 (8.2.38)
as above,
V x A->W
(u, A)i L u,
x
where V, W are Banach spaces and A is an open subset of some Banach

space, and L\u is smooth in u and A. We suppose that
L0
Xo = 0,
and we wish to find the solutions of L\u = 0 in the vicinity of 0 as A

varies in the vicinity of Ao.

With
J\{u)v := (D L\(u))v u
L {u
= -j;L x + *v)| t = 0 ,
dt
we assume that there is a canonical isomorphism
ker J A = coker J x (see (8.2.3)), (8.2.39)
and we let
ir:V^kevJ Xo (JXo = J (0))

Xo
be a projection as defined above (see (8.2.8)-(8.2.10)). We assume fur

thermore that
dim ker J A o = 1 (see (8.2.7)). (8.2.40)
We assume that there exists some ft with
a 0 : = 7T(d u)ft
x (= j v(u(Q, t Ao + t/2))| =o) ^ 0t (8-2.41)
(see Lemma 8.2.2 below), and also
2 a i :=7T(dlu)^0 (8.2.42)
(nonvanishing of the third variation, see Remark 8.2.1). Then there exist
2
e > 0 and a variation X = Ao - f t ft of Ao with the property that for
t
0 < t < e, there exists a neighbourhood U of 0 in V such that the t
number of solutions u U of t
Lu Xt = 0 (8.2.43)
equals the number of solutions of the quadratic equation

2
a + aT
0 X = 0. (8.2.44)
q.e.d.
Remark 8.2.2. Since kerJA , the image of 7r, is assumed to be one-

0
2
dimensional, we have simply considered ir(d u), 7r(d u) as scalar quan x
tities.
We now consider the case where

2
ir(d u) a = 0, (8.2.45)
but
n(dlu) ^ 0. (8.2.46)
(8.2.23) then becomes
3 w a i
0 = ir(d\u)ii 4- ^7r(dlu)a + 7 r ( ^ a , A ) / + higher order terms. (8.2.47)
For a complete description of the bifurcation behaviour, this time we

need to consider a two parameter variation. We assume that there exist
/ii,/i 2 with
7r(9 w)/ii + 0
A (see Lemma 8.2.2 below) (8.2.48)
and
K(dl u)ti2
)X + 0, but ir(d u)ii x 2 = 0- (8.2.49)
3 2
We put a : = t r , /x = 6 i / i i -f 6 / i , with parameters b\, 6 , and rewrite 2 2 2
(8.2.47) as
3
0 = t (7r(a w)/ii6i + TT(dl u)fi b T A )X 2 2 (8.2.50)
3
+ h{dlu)T + Z{t,T^ fl )) U 2
3 3
=: c t (a -f a i r -f r
0 0 - f E(, r, / i i , / i ) ) , 2
3
with c = ^ T T ( ^ ^ )
0 0 (see (8.2.46)). For t 7^ 0, this is equivalent to
3
0 = a + air + r 0 + E(, r, / i i , / i ) . 2 (8.2.51)
Again
E ( t , r , / i i , / i ) = O(t) as -> 0, for fixed r, / i i , / i .

2 2 (8.2.52)
As before, we may thus invoke the implicit function theorem to conclude

that the qualitative description of the bifurcation behaviour is furnished
by the solution structure of the cubic equation
3
0 = a -f air4-r .
0 (8.2.53)
In particular, locally there exist at most three solutions. We summarize

our result in:
T h e o r e m 8.2.3. As in Theorem 8.2.2, assume the general conditions

(8.2.38)-{8.2.40). Furthermore, we assume that there exist parameter
variations / i i , / i 2 with
7T(9AU)/XI + 0, (8.2.54)
*{tiL \v)ii2t + 0, but 7r(<9 u)/i = 0 A 2 {see {8.2.48), {8.2.49)). (8.2.55)

Then there exist e > 0 and a two-parameter variation of Ao,

2
X = A + t%fn
t 0 +t b ^ 2 (8.2.56)
suc/i that for 0 < < e, there exists a neighbourhood Ut of 0 in V for

which the number of solutions u G Ut of
Lu
Xt = 0 (8.2.57)
equals the number of solutions of the cubic equation

3
a + aiT + r
0 = 0. (8.2.58)
(ao = ^TT(d\u)biî/u(dl u),ait = 2

6Tx(d u)b fi2/n(daU),
x 2 noting Remark
8.2.2 again.) q.e.d.
What we are seeing in Theorem 8.2.3 is the so-called cusp catastrophe

(in the language of R. Thorn's theory of catastrophes), the bifurcation of
the zero set of a cubic polynomial depending on the parameters ao, a\. In
the same manner, one may also identify conditions where the bifurcation
behaviour is described by other so-called elementary catastrophes, as
classified by R. Thorn (see e.g. T h . Brocker, Differentiable Germs and
Catastrophes, LMS Lect. Notes 17, Cambridge Univ. Press, Cambridge,
1975). The higher the order of the polynomial involved, however, the
more independent parameters one needs. The general idea is that the
singular behaviour at a bifurcation point, in particular the nonsmooth
structure of the solution set at such a point, is simply the result of the
projection of a smooth hypersurface in the product of the solution space
and the parameter space onto the solution space. The singularity arises
because that hypersurface happens to have a vertical tangent plane over
the solution space at the bifurcation point.
In order to discuss the assumption (8.2.41), (8.2.54), we provide
L e m m a 8.2.2. Assume that for every there exists some \i with
(3 = TX(D L U(0,
X XQ AO) fa) (:= 7 r ( i L A o + t M u(0, A ))|
0 T = 0 ). (8-2.59)
(Again, we write (3 in place of /3o and consider it as a scalar quantity,

as the image of TX is assumed to be one-dimensional.) Then for every
J S G R , there exists some /i with
Tx((d u)v)
x - /?. (8.2.60)
Proof. We abbreviate
At = AQ - f t/j( as A is open, AQ - f tfi A for sufficiently small \t\).

By (8.2.14)
LA.(, A T ) + A ))
T = .
Taking the derivative w.r.t. t at t = 0 and = 0 gives
n((d u)fi)
x = -^(L A t w(0, A ))| t t = 0
L 0 A
= ~ ~ ( ; ^ * t M ' o)|t=o
-(D L u A o )û(0, A )| t t = 0 .
Since D L\ U = J , and 7r o J
0 A o = 0 by definition of 7r, applying 7r to
A o
both sides of the preceding equation gives
ir((d\u)ti) = -n{D Lx )u(0, x o A ),

0
and by assumption (8.2.59), we may find fi for which the right-hand side
becomes /3o- (We take -/? in place of (3 in (8.2.59).)
a.e.d.
The approach to bifurcation theory presented here originated with

L. Lichtenstein, Untersuchung (iber zweidimensionale regulare Vari-
ationsprobleme, Monatsh. Math. Phys. 28 (1917), and was developed in
X. Li-Jost, Eindeutigkeit und Verzweigung von Minimalflachen, Thesis,
Bonn, 1991, see also X . Li-Jost, Bifurcation near solutions of variational
problems w i t h degenerate second variation, Manuscr. Math. 86 (1995),
1-14, J. Jost, X . Li-Jost, X . W . Peng, Bifurcation of minimal surfaces
in Riemannian manifolds, Trans. AMS 347 (1995), 51-62, Correction
ibid. 349 (1997), 4689-90.
The reduction of a bifurcation problem in an infinite dimensional set
ting to a finite dimensional one is an example of the Lyapunov-Schmid
reduction which we now wish to discuss.
As before, we consider a parameter dependent family of equations
L\u = 0 (8.2.61)
with
V xA-*W
(u, A) i- L\u.
(V, W Banach spaces, A an open subset of some Banach space) near

( u , A ) with
0 0
L uo
Xo = 0. (8.2.62)
Again, we assume that J$u) = D L\(u) U is a Fredholm operator. Thus
V : = ker
0 J\ {u )
0 0
is finite dimensional, and we have decompositions
V = V 0 Vi0 (8.2.63)
W = W Wi, with W i = R{J\ {u )),
0 0 0 W0 finite dimensional.
(8.2.64)
(R denotes the range of an operator as in Definition 2.3.1.) We let
TV : W -+ Wo
be the projection defined by the decompostion (8.2.64). Then our equa

tion L\u = 0 is equivalent to
TXL U
X = 0 and {Id - TT)L U X = 0. (8.2.65)
We first consider
{Id-n) : Vi x V x A W i , 0
and with X := VQ X A, we write

;
L\(v" + t / ) = 0(1/", A) with v V , 0 V i , A e A.
Then, at v" + v' = uo,
D g(v",v v', A ) = D L {v"

0 v Xo 4-1/) : V x Wi
is an isomorphism by definition of V i , W\\ namely it is simply J\ (uo), 0
considered as a map from Vi to W\. Therefore, by the implicit function

Theorem 2.4.1, near ( n , A ) , we may find a unique
0 0
v" = <p(v',$
with
{Id - TT)L {V' X 4- <p{v', A)) - 0. (8.2.66)
Thus u = v' 4- <p(v', A) solves L\u = 0 if and only if
irL\{v' + <p{v\\)) -0. (8.2.67)
Equation (8.2.67) is a finite dimensional system of equations, because

the image of TT, Wo, is finite dimensional. This is a Lyapunov-Schmid
reduction, and we have seen an instance of this in detail in the preceding
for the case where Vo and Wo are one-dimensional. A general reference
for this and other topics and methods in bifurcation theory is S. N . Chow,
J. Hale, Methods of Bifurcation Theory, Springer, New York, 1982.
8.3 T h e existence o f catenoids as a n e x a m p l e o f a b i f u r c a t i o n

process
We consider the variational problem
I(u)= j F(t,u(t),u(t))dt (8.3.1)

Ja
with
2
F(, u, u) = u \ A + u . (8.3.2)
This variational problem is of the type considered in Section 1.1 of Part I .

