You are on page 1of 53

Input-to-State Stability for PDEs Iasson

Karafyllis
Visit to download the full and correct content document:
https://textbookfull.com/product/input-to-state-stability-for-pdes-iasson-karafyllis/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Finite-time stability: an input-output approach First


Edition Amato

https://textbookfull.com/product/finite-time-stability-an-input-
output-approach-first-edition-amato/

Optimal Transport for Applied Mathematicians Calculus


of Variations PDEs and Modeling Filippo Santambrogio

https://textbookfull.com/product/optimal-transport-for-applied-
mathematicians-calculus-of-variations-pdes-and-modeling-filippo-
santambrogio/

Optimal Transport for Applied Mathematicians Calculus


of Variations PDEs and Modeling 1st Edition Filippo
Santambrogio (Auth.)

https://textbookfull.com/product/optimal-transport-for-applied-
mathematicians-calculus-of-variations-pdes-and-modeling-1st-
edition-filippo-santambrogio-auth/

Analysis Modeling and Stability of Fractional Order


Differential Systems 1 The Infinite State Approach 1st
Edition Jean-Claude Trigeassou

https://textbookfull.com/product/analysis-modeling-and-stability-
of-fractional-order-differential-systems-1-the-infinite-state-
approach-1st-edition-jean-claude-trigeassou/
Analysis Modeling and Stability of Fractional Order
Differential Systems 2 The Infinite State Approach 1st
Edition Jean-Claude Trigeassou

https://textbookfull.com/product/analysis-modeling-and-stability-
of-fractional-order-differential-systems-2-the-infinite-state-
approach-1st-edition-jean-claude-trigeassou/

A Compact Course on Linear PDEs 1st Edition Alberto


Valli

https://textbookfull.com/product/a-compact-course-on-linear-
pdes-1st-edition-alberto-valli/

Finite Difference Computing with PDEs A Modern Software


Approach 1st Edition Hans Petter Langtangen

https://textbookfull.com/product/finite-difference-computing-
with-pdes-a-modern-software-approach-1st-edition-hans-petter-
langtangen/

Virtual Equivalent System Approach for Stability


Analysis of Model based Control Systems Weicun Zhang

https://textbookfull.com/product/virtual-equivalent-system-
approach-for-stability-analysis-of-model-based-control-systems-
weicun-zhang/

Arms Sales and Regional Stability 1st Edition The


International Institute For Strategic Studies (Iiss)

https://textbookfull.com/product/arms-sales-and-regional-
stability-1st-edition-the-international-institute-for-strategic-
studies-iiss/
Communications and Control Engineering

Iasson Karafyllis · Miroslav Krstic

Input-to-State
Stability for
PDEs
Communications and Control Engineering

Series editors
Alberto Isidori, Roma, Italy
Jan H. van Schuppen, Amsterdam, The Netherlands
Eduardo D. Sontag, Boston, USA
Miroslav Krstic, La Jolla, USA
Communications and Control Engineering is a high-level academic monograph
series publishing research in control and systems theory, control engineering and
communications. It has worldwide distribution to engineers, researchers, educators
(several of the titles in this series find use as advanced textbooks although that is not
their primary purpose), and libraries.
The series reflects the major technological and mathematical advances that have
a great impact in the fields of communication and control. The range of areas to
which control and systems theory is applied is broadening rapidly with particular
growth being noticeable in the fields of finance and biologically-inspired control.
Books in this series generally pull together many related research threads in more
mature areas of the subject than the highly-specialised volumes of Lecture Notes in
Control and Information Sciences. This series’s mathematical and control-theoretic
emphasis is complemented by Advances in Industrial Control which provides a
much more applied, engineering-oriented outlook.
Publishing Ethics: Researchers should conduct their research from research
proposal to publication in line with best practices and codes of conduct of relevant
professional bodies and/or national and international regulatory bodies. For more
details on individual ethics matters please see:
https://www.springer.com/gp/authors-editors/journal-author/journal-author-
helpdesk/publishing-ethics/14214

More information about this series at http://www.springer.com/series/61


Iasson Karafyllis Miroslav Krstic

Input-to-State Stability
for PDEs

123
Iasson Karafyllis Miroslav Krstic
Department of Mathematics Mechanical and Aerospace Engineering
National Technical University of Athens University of California
Athens San Diego, CA
Greece USA

ISSN 0178-5354 ISSN 2197-7119 (electronic)


Communications and Control Engineering
ISBN 978-3-319-91010-9 ISBN 978-3-319-91011-6 (eBook)
https://doi.org/10.1007/978-3-319-91011-6
Library of Congress Control Number: 2018940658

© Springer International Publishing AG, part of Springer Nature 2019


This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

Partial differential equations (PDEs) are the most elegant way to model the
dynamics of not only much of the physical world around us but, increasingly, many
of the social systems we participate in. Consequently, PDEs are the key tool in
many sciences.
Control theory, as one such science, has historically been developed for ordinary
differential equation (ODE) models. This is natural as control theory grew out of
electric circuit theory and advanced due to the needs associated with aerospace
vehicles and other applications where ODE models are suitable. However, the
breadth of phenomena in physical, life, and social sciences demands that control
theoretic questions be answered in the PDE setting.
One of the notions that has had the greatest impact in the study of control
systems is Input-to-State Stability (ISS). Developed by E. D. Sontag in the late
1980s, for systems described by ODEs, ISS has enabled the solution of numerous
robust nonlinear control problems that were previously inconceivable, not to
mention the robustness studies of nonlinear systems with disturbances and non-
linear feedback systems. No other foundational block in the analysis of forced and
interconnected nonlinear systems is of comparable significance.
Extensions of ISS have been carried out from ODEs to discrete-time systems,
time-delay systems, stochastic nonlinear systems, and hybrid systems. The exten-
sion of the ISS to PDEs has remained elusive the longest because of the unbounded
nature of the input operators in PDEs with boundary inputs. This book deals with
the extension of ISS to PDEs, with the expectation that ISS will enable similar
advances in the control theory of PDEs as it has for finite-dimensional systems.
On the design end of the methodological spectrum, backstepping has played a
comparably enabling role for the rigorous synthesis of stabilizing feedback laws,
first for nonlinear ODEs in the 1990s, and then since about 2000 for PDEs. In
backstepping designs for PDEs, feedback laws are expressed by non-local terms,
and consequently, the closed-loop system is described by PDEs with non-local
terms. More specifically, the non-local terms in the boundary feedback laws give
rise to boundary conditions with non-local terms. Consequently, for any extension
of ISS for PDEs to be useful for the study of robustness of the closed-loop system,

v
vi Preface

they should allow the presence of boundary disturbances and non-local intercon-
nections at the boundaries. Similar structures arise in some physical systems,
regardless of whether control is employed.
This book is a first effort in the development of a theory of ISS in 1-D PDEs that
may include non-local terms and disturbances both in the domain and the boundary
conditions. Moreover, the book is a first effort in the development of a theory for
ISS in PDEs which is reasonably accessible to broader audiences of engineers and
scientists who have not undergone years of mathematical training in abstract
analysis techniques for PDEs. This combination of applicability and accessibility
should equip researchers from a broad range of fields to solve new problems with
ISS techniques for PDEs, just as had previously been the case with ISS for ODEs.
While emphasizing linear PDEs and focusing on the challenges arising from the
unbounded linear operators in linear PDEs, nonlinear terms are allowed in many
cases, particularly globally Lipschitz nonlinearities. Furthermore, the book provides
simple formulas for the disturbance gains that are actually computable. This is
important even if the formula provides a conservative bound: Trade-offs between
various control objectives and qualitative relations are revealed in this way, giving
insights for the discovery of novel approaches.
The book provides ISS estimates for the solutions of PDEs with external inputs
in various spatial Lp norms (with 1  p\ þ1) of the state as well as for the spatial
sup-norm of the state. This is important because—unlike the finite-dimensional
case—these state norms are not equivalent. An emphasis is placed on the spatial
sup-norm, because it can give spatially pointwise estimates of the solution. All
other spatial Lp norms with 1  p\ þ1 cannot give pointwise estimates. For
parabolic PDEs, the book also provides ISS estimates in the spatial H 1 norm in
certain cases. Stability studies in different spatial norms of the state for PDEs are
crucial, because—again, unlike the finite-dimensional case—it is well-known that a
PDE may be stable in one state norm and unstable in some other state norm. To this
purpose, comparison between the stability requirements in different spatial state
norms is performed in many cases: This is a feature that is rarely met in other books
with stability studies for PDEs.

What Does the Book Cover? The book covers all quantitative methodologies that
can be used for showing Input-to-State Stability in systems described by 1-D PDEs
with respect to boundary as well as in-domain disturbances. We emphasize the
adjective “quantitative:” The book leaves out methodologies that would be merely
of a qualitative, existence nature. All methodologies presented in the book lead to
formulas for the gains of the disturbances. Moreover, the PDEs are allowed to
contain non-local, nonlinear terms both in the differential equations and in the
boundary conditions.
The book is divided into three parts. Part I of the book is devoted to hyperbolic
first-order 1-D PDEs. Part II of the book covers the case of parabolic 1-D PDEs.
Both Parts I and II start with a chapter that deals with basic questions about
existence/uniqueness of solutions (Chap. 2 in Part I and Chap. 4 in Part II). The
Preface vii

remaining chapters provide methodologies for proving ISS. A large variety of


methodologies is developed: ISS Lyapunov functionals, transformation to delay
equations, eigenfunction expansions as well as methodologies related to numerical
schemes (ISS Lyapunov Functional Under Discretization). For each result obtained,
the book provides examples which explain in detail all important notions and clarify
many issues.
One of the major advantages of ISS over other stability notions is the fact that
the notion of ISS allowed the development of small-gain analysis. Small-gain
results exploit the structure of a given system, decompose the system into a number
of feedback interconnected subsystems and allow the derivation of stability esti-
mates for the overall system by studying each separate subsystem. Part III of the
book is devoted to the application of the ISS results in Parts I and II to systems
containing at least one PDE. Many kinds of interconnections are studied:
(1) one hyperbolic first-order PDE with ODEs (Chap. 8),
(2) one parabolic PDE with ODEs (Chap. 8),
(3) two hyperbolic first-order PDEs (Chap. 9),
(4) two parabolic PDEs (Chap. 10), and
(5) one hyperbolic PDE with one parabolic PDE (Chap. 11).
All kinds of interconnections are studied: interconnections by means of
boundary or in-domain terms and interconnections by means of local or non-local
terms. Again, existence/uniqueness results are provided in all cases and explicit
small-gain conditions that guarantee ISS or global exponential stability of the
overall system are given. An important advantage of small-gain analysis is the fact
that the small-gain conditions can be used in order to prove stability in the spatial
sup-norm as well as in every other spatial Lp norm (with 1\ p\ þ1). Many
examples and applications that motivate the study of each interconnected system
are also provided.

Who is the Book For? Many engineers and mathematicians, students and
researchers, work on important analysis and control-theoretic aspects of PDEs. The
book should be of interest to all of them.
Systems described by PDEs are abundant in many basic sciences: physics,
biology, chemistry, earth sciences, economics. Stabilization problems and feedback
interconnections arise naturally in many cases, and the results contained in this
book have off-the-shelf applicability to such problems or can be suitably extended
to more complex problems. The applicability in multiple domains, immediate or
with some additional effort, is the reason why the book offers many examples from
various sciences, even though the examples’ primary role is pedagogical—to
illustrate the theorems.
Mathematicians working on stability of PDEs, an area in which ISS and
small-gain ideas are still absent, will find the book particularly interesting and
possibly even startlingly useful.
viii Preface

We assume that the reader is familiar with stability theory for finite-dimensional
systems at a moderately advanced level. The mathematical background needed for
the complete comprehension of the results is kept to a minimum: There is no need
to know anything more than the basics in measure theory and functional analysis.

Acknowledgements. We thank Christophe Prieur and Federico Bribiesca


Argomedo with whom we have shared the initial interest in the common subject of
ISS for PDEs, as well as Andrii Mironchenko, Birgit Jakob, Fabian Wirth and Felix
Schwenninger for the shared interest in the subject at its advanced stage. The
existence of this community of colleagues has stimulated our focus on this topic.
We also thank Jean-Michel Coron for his encouragement following the inaugural
presentation of our results at his sixtieth birthday symposium in Paris in 2016.
Finally, we thank Eduardo Sontag whose seminal launch of the ISS topic and its
subsequent comprehensive development have inspired us and fueled our interest for
decades. Our conversations on ISS for PDEs began during a stroll in Belgrade in
April 2015.
Iasson Karafyllis would like to thank Markos Papageorgiou for the very fruitful
scientific collaboration during the last five years. The author would also like to
thank John Tsinias and Stelios Kotsios for their constant help to many important
things. And special thanks go to three girls in Iasson’s life—Xqirsίma, Oktlpίa,
and Kaseqίma—for many things (too many to be mentioned here).

Athens, Greece Iasson Karafyllis


La Jolla, CA, USA Miroslav Krstic
February 2018
Contents

1 Preview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Topics and Tools in the Book . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Background Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 The ISS Property for Systems Containing PDEs . . . . . . . . . . . . 5
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Part I ISS for First-Order Hyperbolic PDEs


2 Existence/Uniqueness Results for Hyperbolic PDEs . . . . . . . . . . . . 19
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Basic Existence/Uniqueness Results . . . . . . . . . . . . . . . . . . . . . 20
2.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3 ISS in Spatial Lp Norms for Hyperbolic PDEs . . . . . . . . . . . . . . . . 39
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 ISS-Lyapunov Functionals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.3 ISS by Means of IDEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Part II ISS for Parabolic PDEs


4 Existence/Uniqueness Results for Parabolic PDEs . . . ........... 61
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... 61
4.2 Basic Existence/Uniqueness Results . . . . . . . . . . ........... 62
4.3 Nonlinear, Non-local Terms and Interconnections
with ODEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... 71

ix
x Contents

4.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.5 Boundary Inputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
5 ISS in Spatial L2 and H1 Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.2 L2 Norm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.3 H1 Norm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.4 ISS Lyapunov Functionals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.5 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5.5.1 The Temperature of a Solid Bar . . . . . . . . . . . . . . . . . . 133
5.5.2 Gain of R–A–D PDEs With Respect to Inlet
Disturbances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
6 ISS in Spatial Lp Norms for Parabolic PDEs . . . . . . . . . . . . . . . . . 141
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.2 ISS-Lyapunov Functionals Under Discretization . . . . . . . . . . . . 142
6.2.1 The Notion of the ISS-LFUD . . . . . . . . . . . . . . . . . . . . 144
6.2.2 Proof of Theorem 6.4 . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.3 Deriving ISS Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
6.3.1 ISS in the Sup-norm . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.3.2 ISS in L1 Norm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.3.3 ISS in Lp Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
6.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
6.4.1 Robustness of Backstepping with Respect to Actuator
Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
6.4.2 ISS in Taylor–Couette Flow . . . . . . . . . . . . . . . . . . . . . 177
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

Part III Small-Gain Analysis


7 Fading Memory Input-to-State Stability . . . . . . . . . . . . . . . . . . . . . 185
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.2 Two Basic Lemmas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
8 PDE-ODE Loops . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 193
8.1 Introduction . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 193
8.2 Hyperbolic PDE-ODE Loops . . . . . . . .. . . . . . . . . . . . . . . . . . 194
8.3 Parabolic PDE-ODE Loops . . . . . . . . .. . . . . . . . . . . . . . . . . . 201
8.4 Applications . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 209
8.4.1 A Chemical Reactor with a Cooling Jacket . . . . . . . . . . 209
8.4.2 A Water Tank . . . . . . . . .............. . . . . . . . . . 213
Contents xi

9 Hyperbolic PDE-PDE Loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215


9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
9.2 Existence/Uniqueness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
9.3 Small-Gain Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
9.4 Comparison with Bastin–Coron Stability Conditions for
Balance Laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
10 Parabolic PDE-PDE Loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
10.2 Existence/Uniqueness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
10.3 Small-Gain Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
11 Parabolic–Hyperbolic PDE Loops . . . . . . . . . . . . . . . . . . . . . . . . . . 263
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
11.2 Movement of Chemicals Underground . . . . . . . . . . . . . . . . . . . 264
11.3 Combination of Boundary and in-Domain Feedback
Interconnection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
Abbreviations

1-D One-Dimensional
BCs Boundary Conditions
ES Exponentially Stable
IDEs Integral Delay Equations
IOS Input-to-Output Stability
ISS Input-to-State Stability
ISS-LF ISS Lyapunov Function(al)
ISS-LFUD ISS Lyapunov Functional Under Discretization
ODEs Ordinary Differential Equations
PDEs Partial Differential Equations
R–A Reaction–Advection
R–A–D Reaction–Advection–Diffusion
R–D Reaction–Diffusion
SL Sturm–Liouville

xiii
Notation and Definitions

Throughout the book, we adopt the following notation and terminology.


• < þ :¼ ½0; þ 1Þ. For x 2 <, ½x denotes the integer part of x 2 <. Z þ is the set
of non-negative integers.
• Let A  <n be an open set. By C 0 ðA ; XÞ, we denote the class of continuous
functions on A, which take values in X  <m . By Ck ðA ; XÞ, where k  1 is an
integer, we denote the class of functions on A  <n with continuous derivatives
of order k, which take values in X  <m .
• Let I  < be an interval. A continuous function f : ½0; 1 ! < is called piece-
wise C1 on ½0; 1 and we write f 2 PC 1 ð½0; 1Þ, if the following properties hold:
(i) for every x 2 ½0; 1Þ the limit limh!0 þ ðh1 ðf ðx þ hÞ  f ðxÞÞÞ exists and is
finite, (ii) for every x 2 ð0; 1 the limit limh!0 ðh1 ðf ðx þ hÞ  f ðxÞÞÞ exists and
is finite, (iii) there exists a set J  ð0; 1Þ of finite cardinality, where f 0 ðxÞ ¼
limh!0 ðh1 ðf ðx þ hÞ  f ðxÞÞÞ ¼ limh!0 þ ðh1 ðf ðx þ hÞ  f ðxÞÞÞ holds for
x 2 ð0; 1ÞnJ, and (iv) the mapping ð0; 1ÞnJ 3 x ! f 0 ðxÞ 2 < is continuous, i.e.,
f 0 ðxÞ is continuous for every x 2 ð0; 1ÞnJ. Notice that here a piecewise C 1
function is assumed to be continuous on ½0; 1 (this is not assumed in many
textbooks).
• Let r 2 C 0 ð½0; 1; ð0; þ 1ÞÞ be given. L2r ð0; 1Þ denotes the equivalence class of
R 1=2
1
measurable functions f : ½0; 1 ! < for which k f kr ¼ 0 rðxÞjf ðxÞj2 dx
R1
\; þ1. L2r ð0; 1Þ is a Hilbert space with inner product hf ; gi ¼ 0 rðxÞ
f ðxÞgðxÞdx. When rðxÞ  1, we use the notation L2 ð0; 1Þ for the standard space
R 1=2
1
of square-integrable functions and k f k2 ¼ 0 jf ðxÞj2 dx \ þ1 for
f 2 L2 ð0; 1Þ.

xv
xvi Notation and Definitions

• L1 ð0; 1Þ denotes the equivalence class of measurable functions f : ½0; 1 ! <


for which k f k1 ¼ sup0  z  1 ðjf ðzÞjÞ\ þ1 (a Banach space), where sup
  For g 2 C ð½0; 1; ð0; þ1ÞÞ we also use the
0
denotes the essential supremum.
jf ðzÞj
norm k f k1;g ¼ sup0  z  1 gðzÞ . Let p 2 ½1; þ1Þ be given. Lp ð0; 1Þ denotes
the equivalence class of measurable functions f : ½0; 1 ! < for which
R 1=p
1
k f kp ¼ 0 jf ðxÞjp dx \ þ1.
• Let I  < þ :¼ ½0; þ1Þ be an interval. By L1 ðI; UÞ, we denote the space of
measurable and essentially bounded functions uð Þ defined on I and taking
values in U  <m . By L1 loc ðI; UÞ, we denote the space of measurable and locally
essentially bounded functions uð Þ (i.e., essentially bounded functions on every
compact subset of I) defined on I and taking values in U  <m .
• Let u : < þ ½0; 1 ! < be given. We use the notation u½t to denote the profile
at certain t  0, i.e., ðu½tÞðzÞ ¼ uðt; zÞ for all z 2 ½0; 1. When uðt; zÞ is differ-
entiable with respect to z 2 ½0; 1, we use the notation u0 ðt; zÞ for the derivative
of u with respect to z 2 ½0; 1, i.e., u0 ðt; zÞ ¼ @@ uz ðt; zÞ. For an interval I  < þ , the
space C 0 ðI; L2r ð0; 1ÞÞ is the space of continuous mappings
I 3 t ! u½t 2 L2r ð0; 1Þ.
• For an integer k  1, H k ð0; 1Þ denotes the Sobolev space of functions in L2 ð0; 1Þ
with all its weak derivatives up to order k  1 in L2 ð0; 1Þ.
• K is the class of strictly increasing, continuous functions a : < þ ! < þ with
að0Þ ¼ 0. The class of functions K1 is the class of strictly increasing, contin-
uous functions a : < þ ! < þ with að0Þ ¼ 0 and lims! þ 1 aðsÞ ¼ þ1. KL
denotes the class of functions r : < þ < þ ! < þ with the following prop-
erties: (a) for every s  0, the mapping < þ 3 t ! rðs; tÞ is non-increasing with
limt! þ1 rðs; tÞ ¼ 0, and (b) for every t  0, the mapping < þ 3 s ! rðs; tÞ is
of class K.
Chapter 1
Preview