I(u) with F given by (8.3.2) is the area of the surface of revolution ob
tained by rotating the curve u(t), a < t < 6, about the t-axis. Crit
ical points are so-called minimal surfaces of revolution. According to
Theorem 1.1.1 of Part I , the corresponding Euler-Lagrange equation is
computed as
j F (t,u{t),u{t))
t p -F (t,u{t),u{t))
u
= F pp (t, u{t), i ( t ) ) i i ( t ) + F pu ( t , u(i), u(t))tz(t)
+ F (t, u(), u()) - F (t, u{t),

pt u
= 0
which in the present case becomes
/
y/l -f 6 2
dt V v l T ^ 2
2
2
/ 2 3
+ - 7 ^ = ^ " \ A + u - 0,
/ 2
(8.3.3)
(v TT^ ) v IT^
or equivalently
2
nix - i i - 1 = 0. (8.3.4)
By (1.1.7) of Part I , we have F - iiF p = constant, since F = 0, hence

t
2
in our case F = u\/l - f u ,
= constant = : A.
Therefore, for A 7^ 0, the general solution of (8.3.4) is
u{t) = A cosh
A
8.3 Example: bifurcation of catenoids 283
with parameters A, t . Here A ^ 0, and we may assume A > 0 as the ease

0
A < 0 is symmetric to the case A > 0. Also, since to just represents a

translation of the independent variables, we may assume to = 0, i.e.
u(t) = A c o s h ^ . (8.3.5)
The curve u(t) is called a catenary, and the minimal surface of revolution
obtained by revolving u(t) about the t-axis is called a catenoid. For the
sake of normalization, we consider the interval I = [1,1]. In order to
use the general theory of Section 8.2, we need to choose appropriate
Banach spaces V, W and A = E and consider the operator
L x : VxA-^W (8.3.6)
( n , A ) f _ >
(it (vTT6*) " ^ 1 + i 2
^ ~ Aeoshl,?x(-f)
A cosh j
On the right hand side, we have a differential operator of second order

and a Dirichlet boundary condition. The boundary values are real num
2
bers, and so W should contain R as a factor as we have two boundary
points. Otherwise, V and W shouM differ by two orders of differentia
bility. Thus, possible choices are Sobolev spaces
k+2 P k p 2
V = W ' (I), W = W > (I) x E for some k,p
or spaces of differentiable functions

+ 2 k 2
V = C* (I), W = C (I) x E .
Here, we shall take

2 2 2 2
V = W ' (/), W = L (I) x M , (8.3.7)
but the reader should also convince herself or himself that the other
2
choices work as well, although the space L will always play some aux
iliary role.
2 2
In the sequel, we shall denote the scalar product in L (I) by (, - ) L ,
i.e.
(WI,W )L*
2 = J wi(t)w (t) 2 dt
2 2 2
for wi,w 2 G L (I). The scalar product on W = L (I) x E for w x =
2 2
w = (w ,s )
2 2 2 (wi,w 2 G L (I),s ,s x 2 K ),
(wi,W )2 W = (WI,W )L*2 + S I - 8 2
2 2
is obtained from the scalar products on L (I) and on E .
The Jacobi operator is given by
J\(u)v = D L\(u)vu
( 8 3 8 )
- ( M r * ) ' -
by (8.3.5). I n order to determine the kernel of J\(u), we need to solve

the equation
J (u)v
x = 0. (8.3.9)
This is equivalent to
v(t) - f tanh {v(t) + ^v(t) = 0 (8.3.10)

v ( - l ) = v ( l ) = 0. (8.3.11)
The space of solutions of (8.3.10) is generated by
Vi(t) = sinh j
v (t) 2 = cosh j - j sinh j .
(These solutions are simply obtained by differentiating the general solu

1
tion A cosh (^x* ) f (8.3.4) w.r.t. the parameters A and t (at to = 0), 0
cf. Theorem 1.3.3 of Part I.) The boundary condition (8.3.11) cannot be
satisfied by V\, and so we have to find out for which values of A
v(t) := v (t) = cosh { - { sinh {

2 (8.3.12)
satisfies v(l) = v(l) = 0. This is the case precisely i f
A = tanh A. (8.3.13)
We agreed above to consider only positive values of A, and this equation

has precisely one positive solution which we denoted by Ao, and likewise,
we put
uo(t) = A cosh (J^j

0 '
cf. (8.3.5).
The only solutions of (8.3.10), (8.3.11) are av(t) with a G R and v(t)
given in (8.3.12), and so we have
dimker J ( u ) = 1. A o 0 (8.3.14)
We call v a weak solution of the Jacobi equation i f
for all rje Cg(J).
In the sequel, we shall need a little regularity result, namely that any
2
solution v of (8.3.15) of class L (I) is automatically smooth, in fact of
class C(I). As we are dealing with a one-dimensional problem here,
this result is not too hard to demonstrate, but since that would lead us
too far astray, we omit the proof. I t can be found in most good books on
differential equations or functional analysis, e.g. K . Yosida, Functional
Analysis, Springer-Verlag, Berlin, 5th edition, 1978, pp. 177-82.
2
Of course, i f v is of class C , (8.3.15) is equivalent to
for all 17 G C n i ) ,
and by Lemma 1.1.1 of Part I , this is equivalent to v being a solution of

the Jacobi equation.
We shall now identify ker J ( u o ) and coker J (t*o) as required in
Ao Ao
(8.2.3). We shall simply write J in place of J\ (uo). According to our

A o 0
choice (8.3.7), we consider J as an operator A o
22 2 2
J X O : W (I) - L (I) x R .
2 2
If w G Ro(J\ )
0
; =
R(J\ \H > (I))I
0
2 2 L
-- E
there exists v G H ' (I) with
J\ v = w, then for all cp G ker J
0 A o
(w,ip) w = (J\ v,(p) 2

0 L = (v,J cp) 2
Xo L
= 0
(in the same manner as the equivalence of (8.3.15) and (8.3.16) and
noting that cp is smooth and v and <p both vanish on dl.) Thus i f
1 1
w G Ro(J\ ),
0 then also w G (ker J ^ ) - , where - denotes the orthog
2
onal complement in the Hilbert space L (I), as in Corollary 2.2.4. Con
2 2
sequently, i f we denote the closure of a linear subspace M of L (I) x R
by M , then also
Ro(Jxo) C (ker J A J - . 1
1
Conversely, if w G Ro(J\ ) (= 0 ( ^ o ^ A o ) ) " ) , then
2 2
(w, Â ^)vy = 0 0 for all v e H ' (I).
By the regularity result mentioned above, this implies that w is smooth,

and so we may integrate by parts to get
2
(w, J\ v) 0 w = (J w, Xo v) w for all v G H$ (I)
and hence w G k e r J . Altogether, we have shown that

A o
1
ker J A o = RoiJxo)- .
Since, according to Corollary 2.2.4, we have the decomposition
2 L2 1
L (I) = Ro(Jx ) Q (BRoiJxo)- ,
L
we may also consider Ro(J\ )~ 0 as coker J , and so we get the required
A o
identification
ker J A o 2* coker J . A o (8.3.17)
We note that this depends on the fact that J A o is formally self-adjoint

in the sense that
(v,J w)Xo = (J v,w) Xo (8.3.18)

2 2
if e.g. v,weH ' (I).
Remark 8.3.1. The situation here is slightly different from the one in
L
Section 8.2 inasmuch as we identify coker J here with Ro(J\ )' and A() 0
not with R^JXQ)^. Therefore, in the present situation, if IT denotes the

orthogonal projection onto ker J = coker J , we have A o A o
^(Âo^) = 0
2 2 2 2
only for v G H ' (I), but not for all v G H ' (I). This is for example
relevant for the argument of the proof of Lemma 8.2.2.
Regularity theory also implies that R(J ) Xo is closed. Namely, i f for

2 2
(Vn JnGN C (ker J A J 1
C W ' (I), we have
J\o^n fni
2 2 2 2
and f n converges to fo in L {I) x E , then ||f ||vi/ < (/) is bounded. n
By Rellich's Theorem 3.4.1, after selection of a subsequence, v n then

l 2
converges in W ' (I). Prom the equation, i.e.
1 .. d ( 1 \ . 1 1
V V Vn
2 nn + "J! , 2 t2 nn +
1 To 7TT2 = fn,
cosh ^ ' dt ^ c o s h i ^ A2 c o s h ^
2
we then see that v n also converges in L (I). Thus, v n converges in
2 2
W ' (I), and the limit VQ then satisfies
v
J\o o = /o-
Thus, the image of J\ 0 is closed. Thus, J A o is a Fredholm operator of

index 0.
Our aim is now to check that the assumptions of Theorem 8.2.2 hold.
2
In order to verify (8.2.42), i.e. ir(d u) ^ 0, according to Remark 8.2.1,
we need to compute o\7 , i.e. the second derivative of L \ . Starting from
Ao Q
(8.3.3) and inserting (8.3.6), i.e. no = Aocosht/Ao, we obtain
d
J T / w x ( 2
3Atanh^ . 2 \ 1 . 2
By (8.2.27), we have to project this onto the kernel of J\ (uo) and check 0
that the result is nonzero, for our Jacobi field v given in (8.3.12), i.e.
v = cosh tjA t / A s i n h t / A . Since here the projection TT is given by the
2 2
orthogonal projection in the Hilbert space L (I) x E onto ker J ( u o ) , Ao
which is generated by the Jacobi field v, we simply have to verify that

2
the L - product of dJ\ (uo)(v, v) w i t h v is nonzero. Thus, we compute
0
1
cosh A
3Atanh| , 3 0 I
3
/ | cosh ^ cosh' j
by an integration by parts
2
3v(t)
, (Atanh { ?)() - v(t)) dt.
cosh A-
Now with v = cosh j j sinh ^ , we have
A tanh j v(t) - v(t) = - cosh ^ ,

and so
(dJ (u )(v,v),v)
Xo 0 L2 = < 0.
Thus, indeed
3
ir(d u) a > 0. (8.3.19)
We finally consider (8.2.41). Thus, we have to verify that
r a
it{d\u) ^ 0, w i t h d\u = ^t)\t=o f suitable family
A of parameters.
t (8.3.20)
We start with (8.2.14), i.e. in the notations of Section 8.2
L n ( , A t ) + 7r(n(,At)) = .
At (8.3.21)
In the present case L \ is given by (8.3.6), and IT is the orthogonal
2
projection in L (I) onto ker J , the one-dimensional space generated
A o
by v(t) = cosht/Ao t/Ao sinht/Ao (see (8.3.12)), where Ao is so chosen

that v(l) = v(-l) = 0. Thus, this v can be taken as the of Section 8.2. 0
However, since
^(Acoshi) | A = A o =0
by choice of A (see (8.3.13)), we shall need to employ a variation of the

0
parameter somewhat different from the family At = A - f tfi employed in

Section 8.2. Here, we put
i / := A cosh-^
0 0
and choose the family X such that t
At cosh ^ = vo + tfi. (8.3.22)
We then differentiate (8.3.21) w.r.t. t at t = 0, = 0 to obtain
J (u )d u
Xo 0 x -f T J - ^ (d\u,v) 2v L = 0 (8.3.23)
v 2
\\ \\ L ( / )
d\u(l) = d\u(-l) = fi.