1.1 Introduction

In this chapter, we give a preview of the material contained in the book. We deal
with a topic that only recently has attracted attention: the Input-to-State Stability
(ISS) property of systems containing at least one PDE.
While the ISS theory for finite-dimensional systems is a mature topic, this is not
so for infinite-dimensional systems. Infinite-dimensional systems present a number
of technical difficulties, which sometimes are not easily resolved. This is particu-
larly true for systems involving at least one PDE. We illustrate the key challenge
next.
Time-invariant linear systems, whether finite-dimensional or containing PDEs,
usually admit a representation of the form

x_ ðtÞ ¼ AxðtÞ þ BuðtÞ


ð1:1:1Þ
xðtÞ 2 X; uðtÞ 2 U

where xðtÞ is the state, uðtÞ is the input, X is a Banach space (the state space), U is a
normed linear space (the input space), and A; B are linear operators. Under
appropriate assumptions, the solution of (1.1.1) can be expressed by (the extension
of) the “variation of constants” formula,

Zt
xðtÞ ¼ expðAtÞxð0Þ þ expðAðt  sÞÞBuðsÞds ð1:1:2Þ
0

where expðAtÞ denotes the semigroup generated by A and the integral in (1.1.2) is a
Bochner’s integral [technical issues are omitted in this motivating discussion but
the reader may consult [55] for a justification of (1.1.2)]. When the semigroup
generated by A is exponentially stable, i.e., when there exist constants M; r [ 0

© Springer International Publishing AG, part of Springer Nature 2019 1


I. Karafyllis and M. Krstic, Input-to-State Stability for PDEs,
Communications and Control Engineering,
https://doi.org/10.1007/978-3-319-91011-6_1
2 1 Preview

such that kexpðAtÞxð0ÞkX  M expðrtÞkxð0ÞkX for all t  0 and xð0Þ 2 X, then


we can obtain from (1.1.2) the following estimate under the assumption that B is a
bounded operator:

M  
kxðtÞkX  M expðrtÞkxð0ÞkX þ kBk sup kuðsÞkU ; for t  0: ð1:1:3Þ
r 0st

Inequality (1.1.3) is exactly the ISS property for the linear system (1.1.1), and
this type of reasoning is exactly what allows to conclude ISS in the linear
finite-dimensional case. But this is also exactly the point that shows what can go
wrong in the infinite-dimensional case: Estimate (1.1.3) is valid only if B is a
bounded operator. When B is not a bounded operator, which is the case for systems
described by PDEs with boundary inputs, then we can no longer perform the above
steps and we cannot conclude ISS.
While the above discussion illustrated the major challenge to deriving ISS for
PDEs, many more problems arise at the technical level. Section 1.5 of the present
chapter gives the reader a glimpse of the technical difficulties involved: Even the
notion of ISS itself has to be modified in order to be able to talk about the ISS
property for PDEs.
Section 1.2 of the chapter overviews the topics covered in the book. The
structure of the book and a brief presentation of the contents of all subsequent
chapters are also given in Sect. 1.2.
This book demonstrates the use of ISS in a large number of applications that are
at the core of many scientific disciplines. All applications contained in the present
book are listed in Sect. 1.3, where it becomes apparent to the reader that ISS is a
stability notion that can be used productively by various scientists and engineers.
Finally, Sect. 1.4 of the present chapter states the required background for a
reader of the book. It is emphasized that, unlike other books on PDEs, this book
does not require a mathematical background on measure theory, semigroup theory,
and functional analysis.

1.2 Topics and Tools in the Book

The present book supplies various new tools in the theory of PDEs:
• existence/uniqueness results that can be applied to PDEs with non-local terms,
• existence/uniqueness results for feedback interconnections of a PDE with a
system of ODEs or another PDE,
• quantitative methodologies that can be used for the derivation of ISS estimates
for PDEs with boundary and distributed disturbances,
• a methodology, based on small-gain arguments, for the stability analysis of
systems containing at least one PDE.
1.2 Topics and Tools in the Book 3

It should be emphasized that the book contains only quantitative methodologies


for the derivation of ISS estimates, i.e., methodologies that allow the computation
of the gain functions and the convergence rate. This is in contrast with qualitative
methodologies, which guarantee that an ISS estimate for a system holds but do not
allow the computation of the gain functions and the convergence rate. The present
book does not cover qualitative methodologies for ISS in systems containing PDEs.
The applicability of the results in the present book to PDEs with non-local terms
is important. Control theory has reached a point where standard methodologies
exist for the design of feedback stabilizers for a large number of PDEs. However, in
(almost) all such cases the feedback law (either boundary feedback or in-domain
feedback) is expressed as a non-local term. Therefore, the closed-loop system
becomes a PDE with non-local (boundary or in-domain) terms. Consequently, the
present book is expected to be useful to control theorists and engineers that study
control problems with PDEs.
The book is divided into three parts. Part I is devoted to the study of first-order
hyperbolic PDEs. Chapter 2 contains all existence/uniqueness results that can be
applied to first-order hyperbolic PDEs with non-local terms. Chapter 3 presents two
different methodologies that allow the derivation of ISS estimates for first-order
hyperbolic PDEs: the ISS Lyapunov functional methodology and the methodology
that uses an equivalent representation of the PDE as a system of Integral Delay
Equations (IDEs). The use of both methodologies is explained by means of various
examples.
Part II of the book is dealing with parabolic PDEs. Chapter 4 provides the
existence/uniqueness results that allow the study of parabolic PDEs with non-local
terms. Chapter 5 focuses on the derivation of ISS estimates in the spatial L2 and H 1
norms. For such spatial norms, two different methodologies are presented: the
eigenfunction expansion and ISS Lyapunov functional methodology. This is in
sharp contrast with the sup-norm, where the aforementioned methodologies cannot
be applied. Chapter 6 introduces a novel methodology that can be applied to this
case: the ISS Lyapunov Functional Under Discretization. Various ISS estimates for
parabolic PDEs are derived in Chap. 6 in the sup-norm as well as in weighted Lp
norms (with 1  p\ þ 1).
Part III of the book is devoted to the small-gain analysis of composite systems that
contain at least one PDE. The first chapter of Part III of the book (Chap. 7) introduces
the reader to two basic lemmas that are used in all subsequent chapters: the lemmas
that allow the derivation of fading memory ISS inequalities. The remaining chapters
illustrate the use of the small-gain analysis for various types of loops. Chapter 8 is
dealing with the case where a parabolic or a hyperbolic PDE is feedback intercon-
nected to a system of ODEs. Chapter 9 studies the case where two first-order
hyperbolic PDEs are feedback interconnected. The feedback interconnection of two
parabolic PDEs is studied in Chap. 10. Finally, Chapter 11 provides results for a
topic that has not been studied in the literature: the feedback interconnection of a
parabolic PDE to a first-order hyperbolic PDE. Chapters 9, 10, and 11 contain
additional existence/uniqueness results for systems containing two PDEs.
4 1 Preview

Certain additional technical issues should also be emphasized at this point:


• Almost, all nonlinearities considered in the results of the present book are
globally Lipschitz. This is restrictive, because in applications a globally Lipschitz
nonlinearity is a rare case. However, the obtained results allow the study of a
large number of applications (see next section). It is expected that future research
will allow the consideration of more general types for the nonlinear terms.
• All PDEs studied in the present book have only one spatial dimension. Again,
this is restrictive because in many applications, models with more than one
spatial dimension are encountered. However, the reader can realize that some of
the methodologies of this book can be generalized without much effort to the
case of an arbitrary number of spatial dimensions.
• All PDEs studied in the present book have a linear differential part (i.e., they are
semilinear). Nonlinearities and non-local terms are met only in the reaction
(in-domain) terms or in the boundary terms. Again, this is a restrictive feature of
the book, because many interesting PDEs (e.g., Burgers’ PDE) do not have a
linear differential part. It is expected that future research will allow the study of
more general types of PDEs.
• The feedback interconnections terms considered in the present book can appear
both as in-domain terms as well as boundary terms. However, we do not con-
sider feedback interconnection terms that affect the differential part of a PDE:
The interconnection terms appear only as reaction terms or as boundary terms.

1.3 Applications

The ISS property has been used so far mainly by control theorists. However, it is
clear that the sensitivity of a system with respect to external disturbances is a
fundamental property of a system. ISS is the best way to express quantitatively the
sensitivity with respect to external disturbances and consequently, ISS can be used
by a wide variety of scientists and engineers.
This book demonstrates the use of ISS to a large number of applications that are
in the core of many scientific disciplines. More specifically, the book contains
applications involving:
• Heat transfer phenomena: the water tank (Sects. 4.4 and 8.4), the temperature of
a solid bar (Sect. 5.4), and the counterflow heat exchanger (Sects. 9.2 and 9.3).
• Heat transfer phenomena with chemical reactions: the chemical reactor with a
cooling jacket (Sects. 2.3 and 8.4) and the chemical reactor with a cooling jacket
and non-negligible conduction (Sects. 4.4 and 8.4).
• Fluid mechanics: the Taylor–Couette flow (Sect. 6.4).
• Biological phenomena: the age distribution of a microbial population (Sect. 2.3).
• Hydrogeology: the flow of a liquid in a fissured porous medium (Sect. 10.3).
• Mass transfer phenomena: the movement of chemicals underground (Sect. 11.2).
1.3 Applications 5

In addition to the above applications, the sensitivity of reaction–advection–


diffusion equations with respect to inlet disturbances is studied in Sect. 5.5.
Systems of reaction–diffusion equations are also studied in Chap. 10.

1.4 Background Material

The present book assumes that the reader is familiar with the notion of ISS for
finite-dimensional systems described by ODEs. Moreover, the book also assumes
that the reader is familiar with basic stability theory: Terms like (global) exponential
stability, functions of class K; K1 ; KL are used throughout the book. The book
deals mainly with global exponential stability notions.
However, unlike other books on PDEs, this book does not require a mathe-
matical background on measure theory, semigroup theory, and functional analysis.
Moreover, the book does not require knowledge of technical results about weak,
mild, and strong solutions of PDEs. This is made on purpose: Many scientists and
engineers do not understand easily the notions involved in such areas of mathe-
matics. However, a basic familiarity with the theory of metric spaces is assumed as
well as a good knowledge of real analysis. Banach’s fixed point theorem is used
many times (almost in every existence proof of the book), as well as notions of
uniform continuity, uniform and absolute convergence of series of functions.
Moreover, only a basic knowledge of Lp spaces ð1  p  þ 1Þ and H k spaces is
required.
It is clear that the comprehension of an application requires a slight knowledge
of the corresponding scientific laws that determine the evolution of the phe-
nomenon. For example, the mass transfer application in Chap. 11 involves the use
of Fick’s law, and all heat transfer applications involve Fourier’s law and/or
Newton’s law. However, the reader who is not familiar with the specific scientific
law may skip the derivation of the mathematical model of the application and focus
on the analysis of the derived model.
A final comment about the present book is in order. The book contains a huge
number of examples. The examples of the book have been chosen carefully and
illustrate all aspects of the theory. It is suggested to the reader to study the examples
in detail because in this way the reader will understand deeply all available results.

1.5 The ISS Property for Systems Containing PDEs

The present paragraph aims to the conceptual clarification of the ISS property for
infinite-dimensional systems that contain at least one PDE in one spatial variable.
The ISS property for PDEs requires certain important conceptual modifications of
6 1 Preview

the standard ISS property that was used so far in the literature for finite-dimensional
systems or for delay systems.
E. D. Sontag formulated the ISS property for time-invariant finite-dimensional
systems of the form

x_ ðtÞ ¼ f ðxðtÞ; dðtÞÞ


ð1:5:1Þ
xðtÞ 2 <n ; dðtÞ 2 <m

where f 2 C0 ð<n  <m ; <n Þ is a locally Lipschitz vector field with f ð0; 0Þ ¼ 0.
The ISS property requires the existence of functions r 2 KL; c 2 K and norms k  k
of <n ; jj of <m , so that for every x0 2 <n ; d 2 L1 loc ð< þ ; < Þ, the following
m

inequality holds for the solution xðtÞ 2 < of the initial-value problem (1.5.1) with
n

xð0Þ ¼ x0 :

kxðtÞk  rðkx0 k; tÞ þ sup ðcðjdðsÞjÞÞ; for all t  0: ð1:5:2Þ


0st

In the finite-dimensional setting, the selection of the norms kk of <n ; jj of <m is
indifferent (because of the equivalence of norms in finite-dimensional spaces).
The ISS property was later extended to (infinite-dimensional) time-delay systems
of the form

d
ðxðtÞ  qðxt ÞÞ ¼ f ðxt ; dðtÞÞ
dt ð1:5:3Þ
xðtÞ 2 <n ; dðtÞ 2 <m

where xt 2 C0 ð½r; 0; <n Þ is defined by ðxt ÞðsÞ ¼ xðt þ sÞ for s 2 ½r; 0; r [ 0 is a
constant, f : C0 ð½r; 0; <n Þ  <m ! <n ; q : C 0 ð½r; 0; <n Þ ! <n are continuous
mappings which are Lipschitz on bounded sets with f ð0; 0Þ ¼ qð0Þ ¼ 0. Again, the
ISS property required the existence of functions r 2 KL; c 2 K and a norm jj of
<m , so that for every x0 2 C0 ð½r; 0; <n Þ; d 2 L1 loc ð< þ ; < Þ, the following
m

inequality holds for the solution xðtÞ 2 < of the initial-value problem (1.5.3) with
n

xðsÞ ¼ x0 ðsÞ for s 2 ½r; 0

kxt k  rðkx0 k; tÞ þ sup ðcðjdðsÞjÞÞ; for all t  0 ð1:5:4Þ


0st

where kxt k ¼ maxr  s  0 ðjxðt þ sÞjÞ denotes the sup-norm of C 0 ð½r; 0; <n Þ.
Again, the selection of the norm jj of <m is indifferent (because of the equivalence
of norms in finite-dimensional spaces).
It would seem appropriate for a PDE in one spatial dimension, whose state u½t is
a real function on a bounded domain X  <, and which is driven by a (possibly
distributed) disturbance d : < þ ! D, where D is a normed linear space with norm
kkD (the input space), to formulate the ISS property in the following way: to
require the existence of a normed linear space S of real functions on X with norm
1.5 The ISS Property for Systems Containing PDEs 7

kkS , a set MD of locally bounded mappings d : < þ ! D, functions


r 2 KL; c 2 K, so that for every u0 2 S; d 2 MD , the following inequality holds for
the solution u½t 2 S of the PDE with u½0 ¼ u0 :
    
ku½tkS  r ku0 kS ; t þ sup c kd½skD ; for all t  0: ð1:5:5Þ
0st

The first important feature of such a property for PDEs is that, unlike for ODEs,
the selection of the spatial norms kkS ; kkD is not inconsequential: Norms on
functional spaces are not (in general) equivalent. Therefore, it would seem justified
to call such a property “the ISS property in spatial norm kkS ”.
However, many important technical reasons do not allow the formulation of the
ISS property in the way described above.
(A) Boundary Disturbances
When one deals with PDEs accompanied by boundary conditions, some compli-
cations might arise. When the disturbance appears at the boundary condition
(boundary disturbance) then, depending on the notion of the solution that we have
chosen, certain compatibility conditions must hold. The following example illus-
trates this point.
Example Consider the heat equation with a boundary disturbance:

@u @2u
ðt; zÞ ¼ 2 ðt; zÞ; for ðt; zÞ 2 ð0; þ 1Þ  ð0; 1Þ ð1:5:6Þ
@t @z
@u
uðt; 0Þ  dðtÞ ¼ ðt; 1Þ ¼ 0; for t [ 0 ð1:5:7Þ
@z

u½0 ¼ u0 : ð1:5:8Þ
A classical solution for the above initial-boundary value problem is a function
u 2 C 0 ð< þ  ½0; 1Þ \ C1 ðð0; þ 1Þ  ½0; 1Þ with u½t 2 C 2 ðð0; 1ÞÞ for all t [ 0
that satisfies (1.5.6), (1.5.7) and (1.5.8).
Consequently, if classical solutions are considered for the evolution Eq. (1.5.6)
with (1.5.7), then we must require the compatibility condition
u0 ð0Þ ¼ limt!0 þ ðdðtÞÞ. □
The reader might be disturbed by the previous example and consider it as an
artificial example. After all, it is well known that weaker notions of the solution of a
PDE might require less demanding regularity properties for the solution, which may
allow us to avoid the restrictive compatibility conditions. Indeed, this is the case for
the above example. If one considers solutions u 2 C 0 ð< þ ; L2 ð0; 1ÞÞ \
C 1 ðð0; þ 1Þ  ð0; 1ÞÞ, then the compatibility condition u0 ð0Þ ¼ limt!0 þ ðdðtÞÞ is
not required. Moreover, in such a case, we are able to consider a wider class of
initial conditions ðu0 2 L2 ð0; 1ÞÞ instead of the restricted class of initial conditions
required for a classical solution (notice that a classical solution for the
8 1 Preview

initial-boundary value problem (1.5.6), (1.5.7), (1.5.8) requires that the initial
condition u0 be a continuous function on ½0; 1Þ.
However, what happens if one can prove an estimate of the form (1.5.5) only for
classical solutions? Indeed, as we will see in the subsequent chapters, an ISS
estimate in the sup-norm for the evolution Eq. (1.5.6) with (1.5.7) can be proved
only for classical solutions.
It is clear that for PDEs the set of allowable disturbance inputs cannot be (in
general) a fixed set MD of locally bounded mappings d : < þ ! D: The allowable
disturbances are restricted by the initial state u0 . This is a technical complication
that the reader might mistakenly consider “harmless”: Unfortunately, this is not
merely a technical complication. If the set of allowable disturbances depends on the
initial state u0 , then there are serious system-theoretic complications: The con-
catenation property does not hold for arbitrary disturbance inputs in this case. The
concatenation property plays a crucial role in the proof of the existence of ISS
Lyapunov functionals (see [24, 48]).
(B) Determination of State Space and Allowable Disturbances
Another feature that arises very often in the study of PDEs is the difficulty in
determining explicitly the state space and the set of allowable disturbances. The
following examples illustrate this issue.
Example Consider the following disturbance-free, infinite-dimensional system that
is described by a PDE-ODE loop:

1þaþh
x_ ðtÞ ¼  xðtÞ þ huðt; 1Þ ð1:5:9Þ
1þa

@u @2u að1 þ a þ hÞ ah
ðt; zÞ ¼ p 2 ðt; zÞ þ zxðtÞ  zuðt; 1Þ ð1:5:10Þ
@t @z ð1 þ aÞ 2 1þa

@u
uðt; 0Þ ¼ ðt; 1Þ þ auðt; 1Þ ¼ 0 ð1:5:11Þ
@z

where p; a; h [ 0 are constants. System (1.5.9), (1.5.10), (1.5.11) is a special case


of a class of systems studied in Example 4.4.2 (parameterized by a smooth function
k : ½0; 1 ! < with kð0Þ ¼ k 0 ð1Þ þ akð1Þ þ a ¼ 0; here we have kðzÞ ¼  1 az þ a) in
Sect. 4.4 of the book,  where it is shown that for every u0 2 D; x0 2 < with
h 2 D, where D :¼ v 2 H 2 ð0; 1Þ : vð0Þ ¼ v0 ð1Þ þ avð1Þ ¼ 0 ; hðzÞ ¼ pu000 ðzÞ þ
 
1þaþh
az
1þa 1þa x 0  hu 0 ð1Þ for z 2 ½0; 1, there exist unique mappings
x 2 C 1 ð< þ ; <Þ, u 2 C 0 ð< þ  ½0; 1Þ \ C1 ðð0; þ 1Þ  ½0; 1Þ, with u½t 2
C 2 ð½0; 1Þ for all t [ 0; u½0 ¼ u0 ; xð0Þ ¼ x0 satisfying (1.5.10) for all ðt; zÞ 2
ð0; þ 1Þ  ð0; 1Þ and (1.5.9), (1.5.11) for all t  0.
What is the state space for system (1.5.9), (1.5.10), (1.5.11)? In other words,
what is the set S  <  H 2 ð0; 1Þ, for which the implication
1.5 The ISS Property for Systems Containing PDEs 9

“ðx0 ; u0 Þ 2 S ) ðxðtÞ; u½tÞ 2 S for all t  0”

holds? If we define S  < H 2 ð0; 1Þ to be the 


set of u0 2 D; x0 2 < with h 2 D,
00 1þaþh
where hðzÞ ¼ pu0 ðzÞ þ 1 þ a 1 þ a x0  hu0 ð1Þ for z 2 ½0; 1, then we cannot
az

guarantee that ðxðtÞ; u½tÞ 2 S, for all t  0 (notice that a function in S must be at
least three times continuously differentiable with respect to the spatial variable,
while the solution is only guaranteed to be in C2 ð½0; 1Þ). Consequently, we cannot
guarantee that the proposed set S  <  H 2 ð0; 1Þ is the state space for system
(1.5.9), (1.5.10), (1.5.11). □
Example Consider again system (1.5.6), (1.5.7). Theorem 4.11 in Sect. 4.5 shows
that for every u0 2 H 2 ð0; 1Þ; d 2 C 2 ð< þ Þ with u0 ð0Þ  dð0Þ ¼ u00 ð1Þ ¼ 0, there
exists a unique classical solution of the initial-boundary value problem (1.5.6),
(1.5.7), and (1.5.8). It follows that theset of allowable disturbances Uðu0 Þ contains
the set d 2 C 2 ð< þ Þ : dð0Þ ¼ u0 ð0Þ . However, this is a conservative estimation
of the set of allowable disturbances Uðu0 Þ for classical solutions. To see this,
consider the following initial-boundary value problem