Then
dxU
(J\ (u )d u,v) 2
0 0 x L =j ^ c Q s ^ 2 _L ( W^J
+ dxU{t)Ht)dt
^ ^^0
Exercises 289
1 l
2-rd u(t)v{t)\ _
x v
cosh A q
integrating by parts and using J\ (uo)v 0 = 0 v(l)

0 = v(-l).
_ 2
A cosh ~ ^
0
< 0 for fi < 0.
= TT^TQ V)V ^ 0,
v 2
ll ll L ( / )
i.e. (8.3.20).
We thus have verified all the assumptions of Theorem 8.2.2 (for the
family X defined by (8.3.22) in place of the family At = Ao + tfi).
t
Theorem 8.2.2 thus describes the bifurcation behaviour of the solu

tions of (8.3.3) or (8.3.4), i.e. the critical points of (8.3.1), (8.3.2) near
Uo(t) = A o c o s h ^ : For boundary values u(l) = u(1) < A o c o s h ^ ,
there is no solution (at least in the vicinity of no), whereas for u ( l ) =
u(1) > A o c o s h ^ , we may find two solutions. Of course, this may
also be verified directly without going through all the abstract machin
ery of Section 8.2, but hopefully this example can serve to illustrate
the general scheme. The catenoids are frequently discussed in books on
the calculus of variations, e.g. O. Bolza, Vorlesungen uber Variation-
srechnung, Teubner, Leipzig, Berlin, 1909, or M . Giaquinta, St. Hilde-
brandt, Calculus of Variations, Springer, Berlin, 1996, I , p. 366 and
I I , pp. 263-70. A discussion in terms of bifurcation theory also in the
case of not necessarily symmetric boundary conditions (i.e. not requir
ing u(l) = u(1)) is given by H . Wenk, Extremverhalten der Stabilitat
von Catenoiden als Rotationsminimalflache, Diplom thesis, Bochum,
1994.
Exercises
8.1 How many parameters are needed for a complete description of

the bifurcation behaviour of the roots of a fourth-order polyno
mial?
8.2 Consider the problem of finding critical points
2
I(u) = J u(t)^/l + u(t) dt
U(K) = U(K) = 1
for a parameter n > 0. Determine the value no for which a

bifurcation occurs.
(Hint: This problem can be reduced to the one considered in
Section 8.3.)
2
8.3 Consider geodesies on S as in Chapter 2 of Part I . More pre
2
cisely, we take two points p,q G S w i t h distance d(p, q) = A,
and consider geodesic arcs between p and q of length A, i.e.
length minimizing arcs. What happens at A = 7r? Does this fit
into the framework described in Section 8.2?
9
The Palais-Smale condition and unstable
critical points of variational problems
9.1 T h e P a l a i s - S m a l e condition
In this chapter, we take up a direction that has already been presented in

Chapter 3 of Part I , namely the search for nonminimizing critical points
of variational problems. This chapter will consequently be independent
of Chapters 4-8 of the present Part I I . I n Section 3.1 of Part I , we pre
sented existence results for unstable critical points of functionals F of
1 d
class C on some finite dimensional Euclidean space R . We only needed
a coercivity condition on the functional guaranteeing that a critical se
quence ( x ) N (i.e. satisfying DF(x )
n n n 0, | F ( x ) | bounded) stayed
n
d
in a bounded set. The local compactness of R then allowed the extrac
tion of a convergent subsequence whose limit XQ satisfied DF(xo) = 0,
because of the continuity of DF. I n Sections 2.3 and 3.2 of Part I , we
also presented examples where variational problems could be reduced to
such finite dimensional problems. The domain was a little more compli
d
cated than E , but being finite dimensional, i t was still locally compact
so that we had no difficulties finding limits of subsequences for critical
sequences. I n the remainder of this book, however, we have had am
ple opportunity to realize that variational problems are typically and
naturally posed on some infinite-dimensional Hilbert or Banach space.
Such a space is not locally compact anymore w.r.t. its Hilbert or Banach
space topology, so that the previous strategy encounters a serious prob
lem. Also weak topologies do not help much as the functionals under
consideration typically are not continuous w.r.t. the weak topology. I f
one searches for minimizers, this problem can be overcome by introduc
ing convexity assumptions as we have seen in Chapters 4 and 8, but any
convexity assumption excludes the existence of critical points other than
minima.
291
292 The Palais-Smale condition
Nevertheless, the lack of compactness of the underlying space must

be compensated by an assumption on the functional that guarantees the
appropriate compactness of critical sequences. I n other words we do not
require the compactness of arbitrary bounded sequences on our space
which is impossible as argued but only of critical sequences. This is
the idea of the P a l a i s - S m a l e condition which we now formulate:
Definition 9.1.1. Let (V, ||-||) be a Banach space, F : V > E a func

1
tional of class C . We say that F satisfies the Palais-Smale condition,
abbreviated as (PS) , if any sequence ( x ) n N C V satisfying n
(i) | F ( x ) | < c
n for some constant c
(ii) \\DF(x )\\->0 n /orrwoo
contains a convergent subsequence.
Note that a limit XQ of such a subsequence satisfies DF(xo) = 0 (i.e.

is a critical point of F) because DF is continuous.
A direct consequence of the definition is:
L e m m a 9.1.1. Suppose F : V E satisfies (PS). Then for any a e E
K a := {x e V : F(x) = a, DF(x) = 0}
(the set of critical points of F with value a) is compact.
q.e.d.
We also have:
L e m m a 9.1.2. Suppose F : V > E satisfies (PS). For a G E, we put

Ua, :=
P | J {zeV\ ||x-*||<p} (p>0),
xeK a
N ,6 := {y e V | \F(y) - a | < 6 , \\DF(y)\\

a < 6} (6 > 0).
Then the families (U ) o and (N ^)s>o are fundamental systems of

ayP p> a
neighbourhoods of K (i.e. each neighbourhood of K contains some U ,

a a a p
and some N j). a
Proof. I f is clear that U , and N j are neighbourhoods of K for p > 0

a P a a
respectively <5 > 0. I t follows from the compactness of K that each a
neighbourhood of K contains some U , - Concerning the same prop

a a P
erty of the i V 5 , let us assume on the contrary that there exist a neigh
a <
bourhood U of K and a sequence (y ) ?$

a w i t h y e i V \ (UDN L) n ne n a a
for all n. (PS) implies that a subsequence of (y )nen n converges to some

yo e K C U, contradicting the openness of U.
a
q.e.d.
In our applications below, we shall also encounter the situation where

we want to find critical points of the restriction of some functional F to
the level hypersurface G(x) = (3 of some other functional G. For that
purpose, we shall need a relative version of the Palais-Smale condition
which we shall formulate only for the case of a Hilbert space:
D e f i n i t i o n 9.1.2. Let (H,< , >) be a Hilbert space, F,G : H - * R

1
functionals of class C , (3 G E . Suppose
DG{x) ^ 0 for all x with G(x) = (3.
We say that F satisfies (PS) relative to G = (3 if every sequence

( x ) n N C H with G(x ) = (3 and satisfying
n n
(i) | F ( x ) | < c
n for some constant c
0 for n + oo
2 (
\\DG(x )\\
n
contains a convergent subsequence.
A limit xo of such a subsequence then satisfies
G(x )
Q = f3 (9.1.1)
and
2
\\DG(x )\\ 0
i.e. is a critical point of the restriction of F to G(x) = (3. Of course, re

sults analogous to Lemmas 9.1.1 and 9.1.2 hold in the relative case. One
simply intersects the corresponding sets w i t h {G(x) = (3} and replaces
DF by its projection to that level set.
As in Sections 2.3, 3.1, 3.2 of Part I , in order to find critical points of
a functional, one needs to construct (local) deformations that decrease
the value of the functional except at or at least away from critical points.
We shall now do so in stages of increasing generality.
We start w i t h a functional
F : H R
2
of class C on some Hilbert space (H, (, )) that satisfies (PS). For each
u E H, DF(u) is a linear functional on H, and by Corollary 2.2.3, i t can

therefore be identified w i t h an element V F ( w ) of H , called the gradient
of F at u. Thus, VF(u) satisfies
2
DF(u)(VF(u)) = ||>F(u)|| (9.1.3)
\\VF(u)\\ = \\DF(u)\\. (9.1.4)
2 1
Since F is assumed to be of class C , DF and hence V F are of class C
in their dependence on u. I n particular, V F is locally Lipschitz.
We now consider the (negative) gradient flow induced by F :
i/>(u, t) = -VF(V>(u, 0 ) for t > 0 (9.1.5)

ot
V>(u,0) = u. (9.1.6)
Because of the Lipschitz property, by Theorem 2.4.2 and Corollary 2.4.2,

for small t > 0, there exists a unique solution tp(u, t) satisfying the
semigroup property
V>(M + s) = ^ ( M ) , * ) (9.1.7)
for sufficiently small 5, > 0.

Moreover,
ip(u,t) = u for all u w i t h V F ( u ) = 0, i.e. for all critical points of F .

(9.1.8)
Finally
F ( ^ ( u , i ) ) = F(u) + J* ~F{^{u, r))dr
= F(u) + jT DF(i/>(u,t))-îl>(u,T)dT
= F(u)- f \\DF(ip(u,T))\\ 2
dr by (9.1.5), (9.1.3)
Jo
< F(u) for t > 0, i f >F(u) = D F ( ^ ( u , 0 ) ) ^ 0, (9.1.9)
i.e. i f u is not a critical point of F .
Thus, we have found the prototype of a deformation that decreases the

value of F except at its critical points. For technical reasons, however,
the above flow will need some modifications and generalizations.
First of all, a solution of (9.1.5) need not exist for a l H > 0 because i t
may become unbounded in finite ' t i m e ' t . This can be easily remedied
by using the Lipschitz function
+ +
77 : M -> M
1 for 0 < s < 1
T](S) = { 1
for 5 > 1
s
and putting
VF(u) :=n(\\VF(u)\\)VF{u)
(i.e. V F ( w ) = V F ( u ) for | | V F ( u ) | | < 1 and | | V F ( u ) | | < 1 for all u) and

replacing (9.1.5) by
J^(M) = -VF(iK,t)). (9.1.10)
Of course, we still use (9.1.6).

Since VF(u) < 1 for all u, the solution of (9.1.10), (9.1.6) now exists
for all t > 0, and satisfies (9.1.7) for all s,t > 0. Equation (9.1.8) also
still holds, and as in the derivation of (9.1.9), we get
2
F(iKu,t)) = F(u) - T r / ( | | V F ( ^ ( u , r ) ) | | ) ||DF(^(u, r ) ) | | d r
Jo
< F(u) for t > 0,
if tx is not a critical point of F . More generally, we have
F(-0(u, t)) < F(V>(u, 5)) whenever 0 < 5 < t, for all u.
Next, we wish to localize the construction near a level a. Thus, for given
eo > 0 and a neighbourhood U of K we want to have a flow ip(u, t)a
with (9.1.7), (9.1.8) and also
i)(u,t) = u if|F(u)-a| > c , 0 (9.1.11)
and the following more explicit local decrease of the value of F : For
a E R, we put
F : = {veH
a I F(v) <a}.
We want to find 0 < e < e w i t h 0
^(F a + e \ U, 1) C F _ Q C (9.1.12)
^(f/,l)cF _ U[/, a e (9.1.13)
and of course also
< F(i>(u,s)) if 0 < s < t for all u. (9.1.14)
We let (p : E E be Lipschitz continuous w i t h
(p(s) = 0 for |a - s\ > e 0
<p(s) = 1 for |a s\ < ^~
0 < <p(s) < 1 for all 5
and replace (9.1.10) by
~*P(u, t) = -y>(F(V>(u, t)))VF(1>(u, t)). (9.1.15)
Again, a solution ip(u,t) exists for all > 0 and satisfies (9.1.7) for all
5,t > 0, as well as (9.1.8) and (9.1.14) (for the latter i t was necessary to
require (p > 0). (9.1.11) also is clear from the choice of (p. We now verify
(9.1.12), (9.1.13). I f 0 < c < f and u G F and i f F{i>(u, 1)) > a - e,
a + C
from (9.1.14)
|F(V>(tx, t)) ~a\<e for all 0 < t < 1,
and therefore
<p(F(i/>(u,t))) = 1 for all 0 < t < 1. (9.1.16)
As before, we may now compute
F(i>(u,l))
1
f d
= F(u)+ / F(V(w,r))dr
d T
Jo
; , 2
= F(u)- f v (i W,T))i7(||VF(V'(u,T))||)||Z?i!'W,r))|| dT
JO
2
<a + e- [ min(l,||DF(^(u,r)|| )(ir (9.1.17)
Jo
since we assume u G F a + C , using (9.1.16) and the properties of 77.
By Lemma 9.1.2, we may find 6, p > 0 w i t h

tfa, C U atP C t/ ,a 2 p CU (9.1.18)
(here, we are using (PS)!). From the definition of N j, thus Q
2 2
\\DF(ip(u T)\\
} > 6 whenever ip(u,r) ^ N f Without loss of gen a
erality 6 < 1. (9.1.17) then yields

2
F(i){u, 1)) < a + c - (meas {0 < r < 11 i/>(u, r ) iV }) 6 .
M (9.1.19)
From (9.1.18), we have for v G H \ U
dist(v,N s)
a I := inf w|| ) > p.
1
' wN ai6
Since
8
M
< 1, (9.1.20)
d t ^
therefore, i f u fi /, then also t/>(u, r ) ^ 7V 5 for 0 < r < p, and similarly,
a<
if ^ ( u , 1) fi t / , then also ip(u,r) fi N j for 1 - p < r < 1. Therefore, if

a
either tx fi U or -0(tt, 1) ^ (7, then
meas {0 < r < 1 1 ip(u, r) fi N j} a > p.
Thus, from (9.1.19), i f u fi U or i f ^ ( u , 1) fi U,

2
F(^(u, 1)) < a + e - p<5
2
<a~e i f we choose e < ^p<5 . (9.1.21)
2
Thus, for 0 < e < min(, | p 6 ) , we get (9.1.12), (9.1.13).
In conclusion, we have shown the following deformation result:
2
T h e o r e m 9.1.1. Let F : H > R be a C functional on a Hilbert space
H, satisfying (PS). Let a G R, and put
FA : = {v G H : F ( v ) < a } ,
Ka := {v e H : = a, DF(v) = 0} .
Let eo > 0 and a neighbourhood U of K a be given. Then there exist

0 < e < to and a continuous family
rj): H x [0, oo) -+ H
with the semigroup property ip(ijj(u, s),t) = ip(u,s + t) for all s,t > 0,
u H and with
(i) -0(w, 0) = it /or all u E H

(ii) F(tp(u,t)) is nonincreasing in t for all u G H
(iii) tp(u,t) = u /or a// t whenever DF(u) = 0, i n particular for u G
(iv) 4>(u,t) = u whenever \F(u) a\ > e , for all t 0
(v) </>(F \ U, 1) C F _ , </>(F , 1) C F _ U [/

Q+c a a+c Q C
(vi) IfF(u) is even (i.e. F(u) = F(u) for all u), then also F(IJJ(U, t))
is even in u for all t (i.e. F(ip(u,t)) = F(ip(u,t))).
(Property (vi) follows from the construction: A l l the auxiliary functions

are invariant under replacing u by u i f F is even, and V F ( - w ) =
VF(u) in the even case.) q.e.d.
C o r o l l a r y 9 . 1 . 1 . If under the preceding assumptions, F has no critical

point with value a, i.e. K = 0, then there exist a deformation I/J with
a
the preceding properties and
V> ( F a + C , 1) C F -a for some e > 0. (9.1.22)
Proof. I f K a = 0, we may choose U = 0 in Theorem 9.1.1.

q.e.d.
We shall now extend Theorem 9.1.1 in two directions. First, we con

2
sider the relative case, where in addition to F , we have another C
functional F : H R w i t h
DF(x) ^ 0 for all x with G(x) = 0,
for some given value (3 G R. We wish to find critical points of the re

striction of F to G = /3. We assume that F satisfies the relative (PS)
condition of Definition 9.1.2 on G = f3. We then perform the preceding
construction w i t h
G
V F(u) := VF(u) - ( V i ^ V G ( K ) ) (9.1.23)
K K J 2 K J K
' l|VG()|| '
in place of V F ( u ) . We then have
|GWM))
G G
= -<p(F(i,(u, t)))r,(\\V F(u)\\) (V FW(u, t)), VG(i>(u, t)))
from the chain rule and the analogue of (9.1.15)
= 0,
G
since (V F(v),VG(v)) = 0 for all v G H. Therefore, the flow tp(u,t)
now leaves G = /3 invariant. We obtain:
2
T h e o r e m 9.1.2. Let F,G : H R beC functionals on a Hilbert space
( i f , (, )) with F satisfying (PS) relative to G = (3. Let a G R,
:= {veH\F(v)<a,G(v) = (3},
G
K>P := {v G H | F(v) = a, G(v) = (3, V F(v) = 0} .
Let e > 0, and let U be a neighbourhood of K^ in {G(v) = /3}. Then

0
there exist e > 0 and a continuous semigroup family
4,:{G(v) = p}x[0,oo)^{G(v) = l3}
satisfying
(i) ip(u, 0) = u for all u G {G(v) = (3}

(ii) F(ip(u,t)) is nonincreasing in t
(iii) ip(u, t) = u for all u G
(iv) ip(u,t) = u for all t if \F(u) a\ > eo
(v) \U,1)C F, i>{Ff , 1) C F^ UU t t
(vi) If F and G are even, so is F(i/>(-,)) for all t.
Secondly, we wish to extend the preceding construction to functionals

on Banach spaces. For a functional on a Banach space, in general one
does not have a good notion of a gradient. We therefore need to introduce
Palais' concept of a pseudo-gradient:
D e f i n i t i o n 9.1.3. Let (V, ||-||) be a Banach space, U C V, F : U -+ R

1
afunctional of class C . A pseudo-gradient vector field for F is a locally
Lipschitz continuous vector field v : U V satisfying
(i) H ) | | < m i n ( l , | | D F ( u ) | | )
U
2
(ii) DF(u)(v) > Imin(|| DF( )||,||F( )|| )
J M M
for all u G U.
1
L e m m a 9.1.3. Let F : V R be a functional of class C on the
Banach space V. Then F admits a pseudo-gradient vector field on
V':={ueV \ DF(u)^0}.
Proof. For each u G V, we can find w = w(u) with
|HI <min(l,||DF(u)||) (9.1.24)

2
DF(u)(w) > \ min(||DF(u)||,\\DF(u)\\ ). (9.1.25)
1
Since DF is continuous (as we assume F G C ) , w satisfies (9.1.24),
(9.1.25) also for all v in some neighbourhood N of u. Since {N : u G u u
V'} is an open covering of V, it possesses a locally finite refinement

{M }a f . Let
a G /
p (v)
a :=dist(t;,V"\M ). a
f T h i s holds for any open covering of a paracompact set, see e.g. J . Dieudonne,
Grundziige der Modemen Analysis, 2, Vieweg, Braunschweig, second edition, 1987,
pp. 26-9; V is paracompact for example because it is metrizable.
pa is Lipschitz continuous, and p (v) a = 0 for v M . We put a
/ x Pa(v)
Since each v is only contained in finitely many M ^ , because of the local

finiteness of the covering, the denominator of (p is a finite sum. ((f )aei a a
is a partition of unity subordinate to { M } , i.e. 0 < ip < 1, (p = 0 a a a
outside M , Ylaei a = Also, the (p are Lipschitz continuous. Then

a
V U U W U S O m e U e M
() '' = YlalV<*( ) ( <*) <*
is a convex combination of vectors satisfying (9.1.24), (9.1.25) and hence
satisfies these relations, too.
v(u) thus is a pseudo-gradient vector field for F.
q.e.d.
1 2
Note that we only need to require F G C , and not F G C , in order
to construct a locally Lipschitz pseudo-gradient field. We then have the
1
following deformation for C -functionals on Banach spaces.
1
T h e o r e m 9.1.3. Let F : V > E be a C -functional on a Banach space
V satisfying (PS). Let a G E , eo > 0, U a neighbourhood of K as a
in Theorem 9.1.1. Then there exist 0 < e < 1 and a continuous family
I/J : V x [0, oo) V satisfying the semigroup property w.r.t. t > 0, and
(i) ip(u, 0) = u for allu eV

(ii) F(ip(u, s)) < F ( ^ ( u , t)) whenever 0 < t < s, u G H
(iii) ip(u, t) = u for all t whenever DF(u) 0
(iv) ip(u,t) = u whenever \F(u) - a\ > e , for all t 0
(v) ^(F
a+e \ U, 1) C F _ , 1>(U, 1) C F _ U U
a c a c
(vi) If F(-) is even, so is F(ip(-,t)) for all t.
Proof. The proof is the same as the one of Theorem 9.1.1, replacing
VF(u) by a pseudo-gradient vector field v(u) except for the following
technical point: Lemma 9.1.3 asserts the existence of a pseudo-gradient
field only on {x G V \ DF(x) ^ 0}. We therefore have to choose another
Lipschitz continuous cut-off function 7 : V E with 0 < 7 < 1, 7(1*) =
0 i f u G J V j , 7(14) = 1 for u G V \ N ^. We may then consider
a j a
= - 7 ( ^ , < ) ) v ( F ( V ' K t ) ) ) r y ( | | t ; ( V ( , t ) ) | | ) ( V ( , < ) ) (9-1.26)

U
with </?, 77 as before. This has the additional effect that

dj>(u,t) _
dt
whenever tp(u,t) G JV , which is a neighbourhood of K , while the

Q( a
evolution is the same as before (with v(u) in place of V F ( w ) ) outside

N j- This cut-off near K does not affect the rest of the construction.
a a
If F is even, we may also choose 7 even.

However, there might still exist critical points of F in F \ N s- I n a + C ay
order to take account of those, we strengthen the requirements on the

above cut-off function (p to
p(s) = 0 for |a - s\ > min(e , - ) 0
(p(s) = 1 for |a - s\ < m i n ( y , - ) .
W i t h such a </?, the right-hand side of (9.1.26) vanishes near any critical
point of F , and i t is therefore defined on all of V. I f we then also impose
the additional restriction
4
everything works out as before.
q.e.d.
I t is possible, and not overly difficult, to extend Theorem 9.1.3 to the

relative case and to obtain a result analogous to Theorem 9.1.2. Here,
however, we refrain from doing so.
9.2 T h e m o u n t a i n pass t h e o r e m
W i t h the help of the deformation theorems of the previous section, one
may easily derive existence results for critical points of a functional sat
isfying (PS). To illustrate this point, we start with the trivial
1
L e m m a 9 . 2 . 1 . Let F : V R be a C functional on a Banach space
satisfying (PS). If
a := inf F(u) > - 0 0 ,
then F possesses a critical point UQ with value a (i.e. F(uo) = a,

DF(u )=0).
0
Proof. Suppose that K = 0. Then U = 0 is a neighbourhood of K . Let

Q a
e > 0 be arbitrary. Choose e as in Theorem 9.1.3. From the definition

0
of a,
F a + C ^ 0 , F _ = 0.
a c
Therefore, it is impossible that as Theorem 9.1.3 (v) asserts, the defor

mation tp(-, 1) maps F into F _ . This contradiction implies K ^ 0 ,
a + C a c a
which means the existence of the desired critical point.