@w @2w
ðt; zÞ ¼ 2 ðt; zÞ; for all ðt; zÞ 2 ð0; þ 1Þ  ð0; 1Þ ð1:5:12Þ
@t @z

Z1
@w
wðt; 0Þ  hðsÞwðt; sÞds ¼ ðt; 1Þ ¼ 0; for t [ 0 ð1:5:13Þ
@z
0

w½0 ¼ w0 ð1:5:14Þ

where h 2 C4 ð½0; 1Þ is a given function with hð0Þ ¼ h0 ð1Þ ¼ h00 ð0Þ ¼ 0. System
(1.5.12), (1.5.13) is studied in Sect. 4.3, where it is shown that for every w0 2
R1
H 2 ð0; 1Þ with w0 ð0Þ  0 hðsÞw0 ðsÞds ¼ w00 ð1Þ ¼ 0, there exists a unique mapping
w 2 C0 ð< þ  ½0; 1Þ \ C 1 ðð0; þ 1Þ  ½0; 1Þ satisfying w½t 2 C2 ð½0; 1Þ for all
t [ 0 and (1.5.12), (1.5.13), (1.5.14). It follows that for every w0 2 H 2 ð0; 1Þ with
R1 R
w0 ð0Þ  0 hðsÞw0 ðsÞds ¼ w00 ð1Þ ¼ 0, the mapping dðtÞ ~ :¼ 1 hðsÞwðt; sÞds
0
defined for all t  0 is an allowable disturbance for the initial-boundary value
problem (1.5.6), (1.5.7), and (1.5.8).Notice that d~ does not necessarily belong to
the set d 2 C2 ð< þ Þ : dð0Þ ¼ u0 ð0Þ . □

(C) A Conceptual Modification of the ISS Property


Typically, an existence/uniqueness result for an infinite-dimensional
initial-boundary value problem that contains a PDE on a bounded domain X  <
will be a statement of the form:
10 1 Preview

There exists a non-empty set A of functions on X and a normed linear space D with the
following property: For every u0 2 A, there exists a non-empty set Mðu0 Þ of locally
bounded mappings d : < þ ! D, such that a unique solution u½t of the infinite-dimensional
problem with u½0 ¼ u0 exists for all t  0.

Suppose that the notion of the solution for the infinite-dimensional problem has
been selected in such a way that the semigroup property holds, i.e., for every t0  0
and every solution u½t of the infinite-dimensional problem that corresponds to the
locally bounded mapping d : < þ ! D, the mapping w½t ¼ u½t0 þ t for t  0, is the
unique solution of the infinite-dimensional problem with w½0 ¼ u½t0  correspond-
ing to the shifted input ðdt0 dÞðtÞ ¼ dðt0 þ tÞ for t  0. In every such case, the state
space S and the set of allowable disturbances Uðu0 Þ may be defined implicitly by
the following statements:
The state space S is the set of all functions on X for which there exists at least one locally
bounded mapping d : < þ ! D such that a unique solution u½t of the infinite-dimensional
problem with u½0 ¼ u0 exists for all t  0.
For every u0 2 S, the set of allowable disturbances Uðu0 Þ is the set of all locally bounded
mappings d : < þ ! D for which a unique solution u½t of the infinite-dimensional problem
with u½0 ¼ u0 exists for all t  0.

In other words, the state space as well as the set of allowable disturbances is
determined implicitly by the notion of the solution. Notice that in the above setting,
it holds that A S and that the following implication holds:

“u0 2 S; d 2 Uðu0 Þ ) u½t 2 S and ðdt dÞ 2 Uðu½tÞ; for all t  0”

where dt d is the shifted input ðdt dÞðsÞ ¼ dðt0 þ sÞ for s  0.


Consequently, a convenient formulation for the ISS property for PDEs requires
the existence of a certain normed linear space X of real functions on X with norm
kkX , a normed linear space D with norm kkD , functions r 2 KL; c 2 K, so that for
every solution u½t of the infinite-dimensional problem that corresponds to the
locally bounded mapping d : < þ ! D, the following inequality holds with
u0 ¼ u½0:
    
ku½tkX  r ku0 kX ; t þ sup c kd½skD ; for all t  0: ð1:5:15Þ
0st

Notice that the modified ISS property essentially requires inequality (1.5.15) to
hold for every u0 2 SðS X defined as above) and d 2 Uðu0 ÞðUðu0 Þ defined as
above). However, there is no reference to the state space S X and to the set of
allowable disturbances Uðu0 Þ in the requirements of the ISS property: The weight
has been transferred from the state space S X (which may not be known explicitly)
to the notion of the solution. The notion of the solution (weak, mild, strong,
classical) of the infinite-dimensional problem must always be selected in such a way
that the semigroup property holds. This is exactly the notion of ISS that is used
throughout the book (although we manage to give the actual state space in most
cases).
1.5 The ISS Property for Systems Containing PDEs 11

(D) The Lack of ISS Lyapunov Functionals


Apart from the modification of the ISS property itself, additional complications
arise in the study of PDEs. One of the most convenient tools for establishing the ISS
property for finite-dimensional systems of the form (1.5.1) as well as for delay
systems of the form (1.5.3) is the ISS Lyapunov functional. In general, for a control
system with state space S X; where X is a normed linear space, and input space the
normed linear space D, an ISS Lyapunov functional is a functional V : S ! < þ
with the following properties:
(i) There
 exist functions
 a1 ; a2 2 K1 such that the inequality
a1 kukX  VðuÞ  a2 kukX holds for all u 2 S,
(ii) the mapping < þ 3 t ! Vðu½tÞ is absolutely continuous for every solution
u½t,
(iii) there exists a functional V_ : S  D ! < such that the equation dtd Vðu½tÞ ¼
V_ ðu½t; d½tÞ holds for almost all t  0 and for every solution u½t, and
(iv) there exist a seminorm G of D and a positive-definite   function j 2
C ð< þ ; < þ Þ for which the inequality GðdÞ  j kd kD holds for all d 2 D, a
0

positive-definite function q 2 C 0 ð< þ ; < þ Þ and a function c 2 K such that


the following implication holds: “if ðu; dÞ 2 S  D satisfy VðuÞ  cðGðdÞÞ
then V_ ðu; d Þ   qðVðuÞÞ”.
More specifically, by virtue of Lemma 2.14 on page 82 in [24], a functional
V : S ! < þ with the above properties guarantees the following ISS estimate in the
norm kkX of X:

      
ku½tkX  max r ku0 kX ; t ; sup a1
1 c j kd½skD ; for all t  0
0st

ð1:5:16Þ

for certain r 2 KL, where u½t 2 S is the solution with initial condition u½0 ¼ u0
that corresponds to the locally bounded mapping d : < þ ! D.
The subsequent chapters will provide ISS Lyapunov functionals in some cases.
Moreover, Lyapunov functionals for the corresponding disturbance-free PDE can
provide ISS Lyapunov functionals when the effect of the disturbance is expressed
by means of a bounded operator: This is the case for distributed additive distur-
bances, i.e., disturbances that appear in the PDE in an additive way.
Boundary disturbances present a major challenge, because the transformation of
the boundary disturbance to a distributed disturbance leads to the ISS property with
respect to the boundary disturbance and some of its time derivatives (see, e.g., [4]
and see [1] for the analysis of this stability property in the finite-dimensional case).
This is explained by the use of unbounded operators for the expression of the effect
of the boundary disturbance (see the relevant discussion in [44] for inputs in
infinite-dimensional systems that are expressed by means of unbounded linear
operators). Although the construction of Lyapunov functionals for PDEs has
12 1 Preview

progressed significantly during the last years (see, e.g., [5, 42, 43, 45, 48, 58]), none
of the proposed Lyapunov functionals can be used for the derivation of the ISS
property with respect to boundary disturbances for parabolic PDEs with Dirichlet
boundary conditions. It should be noticed that there is no guarantee that an ISS
Lyapunov functional exists for PDEs with boundary disturbances, since all con-
verse ISS Lyapunov results require the so-called concatenation property for the
allowed disturbances (see [24, 48]). Unfortunately, as noted above the “concate-
nation property” does not hold for boundary disturbances. Therefore, novel math-
ematical results have to be devised in order to be able to prove the ISS property for
PDEs.
One of the subjects of the present book is the development of various mathe-
matical tools that guarantee the ISS property for systems containing PDEs with
respect to distributed and boundary disturbances.
Notes and Comments for Chapter 1
(1) E. D. Sontag formulated the ISS property for finite-dimensional systems
described by ODEs in [61]. As stated by Sontag in [67], ISS was formulated as
an external stability property that takes into account the internal stability
properties of the system (Lyapunov stability + input-output stability). Since
1989, various qualitative properties of this stability notion were analyzed in
[62, 63, 67] and various variants of the ISS property were formulated (see [1,
64–67]). The impact of the ISS in control theory was huge. It seems impossible
in our days to give a course on nonlinear control theory without including ISS.
Many researchers have used ISS in order to prove a number of results that
changed nonlinear control theory and nonlinear stability theory for ever (see for
instance the small-gain theorem in [21] and the use of ISS for nonlinear
adaptive control in [32]). The research related to ISS gave birth to a number of
important qualitative system-theoretic properties (e.g., Bounded-Input-
Bounded-State, Converging-Input-Converging-State, asymptotic gain), which
are also used frequently in recent control studies (e.g., see the use of the
asymptotic gain property in [30]).
(2) The study of ISS and IOS for delay systems of the form (1.5.3) was performed
in [23, 24, 26, 56, 57]. This is still a topic of ongoing research (see [51]). ISS
concepts have been also applied to hybrid, switched, and impulsive systems:
see [6, 10, 17, 68].
(3) The first paper that studied the ISS property for infinite-dimensional systems
that could be described by PDEs was the paper [20]. Later, the papers [7–9]
started a systematic qualitative study of the ISS property for PDEs. The effort
for a complete qualitative characterization of ISS for PDEs continues (see [19,
41, 44, 46, 47, 49]). A different direction of study of ISS for PDEs, related to
the construction of ISS Lyapunov functionals, was initiated in [42, 43] and
continued in [5, 45, 48, 58]. Use of semigroup theory for the qualitative study
of ISS of linear abstract infinite-dimensional systems was made in [18, 19].
1.5 The ISS Property for Systems Containing PDEs 13

(4) Most of the existing results in the literature deal with distributed disturbances.
The first papers that studied ISS for PDEs with boundary disturbances were [25,
28]. Additional results were given in [27, 29, 31]. The recent paper [69] gave an
ISS Lyapunov functional for parabolic PDEs with boundary disturbances.
(5) The existence of ISS Lyapunov functionals is closely related to converse
Lyapunov results. For finite-dimensional systems, the results of the seminal
paper [39] were used in [62] for the construction of ISS Lyapunov functions.
For infinite-dimensional systems, converse Lyapunov results have been given
in [22, 24–26, 48, 50].
(6) The interest for the stability analysis of PDEs with non-local terms is strong,
both from the PDE literature as well as from the numerical analysis literature
(see [11, 12, 14–16, 40, 53, 54]). However, recent advances in feedback control
of PDEs have forced the control literature to deal with PDEs containing
non-local terms. This happened because the feedback law itself is a functional
of the solution of the PDE and appears as a non-local term either in the
boundary condition or in the PDE itself (see [33, 59, 60] and references
therein).
(7) A completely different methodology for the derivation of ISS estimates for
parabolic PDEs was proposed in [52]: The methodology is based on monotone
systems theory and can be applied to parabolic PDEs in domains of arbitrary
dimension. However, the methodology in [52] is qualitative and does not
provide accurate estimates for the gains of the disturbance inputs.
(8) Boundary input operators have been studied extensively in the context of
semigroup theory (see [2, 3, 13, 34–38]). The results in this research direction
can handle an arbitrary number of spatial dimensions (where interesting new
phenomena occur either due to geometry or different properties of functional
spaces) and deal with many important problems (regularity, input-output maps,
nonlinearities, etc.). However, the results in this research direction are not
dealing with the ISS property for PDEs.