q.e.d.
Of course, the methods presented in Chapter 4 yield more general

existence results for minimizers of variational problems. The strength
of the Palais-Smale approach rather lies in its capability of producing
nonminimizing critical points. To demonstrate this, we now present the
mountain pass theorem of Ambrosetti-Rabinowitz.
1
T h e o r e m 9.2.1. Let F :V M be a C functional on a Banach space
(V, ll-ll) satisfying (PS). Suppose F(0) = 0 and
(i) 3p > 0,/3 > 0: F(u) > (3 for all u with \\u\\ = p
(ii) 3t*i with \\u\\\ > p and F(u) < (3.
We let
r := { C([0,1], V) | 7(0) = 0,7(1) = Ui} .
7
Then
a := inf sup F(I(T)) (> 0)
7r r [ 0 ) 1 ]
is a critical value ofF (i.e. there exists UQ with F(u ) 0 = a, DF(UQ) = 0).
Proof. Suppose again that K = 0 , and take the neighbourhood U = 0

a
of K . We let e = min(/?, f3 - E(u\)).

a 0 Choose e as in Theorem 9.1.3.
Prom the definition of a, there exists 7 G T with 0
sup F ( 7 O ( T ) ) < a + e,
T[0,1]
while no such 7 can satisfy
sup F(y(t)) < a - e,

[0,11
i.e. satisfy 7Q0,1]) C F _ . However, if we apply the deformation

a c ?/>(, 1)
of Theorem 9.1.3, we obtain a path
7 ( r ) :=V>(7o(r),l)CF _ a c
w i t h 7(0) = 7o(0) = 0 and 7(1) = 7o(l) = u\ by choice of e . This 0
contradiction implies K ^ 0 , i.e. the existence of the desired critical

Q
point.
q.e.d.
Let us summarize the essential features of the preceding reasoning:
(1) One chooses a family of sets, here T, that exploits some properties
of F and is invariant under the deformation ?/>(, 1).
(2) This family yields a minimax value a.
(3) a can be estimated from above w i t h the help of any member of
our family r (a < s u p [ j F ( 7 ( t ) ) ) for any 7 G T), and from
r 0 x
below through the constraints that the members of T have to

satisfy (in Theorem 9.2.1, every 7 G T intersects dB(0, p), and
therefore a > f3 > 0, and therefore in particular, the critical
point produced is different from 0).
(4) A reasoning by contradiction, based on the deformation theorem,
shows that a is a critical value.
As an application of the mountain pass theorem, we consider the fol

lowing example:
d
T h e o r e m 9.2.2. Let Q C R be a bounded domain, 2 < p < ^
(respectively < 00 for d = 1,2). Then the Dirichlet problem
p 2
Au+ \u\ ~ u = 0 in Q (9.2.1)
u = 0 on dn (9.2.2)
admits at least two nontrivial (weak) solutions ('nontrivial' means not

identically 0).
Proof. I f u is a solution, so is u. Therefore, it suffices to verify the ex

istence of one nontrivial solution. (9.2.1), (9.2.2) are the Euler-Lagrange
y2
equations in HQ (ft) for the functional
2
F(u) = \ f \Du\ -- f \uf. (9.2.3)
^ JQ P JQ
2 2
This functional is a continuous functional on HQ' (Q), because J \Du\ Q
p
clearly is continuous there, and J \u\ too, because of the Sobolev Em
bedding Theorem 3.4.3 as we assume p < -j^. F is also differentiable,
with
p 2
DF(u)(<p)= j Du Dip - j \u\ ' u(p. (9.2.4)
JQ JQ
Again
(p 1 / Du Dip
JQ
,2
clearly is continuous on HQ (Q), whereas
Jn
is continuous, because we have by Holder's inequality
p-i i
2 p P P (9.2.5)
I\ur u<p <([\u\A (f M )
<co\\u\\ i\ \\<p\\ i,2 p
H {Q) H {Q)
(9.2.6)
by the Sobolev Embedding Theorem 3.4.3
2 1
for some constant c . Thus F : H^ (Vt)
0 R is of class C .
We shall verify the Palais-Smale condition for F: Suppose ( u ) n n N C
H Q ' ( 0 ) satisfies
2
| F ( u ) | < C i for some constant c\

n (9.2.7)
DF(u ) n ^ 0 for u ^ oo. (9.2.8)
Thus
< ci (9.2.9)
and
sup / Du Dip n u y>| 0

n for n oo.
1,2<1 'O
(9.2.10)
a n
In (9.2.10), we use (p = n^)"^ 2 d obtain
2 p
- J \Du \ n +j \u \ n < c \\u \\ , .
2 n H1 2 (9.2.11)
Since p > 2, (9.2.9) and (9.2.11) imply
2
J \Du \ n < c ||w||i,2 + c .
3 A (9.2.12)
Since by the Poincare inequality (Theorem 3.4.2)
=/ +j \Du \ n
2
<c J 5 \Du \
n
2
, (9.2.13)
we conclude from (9.2.12)
C
HnllHi.(fl) ^ 6 - (9-2-14)
Thus, any 'critical sequence' ( u ) n N is bounded. We now claim that n
such a sequence ( u ) n N contains a convergent subsequence, thereby

n
completing the verification of (PS). We need to show that, after selection

of a subsequence,
2
j \Du n Du \ m 0 for n, ra oo (9.2.15)
(using again the Poincare inequality as in (9.2.13)). Now
J Du D(u
n n - Um) - j |u | ~
n
p 2
u (u
n n - u) m 0 for 71, ra > oo
(9.2.16)
by (9.2.10), (9.2.14).
By the Rellich-Kondrachev theorem (Corollary 3.4.1), we may also as
sume (by selecting a subsequence) that (u ) ^ is a Cauchy sequence in n ne
1^(0,). Then, using Holder's inequality as in (9.2.5),

p-i i
Jy \u \ ~ n
P 2
U (un n - U)
m < (^J \u \ ^j (^J n
P
|tZ -U | ^
n m
P
^ 0
forn,ra->oo. (9.2.17)
J Du n - D(u n u) m 0 for ra, n > oo,
which implies (9.2.15). We have thus verified (PS) for F. We shall now
check the remaining assumptions of Theorem 9.2.1. First of all,
F(0) = 0.
Recalling that by the Sobolev Embedding Theorem 3.4.3 (and the

Poincare inequality, see (9.2.13))
i
M C |D |!
(/n f '(/ " )"' n
we have
F(u) > ( i - C8 |||| ) I M I i . . ( n > (n) > 0 > o

if IM|#i.2(Q) = p is sufBciently small.
i2 p
Finally, take any u G HQ (Q) 2 with / n \u \ 2 ^ 0. Then for sufficiently
large A > 0, u\ Xu satisfies 2
D U 2 P
^(l) = T / \ 2 \ ~ - f \U2\ <0.
2
J P Jn
We have now verified all the assumptions of the mountain pass Theorem
9.2.1, and we consequently get a critical point u of F w i t h
F(u) > (3 > 0.
This is the desired nontrivial solution. (In fact, regularity theory im

plies that any weak solution of (9.2.1) is smooth in fi, see e.g. Gilbarg-
Trudinger, loc. cit.)
q.e.d.
Remark 9.2.1. By the same method, we can also treat the equation
p 2
Au - Xu + \u\ ' u = 0 for any A > 0. (9.2.18)
9.3 Topological indices and critical points

In Section 3.2 of Part I , we have seen an example where a topologi
cal construction permitted to deduce the existence of more than one
(unstable) critical point of a functional. I n the present section, we first
give an axiomatic approach to such constructions and then apply this
in conjunction w i t h the Palais-Smale condition to a concrete variational
problem to show the existence of infinitely many solutions.
Such global topological constructions originated w i t h the work of
Lyusternik. Contributors also include Schnirelman, and, more recently,
Rabinowitz, and many others. The reader will find detailed references
in the monographs quoted at the end of this chapter.
Definition 9.3.1. Let X be a topological space, F : X > R continuous,

x X is called a special point for F, with value a,
x G spec Fa (a then is called a special value)
if x is contained in all A C X with the following property: For each open

U D A there exist e = e(U) > 0 and a continuous
t/> : X x [0,1] ^ X
satisfying
(i) \j)(y,0)=y foryeX

(ii) F(^(y, t)) < F(^(y, s)) for all y G X, 0 < s < t < 1
(iii) For every y G X \ U with
F(y) < a + e,
we have
F(^(,l))<a-c.
Of course, the ip of the preceding definition is an abstract version of

the deformations constructed in Section 9.1, and the notion of special
point is a topological version of the notion of critical point.
Remark 9.3.1. Since the composition of any two deformations ipi,ip 2
satisfying the properties of Definition 9.3.1 continues to satisfy these

properties, the intersection of any two sets A\, A still satisfies the prop
2
erty expressed in Definition 9.3.1 i f Ai,A do. Therefore, i f spec F 0,

2 a
we may take U = A = 0 in Definition 9.3.1 and find a deformation ip

that satisfies (i)-(iii) for all y G X.
In order to illustrate the notion of special point as well as the topo

logical constructions to follow, we now present the simple:
L e m m a 9 . 3 . 1 . Let F : X E be a continuous function on the topo

logical space X. Let M be a (nonempty) class of nonempty subsets of X.
If spec F = 0, we require that M is invariant under the
a deformations
considered in Definition 9.3.1:
IfAeM, then also ip(A, 1) G A i . (9.3.1)
Suppose
-oo < a = inf sup F(y) < oo. (9.3.2)

AeM yeA
Then a is a special value for F, i.e. there exists
xo G spec F. a
Proof. Suppose spec F = 0. According to the preceding remark, we

a
may then take U = 0 and find ^ : l x [ 0 , a n d e > 0 w i t h
F(ip(y, 1)) < a - e whenever F(y) < a + e. (9.3.3)
We may find AQ G M w i t h
sup F(y) < a + e, (9.3.4)

yA 0
but no A G M can satisfy
s u p F ( y ) < a-e. (9.3.5)

yA
However, i f we take A\ := ip(Ao, 1) then A\ G M. by assumption, and

by (9.3.3)
sup F(y) < a - e,
yAx
contradicting (9.3.5). Thus, spec F ^ 0. a
q.e.d.
In order to obtain the existence of further special points, we now shall

introduce the notion of a (topological) index. Such an index is based
on symmetry or invariance properties of the functional under considera
tion. Here, we only consider the case of the simplest nontrivial symmetry
group, namely Z , although the subsequent constructions easily gener
2
alize to any compact group G. We thus make the following symmetry

assumptions:
X is a topological space with a nontrivial involution, i.e. there exists

a continuous map j : X > X , j ^ id, w i t h
2
j = id.
F : X E is continuous and even, i.e.
F(j(x)) = F(x) for all x G X.
M := {A C X I j(A) = A and for all x G A,j(x) ^ x

(i.e. A contains no fixed points of j ) } .
We now also require ip(j(x),t) j(ip(x, t)) for all the deformations of
Definition 9.3.1.
D e f i n i t i o n 9.3.2. An index for ( X , F) is a map
i : M -+ { 0 , 1 , 2 , . . . , 0 0 }
satisfying for all A, A\,A 2 G M:
(i) i(A) = 0 ^ A = 0
(ii) A finite (A ^ 0) i(A) = 1
(iii) 2(^1 U A ) < i(Ai) + i(A )
2 2
(iv) Ai C A =2 < i(A ) 2
(v) t(i4) < i(j(A))

(vi) A compact = 3 neighbourhood U of A in X with U G M,
i(A) = i(U) < oo.
For n { 0 , 1 , 2 , . . . , oo}, we put
M :=
n {AeM\ i(A) >n}.
Remark 9.3.2. More precisely, one should call an i as in Definition 9.3.2

an index for (X, F, Z 2 ) , in order to specify the symmetry group involved.
For n E { 0 , 1 , 2 , . . . , 00}, we define
an := inf sup F(y).

AM n yeA
T h e o r e m 9.3.1. Suppose the above symmetry assumptions hold, an in

dex i for ( X , F) exists, and
00 < a n < oof
(i) Then
spec Q n F^0 (9.3.6)
(ii) If furthermore for some k > 1, a n = c* +i = . . . = a _|_fc, ^ e n

n n
spec F 25 infinite.
Qn
Proof We note that property (v) of Definition 9.3.2 implies that M n
is invariant under (symmetric) deformations ip. Therefore, Lemma 9.3.1

implies spec F ^ 0 . For the second statement, we claim that for Ao =
Qn
spec F ,
Qn
i(Ao) >k+l. (9.3.7)
If k > 1, property (ii) of Definition 9.3.2 then implies the existence of

infinitely many special points w i t h value a . n
Suppose on the contrary that
i(A )
0 < k. (9.3.8)
By Definition 9.3.2 (vi), we may find a neighbourhood U of Ao w i t h

U e M and
t(i4o) = i(U). (9.3.9)
Since A consists of special points, we may find a (symmetric) deforma

0
tion ip w i t h
F(ip(y, 1)) < a n - e for all y e X \ U with F(y)<an +e
f Since the infimum over an empty set is 00, this contains the assumption Mn / 0-
for some e > 0. Since a n a n + k , we may find A e M +k n with
sup F(y) < a n + e,

yA
hence
sup F(z)<a -e. n (9.3.10)
zeif(A\u,i)
We have
i(A\U)>i(A)-i(U) by (iii)
> n + k - k, using (9.3.8), (9.3.9), A e M+
n k
= n.
Thus
A\UeM , n
hence A \ ( 7 ^ 0 by (i). Since, as noted in the beginning, M n is invariant

under we get
i>(A\U,l) G M,n
hence
sup F(y) > a , n
yV(^\t/,l)
contradicting (9.3.10).
q.e.d.
I n order to apply the preceding considerations, we need to construct

an index w i t h the properties listed in Definition 9.3.2. We shall present
here Coffman's version of the genus of Krasnoselskij.
D e f i n i t i o n 9.3.3. Suppose the symmetry assumptions stated before Def

inition 9.3.2 hold. The genus of A ^ 0, A e M is defined as follows:
n
gen(A) := inf { n G { 1 , 2 , 3 , . . . , 0 0 } | 3 continuous f : A -> R \ { 0 }
with f(j(x)) = f(x) for all xeA}
while g e n ( 0 ) : = 0.
As an example, we state:
n 1 n
L e m m a 9.3.2. The genus of the unit sphere S' "" = {||x|| = 1} in R
(with involution j(x) = x) is equal to n.
n 1 n
Proof. The inclusion map S' "" R satisfies the properties of Def
n _ 1 n _ 1
inition 9.3.3, and so g e n ( 5 ) < n . I f n > 2, 5 is connected,
and therefore, by the mean value theorem, there is no continuous map
1 1
/ : S^- -+ R ^ j o } w i t h f(-x) = -f(x) for all x. Hence g e n ^ " ) > 2.
In fact, by the Borsuk-Ulam theoremf, there is no such continuous map
m n 1
to R \ {0} w i t h m < n. Therefore, g e n ( 5 ~ ) > n .
q.e.d.
C o r o l l a r y 9 . 3 . 1 . The genus of the unit sphere S := {x V : \\x\\ = 1}
in an infinite dimensional Banach space (V, ||-||) is oo.

n
Proof. For any n-dimensional subspace V of V,
n
gen(S) > gen(S f l V ) >n by Lemma 9.3.2 .
q.e.d.
T h e o r e m 9.3.2. The genus as defined in Definition 9.3.3 is an index

in the sense of Definition 9.3.2.
Proof. We need to check the properties (i)-(vi) of Definition 9.3.2.
(i) is obvious.
(ii) I f A A I is finite, then A is of the form {x ,j(x ) \ v 1 , . . . , k} 1/ 1/
1
for some k. We define / : A -+ R \ {0} by f(x) = 1, f{j{x )) = u
- 1 for all v (of course, we may assume x ^ j(x ) for all fi, v). p u
(hi) Let gen(A) = n < oo, v = 1,2, and let the continuous f : v
n
A -> R " \ { 0 } satisfy U(j(x))
v = ~U(x) for all x. By the
Tietze extension theorem^, f can be continuously extended to
v
n
f v :X-+R ".
By considering \(fv(x) ~ fv(j{x))) in place of f , we may assume

v
that the extension still satisfies
/ ( j ( x ) ) = -f {x) v for all x.

n i + n 2
The map ( / i , / ) : A U A
2 x 2 R \ {0} then shows that
gen(Ai U A ) < n i + n = gen(Ai) 4- gen(A ).

2 2 2
(iv) is obvious.
(v) follows, since / o j shares the necessary properties w i t h / .
f See e.g. E . Zeidler, Nonlinear Functional Analysis and its Applications, I , Springer,
New York, 1984, p. 708, for a proof.
$ See E . Zeidler, loc. cit., p. 49.
(vi) Let A M be compact. Since j(x) ^ x for all x A (by the

properties of A i ) , for each x A, we may find a neighbourhood
/(#) w i t h U(x) n j(U(x)) = 0. Since A is compact, i t can be
covered by finitely many such neighbourhoods U , v = 1 , . . . , n.
v
For each we choose a continuous function 0 for x /|,, <p (x) = 0 for sc E X \ t/. We then define
u
n n
h = (h\...,h ): A^R \{0}by
^( ).= J
x forxt/^
\ <Pv{x) for x A \ U , in particular for x
y j{U ).
u
(Since every x A is contained i n some we have h(x) ^ 0 for

all x A . )
Thus gen(A) < n < 0 0 .
If A M is compact with gen(A) = n, and
n
/ : A -> R \ {0} is continuous with f(j(x)) = -/(x),
n
we may extend / as before to / : X R (with the same symme
try property). Since A is compact, so is / ( A ) , and therefore, we
n
may find an open neighbourhood V of f(A) w i t h V C R \ { 0 } .
l
Then U := f (A) satisfies
n = gen(A)
< gen(U) by (iv)
n
< n since J(U) is contained in V C R \ { 0 } .
Thus gen(U) = gen(A) as required.

q.e.d.
We may now obtain a general existence theorem for critical points of

functionals satisfying (PS):
2
T h e o r e m 9.3.3. Let F, G : H R be C functionals on a Hilbert
space (H, (, )) that are even, i.e. F(x) = F(x), G(x) = G(x) for all
x H. Suppose F satisfies (PS) relative to G = (3, and is bounded from
below. Let
M:={Ac{G(x) = (3}\0<A and (xeA^-xeA)}.
Let 7 0 : = sup{gen(if) | K M compact} (< 0 0 ) . Then F possesses at

least 70 critical points relative to G = (3.
Proof. Since (PS) holds, by Theorem 9.1.2, all special points (in the
sense of Definition 9.3.1) for the restriction of F to X := {x H | G(x) =

0) are critical points for F relative to G = (3. Hence, it suffices to pro
duce 70 special points of F on 1 . Let
a n := inf sup F(x).

AM,gen(A)>n XA
Since F is bounded below, and since in the definition of 70, we only

consider compact sets, we have
00 < a n < 0 0 whenever n < 7 0 .
By Theorem 9.3.2, we may apply Theorem 9.3.1 to the genus as an index.

We have in fact
00 < ot\ < cx<i < < a n < - - - < 0 0 whenever n < 7 0 .
If we always have strict equality, then the
xn spec F Qn
produced by Theorem 9.3.2 (i) are all different, because their values
F(x )
n are all different. I f however any two such numbers a _ i and a n n
are equal, then by Theorem 9.3.2 (ii) we even obtain infinitely many
special points. Thus, in any case, we have at least 70 special, hence
critical points.
q.e.d.
As an application of Theorem 9.3.3, we consider the example of the

previous section:
d
C o r o l l a r y 9.3.2. Let ft C R be a bounded domain, 2 0, the Dirichlet
problem
p 2
Au - Xu + \u\ ~ u = 0 infl (9.3.11)
u = 0 on dft (9.3.12)
admits infinitely many (weak) solutions.
Proof. We consider the even functionals
F{u) = \j^(\Du\ 2
+ \u ) 2
G(u) = lf
P Jn
\u\ . p
We claim that F satisfies (PS) relative to G = 1. The proof is similar

to the argument employed for the demonstration of Theorem 9.2.2: let

(w )nN be a critical sequence, i.e.
n
F(u ) n < ci (9.3.13)
(DGMDFM)
0 forn->oo (9.3.14)
\\DG(u )\f n
2
where all norms and scalar products are from H Q ' ( Q ) . From (9.3.13)
(and the Poincare inequality in case A = 0), we obtain
(9.3.15)
We obtain as in the proof of Theorem 9.2.2 (cf. (9.2.5)), by using Holder's

inequality, that
\DG(u )(u
n n - u )\ m = j \u \ n
p 2
u (u -Um)
n n
2^1
| U n | P P U P P 9 3 1 6
" ( / ) ( / l ^ - - | ) ' ( - - )
Since p < from (9.3.15) and Sobolev's Embedding Theorem 3.4.3,

p
we conclude that / | u | is bounded, whereas (9.3.15) and the Rellich-
n
Kondrachev theorem (Corollary 3.4.1) imply that (u ) ^ is a Cauchy n ne
p
sequence in L ( f i ) . Thus, from (9.3.16)
DG(u )(u
n n Um) > 0 for n, m oo. (9.3.17)
Also
\DG(u )(w)\
n
\\DG(u ) n sup
\DG(u )(u )\ n n
> ||U||1,2
P
!K\
(9.3.18)
||n||fl-.a
> 0 from (9.3.15) and
- / \u \ n
p
= G(u ) n = 1. (9.3.19)
PJ
2
Prom (9.3.17), (9.3.18) we conclude that there exist h n m e H^ (ft) with
DG(u )(u
n n - Um + h ) nm = 0 for all n , m (9.3.20)
l l ^ n m ! ! / ^ - 2
0 for n, ra oo. (9.3.21)
Therefore, from (9.3.14)
DF(u )(u n n - u m 4 h ) nm -+ 0,
i.e.
j {Du n (D(u n - u) m 4 Dh ) nm 4 Xu {u n n - u m + h )) nrn 0
for n, ra ^ oo
and because of (9.3.21) then also
J (Du (D(u n n - u )) m 4 Xu (u n n - Um)) 0.