References

1. Angeli D, Sontag ED, Wang Y (2003) Input-to-state stability with respect to inputs and their
derivatives. Int J Robust Nonlinear Control 13:1035–1056
2. Balakrishnan AV (1981) Applied functional analysis, 2nd edn. Springer, New York
3. Bensoussan A, Da Prato G, Delfour MC, Mitter SK (2007) Representation and control of
infinite dimensional systems, 2nd edn. Series: Mathematics, systems & control: foundations &
applications. Birkhäuser, Boston
4. Bribiesca Argomedo F, Witrant E, Prieur C (2012) D1-input-to-state stability of a
time-varying nonhomogeneous diffusive equation subject to boundary disturbances. In:
Proceedings of the American control conference, Montreal, QC, pp 2978–2983
5. Bribiesca Argomedo F, Prieur C, Witrant E, Bremond S (2013) A strict control Lyapunov
function for a diffusion equation with time-varying distributed coefficients. IEEE Trans
Autom Control 58:290–303
14 1 Preview

6. Cai C, Teel A (2009) Characterizations of input-to-state stability for hybrid systems. Syst
Control Lett 58(1):47–53
7. Dashkovskiy S, Mironchenko A (2010) On the uniform input-to-state stability of
reaction-diffusion systems. In: Proceedings of the 49th conference on decision and control,
Atlanta, GA, USA, pp 6547–6552
8. Dashkovskiy S, Mironchenko A (2011) Local ISS of reaction-diffusion systems. In:
Proceedings of the 18th IFAC world congress, Milano, Italy, pp 11018–11023
9. Dashkovskiy S, Mironchenko A (2013) Input-to-state stability of infinite-dimensional control
systems. Math Control Signals Syst 25:1–35
10. Dashkovskiy S, Mironchenko A (2013) Input-to-state stability of nonlinear impulsive
systems. SIAM J Control Optim 51:1962–1987
11. Day WA (1982) Extension of a property of the heat equation to linear thermoelasticity and
other theories. Q Appl Math 40:319–330
12. Day WA (1983) A decreasing property of solutions of parabolic equations with applications
to thermoelasticity. Q Appl Math 40:468–475
13. Desch W, Lasiecka I, Schappacher W (1985) Feedback boundary control problems for linear
semigroups. Isr J Math 51:177–207
14. Ekolin G (1991) Finite difference methods for a nonlocal boundary value problem for the heat
equation. BIT 31:245–261
15. Fairweather G, Lopez-Marcos JC (1996) Galerkin methods for a semilinear parabolic problem
with nonlocal boundary conditions. Adv Comput Math 6:243–262
16. Friedman A (1986) Monotone decay of solutions of parabolic equations with nonlocal
boundary conditions. Q Appl Math 44:401–407
17. Hespanha JP, Liberzon D, Teel AR (2008) Lyapunov conditions for input-to-state stability of
impulsive systems. Automatica 44(11):2735–2744
18. Jacob B, Nabiullin R, Partington JR, Schwenninger F. Infinite-dimensional input-to-state
stability and Orlicz spaces. arXiv:1609.09741 [math.OC]
19. Jacob B, Nabiullin R, Partington JR, Schwenninger F (2016) On input-to-state-stability and
integral input-to-state-stability for parabolic boundary control systems. In: Proceedings of
MTNS 2016
20. Jayawardhana B, Logemann H, Ryan EP (2008) Infinite-dimensional feedback systems: the
circle criterion and input-to-state stability. Commun Inf Syst 8:403–434
21. Jiang Z-P, Teel A, Praly L (1994) Small-gain theorem for ISS systems and applications. Math
Control Signals Syst 7:95–120
22. Karafyllis I, Pepe P, Jiang Z-P (2008) Global output stability for systems described by
retarded functional differential equations: Lyapunov characterizations. Eur J Control 14
(6):516–536
23. Karafyllis I, Pepe P, Jiang Z-P (2008) Input-to-output stability for systems described by
retarded functional differential equations. Eur J Control 14(6):539–555
24. Karafyllis I, Jiang Z-P (2011) Stability and stabilization of nonlinear systems. Series:
Communications and control engineering. Springer, London
25. Karafyllis I, Krstic M (2014) On the relation of delay equations to first-order hyperbolic
partial differential equations. ESAIM Control Optim Calc Var 20:894–923
26. Karafyllis I, Pepe P (2015) A note on converse Lyapunov results for neutral systems. In:
Karafyllis I, Malisoff M, Mazenc F, Pepe P (eds) Recent results on nonlinear time delayed
systems. Advances in delays and dynamics, vol 4. Springer, Berlin
27. Karafyllis I, Krstic M (2016) Input-to state stability with respect to boundary disturbances for
the 1-D heat equation. In: Proceedings of the 55th IEEE conference on decision and control,
pp 2247–2252
28. Karafyllis I, Krstic M (2016) ISS with respect to boundary disturbances for 1-D parabolic
PDEs. IEEE Trans Autom Control 61:3712–3724
29. Karafyllis I, Krstic M (2017) ISS in different norms for 1-D parabolic PDEs with boundary
disturbances. SIAM J Control Optim 55:1716–1751
Another random document with
no related content on Scribd:
caillouteux, le plus atrabilaire, le plus froid du monde; enfin marie
une jeune fille à un squelette, et tu comprendras l’existence dont tu
m’interdis de te développer les scènes curieuses: projets de fuite
évanouis à l’aspect de mon père, désespoirs calmés par le sommeil,
désirs comprimés, sombres mélancolies dissipées par la musique.
J’exhalais mon malheur en mélodies. Beethoven ou Mozart furent
souvent mes discrets confidents. Aujourd’hui je souris en me
souvenant de tous les préjugés qui troublaient ma conscience à
cette époque d’innocence et de vertu: si j’avais mis le pied chez un
restaurateur, je me serais cru ruiné; mon imagination me faisait
considérer un café comme un lieu de débauche, où les hommes se
perdaient d’honneur et engageaient leur fortune; quant à risquer de
l’argent au jeu, il aurait fallu en avoir. Oh! quand je devrais
t’endormir, je veux te raconter l’une des plus terribles joies de ma
vie, une de ces joies armées de griffes et qui s’enfoncent dans notre
cœur comme un fer chaud sur l’épaule d’un forçat. J’étais au bal
chez le duc de Navarreins, cousin de mon père. Mais pour que tu
puisses parfaitement comprendre ma position, apprends que j’avais
un habit râpé, des souliers mal faits, une cravate de cocher et des
gants déjà portés. Je me mis dans un coin afin de pouvoir tout à
mon aise prendre des glaces et contempler les jolies femmes. Mon
père m’aperçut. Par une raison que je n’ai jamais devinée, tant cet
acte de confiance m’abasourdit, il me donna sa bourse et ses clefs à
garder. A dix pas de moi quelques hommes jouaient. J’entendais
frétiller l’or. J’avais vingt ans, je souhaitais passer une journée
entière plongé dans les crimes de mon âge. C’était un libertinage
d’esprit dont nous ne trouverions l’analogue ni dans les caprices de
courtisane, ni dans les songes des jeunes filles. Depuis un an je me
rêvais bien mis, en voiture, ayant une belle femme à mes côtés,
tranchant du seigneur, dînant chez Véry, allant le soir au spectacle,
décidé à ne revenir que le lendemain chez mon père, mais armé
contre lui d’une aventure plus intriguée que ne l’est le Mariage de
Figaro, et dont il lui aurait été impossible de se dépêtrer. J’avais
estimé toute cette joie cinquante écus. N’étais-je pas encore sous le
charme naïf de l’école buissonnière? J’allai donc dans un boudoir
où, seul, les yeux cuisants, les doigts tremblants, je comptai l’argent
de mon père: cent écus! Évoquées par cette somme, les joies de
mon escapade apparurent devant moi, dansant comme les sorcières
de Macbeth autour de leur chaudière, mais alléchantes,
frémissantes, délicieuses! Je devins un coquin déterminé. Sans
écouter ni les tintements de mon oreille, ni les battements précipités
de mon cœur, je pris deux pièces de vingt francs que je vois encore!
Leurs millésimes étaient effacés et la figure de Bonaparte y
grimaçait. Après avoir mis la bourse dans ma poche, je revins vers
une table de jeu en tenant les deux pièces d’or dans la paume
humide de ma main, et je rôdai autour des joueurs comme un
émouchet au-dessus d’un poulailler. En proie à des angoisses
inexprimables, je jetai soudain un regard translucide autour de moi.
Certain de n’être aperçu par aucune personne de connaissance, je
pariai pour un petit homme gras et réjoui, sur la tête duquel
j’accumulai plus de prières et de vœux qu’il ne s’en fait en mer
pendant trois tempêtes. Puis, avec un instinct de scélératesse ou de
machiavélisme surprenant à mon âge, j’allai me planter près d’une
porte, regardant à travers les salons sans y rien voir. Mon âme et
mes yeux voltigeaient autour du fatal tapis vert. De cette soirée date
la première observation physiologique à laquelle j’ai dû cette espèce
de pénétration qui m’a permis de saisir quelques mystères de notre
double nature. Je tournais le dos à la table où se disputait mon futur
bonheur, bonheur d’autant plus profond peut-être qu’il était criminel;
entre les deux joueurs et moi, il se trouvait une haie d’hommes,
épaisse de quatre ou cinq rangées de causeurs; le bourdonnement
des voix empêchait de distinguer le son de l’or qui se mêlait au bruit
de l’orchestre; malgré tous ces obstacles, par un privilége accordé
aux passions qui leur donne le pouvoir d’anéantir l’espace et le
temps, j’entendais distinctement les paroles des deux joueurs, je
connaissais leurs points, je savais celui des deux qui retournait le roi
comme si j’eusse vu les cartes; enfin à dix pas du jeu, je pâlissais de
ses caprices. Mon père passa devant moi tout à coup, je compris
alors cette parole de l’écriture: L’esprit de Dieu passa devant sa
face! J’avais gagné. A travers le tourbillon d’hommes qui gravitait
autour des joueurs, j’accourus à la table en m’y glissant avec la
dextérité d’une anguille qui s’échappe par la maille rompue d’un filet.
De douloureuses, mes fibres devinrent joyeuses. J’étais comme un
condamné qui, marchant au supplice, a rencontré le roi. Par hasard,
un homme décoré réclama quarante francs qui manquaient. Je fus
soupçonné par des yeux inquiets, je pâlis et des gouttes de sueur
sillonnèrent mon front. Le crime d’avoir volé mon père me parut bien
vengé. Le bon gros petit homme dit alors d’une voix certainement