This implies
2 2
J (\{D(u n - u )\ m 4 A | ( u - u )\ ^ n m -+ 0 forn,ra->oo,
,2
and consequently, ( u ) N is a Cauchy sequence in H o ( n ) . This verifies
n n e
(PS) relative to G = 1.
In order to apply Theorem 9.3.2, we thus only need to check that in

the present case, 70 = 00. However,
p
is the intersection of a sphere centered at the origin in L (ft) with the
,2
subspace HQ (Q). Therefore, the argument of Lemma 9.3.2 easily im
plies 7 0 = 0 0 . Theorem 9.3.2 thus produces infinitely many solutions
n ! 2 y n >
\\DG(u )\\ n
i.e. w i t h
= (DG(u ),DF(u )) n n
a
^' ||r>G()|| '
weak solutions of
P 2
Au n - Xu 4 fi \u \ ~ n n n u =0 n in ft
u n = 0 on dft.
2
If we choose v with z / P ~ / i = 1, then v := v u
n solves (9.3.11),
n n n n
(9.3.12) weakly. Again, we remark that elliptic regularity theory implies

that all u and v are smooth in fi, so that in fact we obtain classical
n n
solutions of (9.3.11), (9.3.12).

q.e.d.
In Theorem 9.2.2 and in Corollary 9.3.1, we had imposed the restric

tion
2d
p < - ( i case d > 3) , n
d 2
and the reader may wonder whether this is necessary. To pursue this
question, we shall now discuss the theorem of Pohozaev:
d
T h e o r e m 9.3.4. Let fi C R be a smooth domain which is strictly star
d
shaped w.r.t. 0 R (this means that the outer normal v of Vt satisfies
(x, v(x)) > 0 for all x dft). Then for X > 0, any solution of
Au - Xu - f M ^ u = 0 in fi (9.3.22)
u = 0 on dft (9.3.23)
vanishes identically.
We shall present a complete proof only for A > 0 and for smooth
solutions u (elliptic regularity implies that any weak solution of (9.3.21),
(9.3.22) is automatically smoothf on fi, but the present book does not
treat this topic):
a n d o b t a i n
We multiply (9.3.22) by YlLi
0 = (Au - Xu + \u\&* uj E ^ ' J ^ (9.3.24)
= div ( W g ~^f~r M 2
+ w*
2 2
+ ^ | D u | + ^ H - - ^ H ^ . (9.3.25)
By (9.3.23), we have w = 0 o n Ofi, hence also X > * | * T = E ^ V ^

1 d
(y = ( z / , . . . , v ) is the exterior normal of fi). Integrating (9.3.25) there
fore yields
2
d-2 f _ l 2 Ad f , l 2 d-2 f , l A W 9u'
/an ^
2>V = o.
(9.3.26)
f See for example Appendix B in M . Struwe, Variational Methods, Springer, Berlin,
2nd edition, 1996.
On the other hand, multiplying (9.3.22) by u leads to
(9.3.27)
Jn Jn Jn
(9.3.28)
Jn Jan du
I f A > 0, this implies u = 0, hence the result. (If A = 0, one still concludes
that | ^ = 0 on dft. Since also u = 0 on dft by (9.3.23) one may invoke a
unique continuation theorem for solutions of elliptic equations to obtain
u = 0 in ft. We omit the details.)
q.e.d.
Theorem 9.3.4 implies that for p j ~ in Theorem 9.2.2 and Corol

lary 9.3.2, the Palais-Smale condition no longer holds. Namely, i f it did,
the proofs of those results would yield the existence of nontrivial solu
tions. I t also shows that i f the Palais-Smale condition fails the whole
scheme developed in the present chapter for producing critical points
breaks down.
Since for p < (PS) does hold, the case p can be considered
as as limit case for (PS). I n fact, such limit cases of the Palais-Smale con
dition occur in many variational problems that are of importance in Rie
mannian geometry, e.g. the Yang-Mills functional on a four-dimensional
Riemannian manifold, two-dimensional harmonic maps, surfaces of con
stant mean curvature, the Yamabe functional etc. The interested reader
is for example referred to
K. C. Chang, Infinite Dimensional Morse Theory and Multiple Solution

Problems, Birkhauser, Boston, 1993,
J. Jost, Riemannian Geometry and Geometric Analysis, Springer, Berlin,
2nd edition, 1998,
M. Struwe, Variational Methods, Springer, Berlin, 2nd edition, 1996,
and the references contained therein.

The basic references that have been used in writing the present chapter
are the monograph of M.Struwe just quoted, as well as
P. Rabinowitz, Minimax Methods in Critical Point Theory with Applications

to Differential Equations, CBMS Reg. Conf. Ser. 65, AMS, Providence,
1986
and
E. Zeidler, Nonlinear Functional Analysis and its Applications, I I I , Springer,

Berlin, 1984.
These three monographs contain not only detailed bibliographical ref

erences which the reader is urged to consult i n order to find the
original sources of the results of the present chapter but also many
further results and examples concerning the Palais-Smale condition and
index theories.
Exercises
9.1 W h y is Theorem 9.2.1 called 'mountain pass theorem'? Hint:
Try to find an analogy between the statement of that result and
the geometry of mountain passes.
9.2 Try to find conditions for a function
/ : ft x R->R
so that the reasoning of Theorem 9.2.2 can be extended to the

Dirichlet problem
Au(x) = f(x,u(x)) for x G ft

u(x) = 0 for x dft
in a smooth bounded domain ft. ( A n answer can be found in

Theorem 6.2 of the quoted monograph of M.Struwe.)
9.3 Develop an index theory for a general compact group G i n place
of Z .
2
9.4 Extend Theorem 9.1.3 to the relative case as indicated at the

end of Section 9.1.
Index
x y = J2i=i = x v
|cc| = X X, X V2 meas, 118

f f(x)dx of / , 120
u(t) = u ( t ) , X V A
l
k
C (Q), xv C {A), 120
d
fc d 0 } , 130
b
/(u) = / F(t,u(t),u(t))ift,3 0 H/IL :=sup^ i^|l,133 0
D/, 4
d
V * := { / V R linear with :
r7Ci([a,6],IR ),5
| | / I U < o o } , 133
F , 5 u
(V*)* = : V * * , 133
i^P, 5 3Jn * X , 135
6 J ( U , I J ) := / ( u 4 - s r / ) , _ 1 0 a = 0 M := x
AC{[a,b\), 11 {a: # : (x, y) = 0 for all y A / } ,

1
D ( / , R * ) , 11 <
141
13 | | T | | := s u p ^ l | ^ f l e R + U {oo}, 144
x 0
2
6 / ( U , T J ) := ^ / ( u + t y ) , ^ 1 9 = 0 L ( V , W), 145
k e r T : = {a; 6 V : To; = 0 } , 145
Fpipj'Hi'nj = 1 2 ^ = 1 FpipiViVj, 19
V = V i e V , 146 2
c(t) : = g ( t ) , 32
coker, 147
^(c) :=/ |c(t)|dt = 0
T
H ( T ) , 147
T 2
i n d T , 147
/ o ( E i i ( 0 ) * * , 32 F ( V , W ) , 147
E(c) :=yT\c(t)\ dt 2
= DF(u), 150
1
C , 150
2
C , 150
2
> F ( u ) , 150
:= 3 9
ODE, 155
fl'y.fc ~$t9v> I M | o : = s u p | | y ( t ) | | , 156
c t /
r : =
Jfc ^(gjitk+gku-gjid), 39 H/llp = H / H L P ) : = :
U \f(x)\ dx)p, M A
p
159
n
esssup f(x) := inf { A G R | f(x) <
x j 4
5 :=
X for almost all xeA}, 162
39 C (n), 1660
d(p,q) := i n f { L ( c ) | c : [a,6] supp^J, 166

M rectifiable curve with c(a) =
p,c(b) = q}, 51
319
320 Index
ft' cc ft, 167 Borel cr-algebra, 117

fh =3 / , 167 brachystochrone, 4
canonical equation, 85, 89, 95, 97,

a := ( a , . . . , a ) ,
x d 171
99-101, 111
ll 171 canonical equations, 80, 93
t; := D w , 171 a canonical system, 80
fc
W >*>(ft), 171 canonical transformation, 95-100, 103
Cantor diagonalization, 135
N l l V * . p ( n ) (E|a|<fc Jn \D u\ ) a
p P
, catenary, 283
171 catenoid, 283
k
H >P(Q), 171 Cauchy sequence, 126
f e
H ' ( n ) , 171
0
p
characteristic function, 119, 211, 243
Diu, 172 Christoffel symbols, 39
Dw, 172 classical calculus of variations, 3
Isc, 185 closed geodesic, 67
A
F ( x ) := i n f y X
2
( A F ( y ) + d ( * , t / ) ) , 190 coarea formula, 250, 257
x
J (x), 191 coercive, 186
coercivity condition, 291
(A:=Eti(afV),199 Coffman, 310
s c ~ F , 208 cokernel, 147
i , 210
A compactness condition, 183
9 " / , 216 compactness of critical sequences, 292
r-limn-oo F, n 225 complementary subspace, 146
V(ft), 242 complete, 126, 134
\\Du\\, 242 complete integral, 84, 93
Dw n 2 4 2
N I B V ( Q ) '= \M\mn) + H l l ( )> completely integrable, 100
l^l -i,243 d
conjugate, 22, 24
P(E,Q) :=\\D E\\(n), 244 conjugate point, 43
X
* u(x), 246 conservation law, 26

G / ( d , R ) , 257 conserved quantities, 26
0 ( d , R ) , 257 constant of motion, 80, 99
J\{u)v, 270 continuous linear functional, 133
(,) 2,283 L
continuous linear operator, 144
( P S ) , 292 control condition, 109
K , 292
a
control equation, 106, 108, 109, 111, 207
V F ( w ) , 294 control parameter, 104
s p e c , 306 a
control problem, 109
gen(A), 310 control restriction, 105
accessory variational problem, 19 control variable, 111, 207
accumulation point, 185, 208 converge, 125
Ambrosetti, 302 convex, 68, 127, 130, 143, 186, 191, 193,
angular momentum, 26, 28, 30 214, 219, 222
arc-length, 3 convex combination, 142
Arzela-Ascoli theorem, 176 convex curve, 68
convex function, 122
Banach fixed point theorem, 150, 152 convex functional, 188
Banach space, 126, 129, 132-134, 138, convexity, 291
145, 161, 162, 270, 291, 292, coordinate transformation, 36
299-301 cost, 105
Banach spaces, 150 cost function, 207
Bellman equation, 105, 108 countable base, 184
Bellman function, 105, 107 countably additive, 118
Bellman's method, 106 critical family, 75
bifurcation theory, 268, 270 critical point, 5, 62, 66, 293, 294, 298,
Borel measure, 118 301, 303, 306, 307, 312, 317
Borel set, 117 critical sequence, 291, 292, 304, 314
Index 321
critical value, 302, 303 fundamental lemma of the calculus of

cusp catastrophe, 279 variations, 5
de Giorgi, 225 T-convergence, 225, 227, 229, 231

deformation, 293, 294, 297, 298, 302, generating function, 100
307-309 genus, 310, 311, 313
dense, 169 genus of Krasnoselskij, 310
diffeomorphism, 34, 95 geodesic, 39, 43, 45, 50, 51, 55, 57, 58,
differentiable, 150 60, 88, 102
differentiable map, 150 geodesic distance, 82, 90, 93
differentiation under the integral, 124 geodesic parallel coordinates, 45, 49
Dirac delta distribution, 173 geometric optics, 86
Dirac distribution, 166 gradient, 294, 299
direct method, 183 gradient flow, 294
Dirichlet boundary condition, 3, 26, great circle, 42
183, 190
Dirichlet principle, 199 Holder continuous, 179
Dirichlet's integral, 199, 203 Holder's inequality, 160, 163
distance, 51 Hahn-Banach theorem, 129, 134, 137,
distance function from a smooth 143, 166
hypersurface, 262 Hamilton-Jacobi equation, 83-86, 89,
distributional derivative, 173 92, 93, 101
dual space, 133, 163 Hamilton-Jacobi theory, 111
Hamiltonian, 80, 89
Hamiltonian flow, 95, 98
eiconal, 82 harmonic, 199, 201
eiconal equation, 83, 86, 90 harmonic oscillator, 87
elementary catastrophes, 279 Hessian, 4
ellipticity assumption, 198 Hilbert space, 126, 128, 141, 162, 293,
energy, 26, 30, 32, 34 297
e-minimizer, 229 Hilbert's invariant integral, 92
equivalence classes of functions, 159 homogenization, 232
essential supremum, 162
Euler-Lagrange equation, 6, 8-10, 16, implicit function theorem, 151, 152
17, 19, 21-23, 29, 38, 60, 79, 80, 83, index, 147, 308, 311, 313, 318
88, 89, 111, 197, 267, 282, 303 indicator function, 210
example of Bolza, 206 inner radius, 70
extension, 130 insulating layer, 235
integrable, 120
Federer, 261 integral, 155
feedback control, 109 integral of motion, 27
Fermat's principle, 4 integral of the Hamiltonian flow, 99
field of geodesies, 46 invariant integral, 93
field of solutions, 90, 93 inverse function theorem, 154
finite perimeter, 244 inverse operator theorem, 145
first axiom of countability, 137, 184, involution, 308
185, 209, 225, 227, 228 isometry, 34
first conjugate point, 23
first integral of motion, 30 Jacobi, 22
flow, 298 Jacobi equation, 20, 24, 268
foliated by tori, 100 Jacobi field, 20-22, 24, 269
Frechet differentiable, 150 Jacobi identity, 103
Fredholm alternative, 149 Jacobi operator, 268, 284
Fredholm operator, 147-149, 270, 281, Jacobi's method, 99
287 Jensen's inequality, 122
free boundary condition, 26 Jordan curve, 35
Friedrichs mollifier, 166 Jordan curve Theorem, 68
322 Index
Kakutani, 139 Moreau-Yosida transform, 212

Kepler problem, 102 Morrey, 222
Kolmogorov-Arnold-Moser theory, 100 mountain pass theorem, 302, 303, 306,
Kondrachev, 175 318
Lagrange multiplier, 9 neighbourhood system, 184

Laplace operator, 199, 200 Newtonian motion, 81
Lebesgue integral, 117, 120 Noether, 26
Lebesgue measure, 117, 118 nonminimizing critical point, 291, 302
Legendre condition, 20, 112 norm, 125
Legendre transformation, 79, 88 norm convergence, 125, 132
length, 32, 34 norm of a linear functional, 133
length minimizing curve, 8 normed space, 125
light ray, 4 null class, 159
limit cases of the Palais-Smale null function, 159
condition, 317
linear functional, 132, 133, 241 optimal control theory, 111, 207
linear functionals, 129 ordinary differential equation, 155
linear operator, 144 ordinary differential equations in
Lipschitz continuous, 155, 203 Banach spaces, 155
local chart, 25, 32 orthogonal, 90
local minimum, 22 orthogonal complement, 141
lower semicontinuity, 184
lower semicontinuous, 185, 186, 188,
Palais, 299
193, 208, 230
Palais-Smale condition, 77, 292, 293,
lower semicontinuous w.r.t. weak
304, 306, 312, 317
convergence, 187
parallel surfaces, 92
lower semicontinuous envelope, 208
parallelogram identity, 128
Lyapunov-Schmid, 280
parameterization invariant, 34
Lyapunov-Schmid reduction, 269
parameterized by arc-length, 8, 35, 36,
Lyusternik, 306
Lyusternik-Schnirelman, 67 43, 88
parameterized proportionally to
arc-length, 35, 38, 55, 89
mean curvature, 263
perimeter, 244
mean value property, 201
phase space, 98, 100
measurable, 118-120
phase transition, 254
measure, 117
Picard-Lindelof theorem, 155
metric tensor, 33, 47
Poincare* inequality, 177, 304
minimal hypersurface, 255
Poisson bracket, 102
minimal hypersurfaces, 203
polar coordinate, 49
minimal surface of revolution, 282
polar coordinates, 48
minimax value, 303
Pontryagin function, 110, 111
minimizer, 4-6, 12, 183, 186, 229, 291,
Pontryagin maximum principle, 110-112
302
principal curvature, 263
minimizer of a convex variational
projection theorem, 142
problem, 189
proper, 62
minimizing, 3
pseudo-gradient, 299, 300
minimizing sequence, 183
Minkowski functional, 143
Minkowski's inequality, 160 quasiconvex, 219, 222
Modica, 254 quasilinear partial differential equation,
Mobius strip, 75, 76 198
mollification, 167, 174, 175, 200, 245
momenta, 80 Rabinowitz, 302, 306
momentum, 26, 28, 30 Radon measure, 118, 241
monotonically increasing sequence, 122 range, 147
Moreau-Yosida approximation, 190 rectifiable, 35
Index 323
reflexive, 134, 135, 137-139, 163, 174, theorem of Clarkson, 164
186 theorem of de Giorgi and Nash, 198
regularity, 11 theorem of E . Noether, 28
regularity theory, 198, 286, 306, 316 theorem of Fatou, 123
regularizing term, 255 theorem of Fubini, 122
relative minimum, 62, 66 theorem of Helly, 132
relatively compact, 167 theorem of Jacobi, 84, 93, 101
relaxation, 208 theorem of Kondrachev, 180
relaxed function, 208, 214 theorem of Lebesgue, 123
relaxed functional, 209 theorem of Liouville, 98
Rellich, 175 theorem of Lyusternik-Schnirelman, 67
Rellich-Kondrachev theorem, 305 theorem of Mazur, 142
reparameterization, 8 theorem of Milman, 139
Riccati equation, 86, 108 theorem of Modica-Mortola, 248
Riemannian manifold, 43, 52, 53 theorem of Morrey, 179
Riemannian normal coordinates, 48 theorem of Picard-Lindelof, 39, 155
Riemannian polar coordinate, 49, 51, 60 theorem of Pohozaev, 316
Riesz representation theorem, 241 theorem of Rellich, 175
rotational invariance, 200 theorem of Riesz, 141
theorem of Riesz-Fischer, 161
Sard's theorem, 250, 257
theorem of Sobolev, 179
scalar product, 126
theorem on dominated convergence, 123
Schnirelman, 306 theory of catastrophes, 279
Schwarz inequality, 35, 127 Thorn, 279
second axiom of countability, 184 topological space, 185
second variation, 18, 23 translation invariance, 118
semigroup family, 299 transversality condition, 110
semigroup property, 157, 294, 297, 300 triangle inequality, 125, 126, 159
separable, 135, 169, 173, 184, 186
shortest geodesic, 52, 53, 55 uniform convergence, 126, 168
shortest length, 50 uniformly continuous, 168
cr-algebra, 117 uniformly convex, 127, 129, 139, 157,
signed measure, 242 164
simple function, 119 unstable critical point, 291
smoothing kernel, 166 variational problem, 9
Sobolev Embedding Theorem, 175, 179, volume preserving, 98
303, 305
Sobolev inequalities, 179 weak convergence, 135-137, 142, 174,
Sobolev space, 171, 173 186, 214
special point, 306-309, 312 weak* convergence, 135
special value, 306, 307 weak* convergent, 135
sphere, 39 weak derivative, 171, 172
star shaped, 316 weak limit, 138
state variable, 207 weak solution, 306, 315
step function, 119 weak solution of the Jacobi equation,
strictly normed, 157 285
strong convergence, 125, 174 weak topology, 291
submanifold, 24, 32, 43, 52, 53 weak* topology, 137
summation convention, xv, 19 weakly convergent, 135, 136
support, 166 weakly lower semicontinuous, 222
surface of revolution, 60, 282 weakly proper, 185
symmetry assumption, 308-310 Weierstrafi, 46
symplectic geometry, 96 Weierstrass approximation theorem, 170
symplectomorphism, 97 Weierstrafi condition, 112
Weyl's lemma, 199
Taylor expansion, 274
Young's inequality, 160
test functions, 166
theorem of B . Levi, 122 Zorn's lemma, 131

Calculus of Variations PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Calculus of Variations PDF

Uploaded by

Copyright:

Available Formats

CAMBRIDGE STUDIES IN

Jiirgen Jost and Xianqing Li-Jost

CAMBRIDGE UNIVERSITY PRESS

Cambridge University Press 1998

T h i s book is in copyright. Subject to statutory exception

First published 1998

Typeset in Computer Modern by the authors using I A l ^ X 2e

A catalogue record of this book is available from the British Library

Library of Congress Cataloguing in Publication data

Jost, Jurgen, 1956-

I S B N 0 521 64203 5 hardback

Transferred to digital printing 2003

Preface and summary page x

P a r t one: One-dimensional variational problems 1

2 A geometric example: geodesic curves 32

3 Saddle point constructions 62

P a r t two: Multiple integrals in the calculus of

1 Lebesgue measure and integration theory 117

4 T h e direct methods in the calculus of variations 183

5 Nonconvex functionals. Relaxation 205

6.2 Homogenization 231

7 BV-functionals and T-convergence: the example of

Appendix A The coarea formula 257

8 Bifurcation theory 266

9 T h e PalaisSmale condition and unstable critical

The calculus of variations is concerned with the construction of optimal

also some examples are presented in detail that occurred in recent ap

to develop some general theory first. I n Chapter 1 of that part, Lebesgue

The present book is self-contained, with very few exceptions. Prere

For a function u(t), we write

In Part I , the independent variable is usually called t, because in many

for the corresponding spaces.

to indicate that the expression on the left of this symbol is defined by

One-dimensional variational problems

The classical theory

Also, one needs to specify a class of admissible functions among which

(1) We want to minimize the arc-length of the graph of a function u :

(2) Historically, the calculus of variations started with the so-called

(3) A generalization of (1) and (2) is

where 7 : [a, 6] x R R is a given positive function. This vari

If one seeks a minimum of a smooth function

one knows that at a minimizing point Zo fi, one necessarily has

where Df is the derivative of / . The first variation of / actually has

In order to investigate this more closely, we assume that F is of class

I(u + sri) > I(u).

I(u + sri)= I F(t,u(t) + sri(t),u(t) + sf}(t))dt.

I(u + sr,)Uo (1.1.1)

of u, and F the one w.r.t. the components of u(t).

f This means that r) is continuously differentiable as a function on [a, b) with values

L h(t)(p(t)dt = 0 for all <p G CQ ((a, 6), R ) , D

then h = 0 on (a, 6).

<^(t)=0 fort^to, t{l,...,d}.

Written out, the Euler-Lagrange equations are

i. e. a system of d ordinary differential equations of second order that are

Let us compute the Euler-Lagrange equations for our preceding three

meaning that u has to be a straight line, a fact that we know of

(2) We just need to insert 7 = yj2gu(t) into (1.1.6) to obtain

Actually, (2) is an example of an integrand F(t,u,ii) that does not

and hence every solution of the Euler-Lagrange equation (1.1.4) satisfies

F(t,u{t),u{t)) -u(t)F (t,u(t),u(t))

Conversely, every solution of (1.1.7), w i t h the exception of ii = 0, i.e.

The Euler-Lagrange equations in this case are

We also know that a reparameterization of a curve g(t) does not change