angélique: «Tous ces messieurs avaient mis,» et paya les quarante
francs. Je relevai mon front et jetai des regards triomphants sur les
joueurs. Après avoir réintégré dans la bourse de mon père l’or que
j’y avais pris, je laissai mon gain à ce digne et honnête monsieur qui
continua de gagner. Dès que je me vis possesseur de cent soixante
francs, je les enveloppai dans mon mouchoir de manière à ce qu’ils
ne pussent ni remuer ni sonner pendant notre retour au logis, et ne
jouai plus.—Que faisiez-vous au jeu? me dit mon père en entrant
dans le fiacre.—Je regardais, répondis-je en tremblant.—Mais, reprit
mon père, il n’y aurait eu rien d’extraordinaire à ce que vous eussiez
été forcé par amour-propre à mettre quelque argent sur le tapis. Aux
yeux des gens du monde, vous paraissez assez âgé pour avoir le
droit de commettre des sottises. Aussi vous excuserais-je, Raphaël,
si vous vous étiez servi de ma bourse... Je ne répondis rien. Quand
nous fûmes de retour, je rendis à mon père ses clefs et son argent.
En rentrant dans sa chambre, il vida la bourse sur sa cheminée,
compta l’or, se tourna vers moi d’un air assez gracieux, et me dit en
séparant chaque phrase par une pause plus ou moins longue et
significative:—Mon fils, vous avez bientôt vingt ans. Je suis content
de vous. Il vous faut une pension, ne fût-ce que pour vous apprendre
à économiser, à connaître les choses de la vie. Dès ce soir, je vous
donnerai cent francs par mois. Vous disposerez de votre argent
comme il vous plaira. Voici le premier trimestre de cette année,
ajouta-t-il en caressant une pile d’or, comme pour vérifier la somme.
J’avoue que je fus prêt à me jeter à ses pieds, à lui déclarer que
j’étais un brigand, un infâme, et... pis que cela, un menteur! La honte
me retint. J’allais l’embrasser, il me repoussa faiblement.—
Maintenant, tu es un homme, mon enfant, me dit-il. Ce que je fais
est une chose simple et juste dont tu ne dois pas me remercier. Si
j’ai droit à votre reconnaissance, Raphaël, reprit-il d’un ton doux
mais plein de dignité, c’est pour avoir préservé votre jeunesse des
malheurs qui dévorent tous les jeunes gens, à Paris. Désormais,
nous serons deux amis. Vous deviendrez, dans un an, docteur en
droit. Vous avez, non sans quelques déplaisirs et certaines
privations, acquis les connaissances solides et l’amour du travail si
nécessaires aux hommes appelés à manier les affaires. Apprenez,
Raphaël, à me connaître. Je ne veux faire de vous ni un avocat, ni
un notaire, mais un homme d’état qui puisse devenir la gloire de
notre pauvre maison. A demain! ajouta-t-il en me renvoyant par un
geste mystérieux. Dès ce jour, mon père m’initia franchement à ses
projets. J’étais fils unique et j’avais perdu ma mère depuis dix ans.
Autrefois, peu flatté d’avoir le droit de labourer la terre l’épée au
côté, mon père, chef d’une maison historique à peu près oubliée en
Auvergne, vint à Paris pour y tenter le diable. Doué de cette finesse
qui rend les hommes du midi de la France si supérieurs quand elle
se trouve accompagnée d’énergie, il était parvenu sans grand appui
à prendre position au cœur même du pouvoir. La révolution renversa
bientôt sa fortune; mais il avait su épouser l’héritière d’une grande
maison, et s’était vu sous l’empire au moment de restituer à notre
famille son ancienne splendeur. La restauration, qui rendit à ma
mère des biens considérables, ruina mon père. Ayant jadis acheté
plusieurs terres données par l’empereur à ses généraux et situées
en pays étranger, il luttait depuis dix ans avec des liquidateurs et des
diplomates, avec les tribunaux prussiens et bavarois pour se
maintenir dans la possession contestée de ces malheureuses
dotations. Mon père me jeta dans le labyrinthe inextricable de ce
vaste procès d’où dépendait notre avenir. Nous pouvions être
condamnés à restituer les revenus par lui perçus, ainsi que le prix de
certaines coupes de bois faites de 1814 à 1817; dans ce cas, le bien
de ma mère suffisait à peine pour sauver l’honneur de notre nom.
Ainsi, le jour où mon père parut en quelque sorte m’avoir émancipé,
je tombai sous le joug le plus odieux. Je dus combattre comme sur
un champ de bataille, travailler nuit et jour, aller voir des hommes
d’état, tâcher de surprendre leur religion, tenter de les intéresser à
notre affaire, les séduire, eux, leurs femmes, leurs valets, leurs
chiens, et déguiser cet horrible métier sous des formes élégantes,
sous d’agréables plaisanteries. Je compris tous les chagrins dont
l’empreinte flétrissait la figure de mon père. Pendant une année
environ, je menai donc en apparence la vie d’un homme du monde,
mais cette dissipation et mon empressement à me lier avec des
parents en faveur ou avec des gens qui pouvaient nous être utiles,
cachaient d’immenses travaux. Mes divertissements étaient encore
des plaidoiries, et mes conversations des mémoires. Jusque-là,
j’avais été vertueux par l’impossibilité de me livrer à mes passions
de jeune homme; mais craignant alors de causer la ruine de mon
père ou la mienne par une négligence, je devins mon propre
despote, et n’osai me permettre ni un plaisir ni une dépense.
Lorsque nous sommes jeunes, quand, à force de froissements, les
hommes et les choses ne nous ont point encore enlevé cette
délicate fleur de sentiment, cette verdeur de pensée, cette noble
pureté de conscience qui ne nous laisse jamais transiger avec le
mal, nous sentons vivement nos devoirs; notre honneur parle haut et
se fait écouter; nous sommes francs et sans détour: ainsi étais-je
alors. Je voulus justifier la confiance de mon père. Naguère, je lui
aurais dérobé délicieusement une chétive somme; mais portant avec
lui le fardeau de ses affaires, de son nom, de sa maison, je lui eusse
donné secrètement mes biens, mes espérances, comme je lui
sacrifiais mes plaisirs; heureux même de mon sacrifice! Aussi,
quand monsieur de Villèle exhuma, tout exprès pour nous, un décret
impérial sur les déchéances, et nous eut ruinés, signai-je la vente de
mes propriétés, n’en gardant qu’une île sans valeur, située au milieu
de la Loire, et où se trouvait le tombeau de ma mère. Aujourd’hui,
peut-être, les arguments, les détours, les discussions
philosophiques, philanthropiques et politiques ne me manqueraient
pas pour me dispenser de faire ce que mon avoué nommait une
bêtise. Mais à vingt et un ans, nous sommes, je le répète, tout
générosité, tout chaleur, tout amour. Les larmes que je vis dans les
yeux de mon père furent alors pour moi la plus belle des fortunes, et
le souvenir de ces larmes a souvent consolé ma misère. Dix mois
après avoir payé ses créanciers, mon père mourut de chagrin. Il
m’adorait et m’avait ruiné; cette idée le tua. En 1826, à l’âge de
vingt-deux ans, vers la fin de l’automne, je suivis tout seul le convoi
de mon premier ami, de mon père. Peu de jeunes gens se sont
trouvés, seuls avec leurs pensées, derrière un corbillard, perdus
dans Paris, sans avenir, sans fortune. Les orphelins recueillis par la
charité publique ont au moins pour avenir le champ de bataille, pour
père le gouvernement ou le procureur du roi, pour refuge un
hospice. Moi, je n’avais rien! Trois mois après, un commissaire-
priseur me remit onze cent douze francs, produit net et liquide de la
succession paternelle. Des créanciers m’avaient obligé à vendre
notre mobilier. Accoutumé dès ma jeunesse à donner une grande
valeur aux objets de luxe dont j’étais entouré, je ne pus m’empêcher
de marquer une sorte d’étonnement à l’aspect de ce reliquat exigu.
—«Oh! me dit le commissaire-priseur, tout cela était bien rococo.»
Mot épouvantable qui flétrissait toutes les religions de mon enfance
et me dépouillait de mes premières illusions, les plus chères de
toutes. Ma fortune se résumait par un bordereau de vente, mon
avenir gisait dans un sac de toile qui contenait onze cent douze
francs, la société m’apparaissait en la personne d’un huissier-priseur
qui me parlait le chapeau sur la tête. Un valet de chambre qui me
chérissait, et auquel ma mère avait jadis constitué quatre cents
francs de rente viagère, Jonathas me dit en quittant la maison d’où
j’étais si souvent sorti joyeusement en voiture pendant mon enfance:
—Soyez bien économe, monsieur Raphaël! Il pleurait, le bon
homme. Tels sont, mon cher Émile, les événements qui maîtrisèrent
ma destinée, modifièrent mon âme, et me placèrent jeune encore
dans la plus fausse de toutes les situations sociales. Des liens de
famille, mais faibles, m’attachaient à quelques maisons riches dont
l’accès m’eût été interdit par ma fierté, si le mépris et l’indifférence
ne m’en eussent déjà fermé les portes. Quoique parent de
personnes très-influentes et prodigues de leur protection pour des
étrangers, je n’avais ni parents ni protecteurs. Sans cesse arrêtée
dans ses expansions, mon âme s’était repliée sur elle-même: plein
de franchise et de naturel, je devais paraître froid, dissimulé; le
despotisme de mon père m’avait ôté toute confiance en moi; j’étais
timide et gauche, je ne croyais pas que ma voix pût exercer le
moindre empire, je me déplaisais, je me trouvais laid, j’avais honte
de mon regard. Malgré la voix intérieure qui doit soutenir les
hommes de talent dans leurs luttes, et qui me criait: Courage!
marche! malgré les révélations soudaines de ma puissance dans la
solitude, malgré l’espoir dont j’étais animé en comparant les
ouvrages nouveaux admirés du public à ceux qui voltigeaient dans
ma pensée, je doutais de moi comme un enfant. J’étais la proie
d’une excessive ambition, je me croyais destiné à de grandes
choses, et me sentais dans le néant. J’avais besoin des hommes, et
je me trouvais sans amis; je devais me frayer une route dans le
monde, et j’y restais seul, moins craintif que honteux. Pendant
l’année où je fus jeté par mon père dans le tourbillon de la haute
société, j’y vins avec un cœur neuf, avec une âme fraîche. Comme
tous les grands enfants, j’aspirai secrètement à de belles amours. Je
rencontrai parmi les jeunes gens de mon âge une secte de fanfarons
qui allaient tête levée, disant des riens, s’asseyant sans trembler
près des femmes qui me semblaient les plus imposantes, débitant
des impertinences, mâchant le bout de leurs cannes, minaudant, se
prostituant à eux-mêmes les plus jolies personnes, mettant ou
prétendant avoir mis leurs têtes sur tous les oreillers, ayant l’air
d’être au refus du plaisir, considérant les plus vertueuses, les plus
prudes comme de prise facile et pouvant être conquises à la simple
parole, au moindre geste hardi, par le premier regard insolent! Je te
le déclare, en mon âme et conscience, la conquête du pouvoir ou
d’une grande renommée littéraire me paraissait un triomphe moins
difficile à obtenir qu’un succès auprès d’une femme de haut rang,
jeune, spirituelle et gracieuse. Je trouvai donc les troubles de mon
cœur, mes sentiments, mes cultes en désaccord avec les maximes
de la société. J’avais de la hardiesse, mais dans l’âme seulement, et
non dans les manières. J’ai su plus tard que les femmes ne
voulaient pas être mendiées. J’en ai beaucoup vu que j’adorais de
loin, auxquelles je livrais un cœur à toute épreuve, une âme à
déchirer, une énergie qui ne s’effrayait ni des sacrifices, ni des
tortures; elles appartenaient à des sots dont je n’aurais pas voulu
pour portiers. Combien de fois, muet, immobile, n’ai-je pas admiré la
femme de mes rêves, surgissant dans un bal! Dévouant alors en
pensée mon existence à des caresses éternelles, j’imprimais toutes
mes espérances en un regard, et lui offrais dans mon extase un
amour de jeune homme qui courait au-devant des tromperies. En
certains moments, j’aurais donné ma vie pour une seule nuit. Eh
bien! n’ayant jamais trouvé d’oreilles à qui confier mes propos
passionnés, de regards où reposer les miens, de cœur pour mon
cœur, j’ai vécu dans tous les tourments d’une impuissante énergie
qui se dévorait elle-même, soit faute de hardiesse ou d’occasions,
soit inexpérience. Peut-être ai-je désespéré de me faire comprendre,
ou tremblé d’être trop compris. Et cependant j’avais un orage tout
prêt à chaque regard poli que l’on pouvait m’adresser. Malgré ma
promptitude à prendre ce regard ou des mots en apparence
affectueux comme de tendres engagements, je n’ai jamais osé ni
parler ni me taire à propos. A force de sentiment ma parole était
insignifiante, et mon silence était stupide. J’avais sans doute trop de
naïveté pour une société factice qui vit aux lumières, et rend toutes
ses pensées par des phrases convenues, ou des mots que dicte la
mode. Puis je ne savais point parler en me taisant, ni me taire en
parlant. Enfin, gardant en moi des feux qui me brûlaient, ayant une
âme semblable à celles que les femmes souhaitent de rencontrer, en
proie à cette exaltation dont elles sont avides, possédant l’énergie
dont se vantent les sots, toutes les femmes m’ont été traîtreusement
cruelles. Aussi, admirais-je naïvement les héros de coterie quand ils
célébraient leurs triomphes, sans les soupçonner de mensonge.
J’avais sans doute le tort de désirer un amour sur parole, de vouloir
trouver grande et forte dans un cœur de femme frivole et légère,
affamée de luxe, ivre de vanité, cette passion large, cet océan qui
battait tempêtueusement dans mon cœur. Oh! se sentir né pour
aimer, pour rendre une femme bien heureuse, et ne pas avoir trouvé
même une courageuse et noble Marceline ou quelque vieille
marquise! Porter des trésors dans une besace et ne pouvoir
rencontrer personne, pas même une enfant, quelque jeune fille
curieuse, pour les lui faire admirer. J’ai souvent voulu me tuer de
désespoir.
—Joliment tragique ce soir! s’écria Émile.
—Eh! laisse-moi condamner ma vie, répondit Raphaël. Si ton
amitié n’a pas la force d’écouter mes élégies, si tu ne peux me faire
crédit d’une demi-heure d’ennui, dors! Mais ne me demande plus
compte de mon suicide qui gronde, qui se dresse, qui m’appelle et
que je salue. Pour juger un homme, au moins faut-il être dans le
secret de sa pensée, de ses malheurs, de ses émotions; ne vouloir
connaître de sa vie que les événements matériels, c’est faire de la
chronologie, l’histoire des sots!
Le ton amer avec lequel ces paroles furent prononcées frappa si
vivement Émile que, dès ce moment, il prêta toute son attention à
Raphaël en le regardant d’un air hébété.
—Mais, reprit le narrateur, maintenant la lueur qui colore ces
accidents leur prête un nouvel aspect. L’ordre des choses que je
considérais jadis comme un malheur a peut-être engendré les belles
facultés dont plus tard je me suis enorgueilli. La curiosité
philosophique, les travaux excessifs, l’amour de la lecture qui,
depuis l’âge de sept ans jusqu’à mon entrée dans le monde, ont
constamment occupé ma vie, ne m’auraient-ils pas doué de la facile
puissance avec laquelle, s’il faut vous en croire, je sais rendre mes
idées et marcher en avant dans le vaste champ des connaissances
humaines? L’abandon auquel j’étais condamné, l’habitude de
refouler mes sentiments et de vivre dans mon cœur ne m’ont-ils pas
investi du pouvoir de comparer, de méditer? En ne se perdant pas
au service des irritations mondaines qui rapetissent la plus belle âme
et la réduisent à l’état de guenille, ma sensibilité ne s’est-elle pas
concentrée pour devenir l’organe perfectionné d’une volonté plus
haute que le vouloir de la passion? Méconnu par les femmes, je me
souviens de les avoir observées avec la sagacité de l’amour
dédaigné. Maintenant, je le vois, la sincérité de mon caractère a dû
leur déplaire! Peut-être veulent-elles un peu d’hypocrisie? Moi qui
suis tour à tour, dans la même heure, homme et enfant, futile et
penseur, sans préjugés et plein de superstitions, souvent femme
comme elles, n’ont-elles pas dû prendre ma naïveté pour du
cynisme, et la pureté même de ma pensée pour du libertinage? La
science leur était ennui, la langueur féminine faiblesse. Cette
excessive mobilité d’imagination, le malheur des poètes, me faisait
sans doute juger comme un être incapable d’amour, sans constance
dans les idées, sans énergie. Idiot quand je me taisais, je les
effarouchais peut-être quand j’essayais de leur plaire. Les femmes
m’ont condamné. J’ai accepté, dans les larmes et le chagrin, l’arrêt
porté par le monde. Cette peine a produit son fruit. Je voulus me
venger de la société, je voulus posséder l’âme de toutes les femmes
en me soumettant les intelligences, et voir tous les regards fixés sur
moi quand mon nom serait prononcé par un valet à la porte d’un
salon. Je m’instituai grand homme. Dès mon enfance, je m’étais
frappé le front en me disant comme André de Chénier: «Il y a
quelque chose là!» Je croyais sentir en moi une pensée à exprimer,
un système à établir, une science à expliquer. O mon cher Émile!
aujourd’hui que j’ai vingt-six ans à peine, que je suis sûr de mourir
inconnu, sans avoir jamais été l’amant de la femme que j’ai rêvé de
posséder, laisse-moi te conter mes folies! N’avons-nous pas tous,
plus ou moins, pris nos désirs pour des réalités? Ah! je ne voudrais
point pour ami d’un jeune homme qui dans ses rêves ne se serait
pas tressé des couronnes, construit quelque piédestal ou donné de
complaisantes maîtresses. Moi, j’ai souvent été général, empereur;
j’ai été Byron, puis rien. Après avoir joué sur le faîte des choses
humaines, je m’apercevais que toutes les montagnes, toutes les
difficultés restaient à gravir. Cet immense amour-propre qui
bouillonnait en moi, cette croyance sublime à une destinée, et qui
devient du génie peut-être, quand un homme ne se laisse pas
déchiqueter l’âme par le contact des affaires aussi facilement qu’un
mouton abandonne sa laine aux épines des halliers où il passe, tout
cela me sauva. Je voulus me couvrir de gloire et travailler dans le
silence pour la maîtresse que j’espérais avoir un jour. Toutes les
femmes se résumaient par une seule, et cette femme je croyais la
rencontrer dans la première qui s’offrait à mes regards. Mais, voyant
une reine dans chacune d’elles, toutes devaient, comme les reines
qui sont obligées de faire des avances à leurs amants, venir un peu
au-devant de moi, souffreteux, pauvre et timide. Ah! pour celle qui
m’eût plaint, j’avais dans le cœur tant de reconnaissance outre
l’amour, que je l’eusse adorée pendant toute sa vie. Plus tard, mes
observations m’ont appris de cruelles vérités. Ainsi, mon cher Émile,
je risquais de vivre éternellement seul. Les femmes sont habituées,
par je ne sais quelle pente de leur esprit, à ne voir dans un homme
de talent que ses défauts, et dans un sot que ses qualités; elles
éprouvent de grandes sympathies pour les qualités du sot qui sont
une flatterie perpétuelle de leurs propres défauts, tandis que
l’homme supérieur ne leur offre pas assez de jouissances pour
compenser ses imperfections. Le talent est une fièvre intermittente,
nulle femme n’est jalouse d’en partager seulement les malaises;
toutes veulent trouver dans leurs amants des motifs de satisfaire leur
vanité; c’est elles encore qu’elles aiment en nous! Un homme
pauvre, fier, artiste, doué du pouvoir de créer, n’est-il pas armé d’un
blessant égoïsme? Il existe autour de lui je ne sais quel tourbillon de
pensées dans lequel il enveloppe tout, même sa maîtresse, qui doit
en suivre le mouvement. Une femme adulée peut-elle croire à
l’amour d’un tel homme? Ira-t-elle le chercher? Cet amant n’a pas le
loisir de s’abandonner autour d’un divan à ces petites singeries de
sensibilité auxquelles les femmes tiennent tant et qui sont le
triomphe des gens faux et insensibles. Le temps manque à ses
travaux, comment en dépenserait-il à se rapetisser, à se chamarrer?
Prêt à donner ma vie d’un coup, je ne l’aurais pas avilie en détail.
Enfin il existe, dans le manége d’un agent de change qui fait les
commissions d’une femme pâle et minaudière, je ne sais quoi de
mesquin dont l’artiste a horreur. L’amour abstrait ne suffit pas à un
homme pauvre et grand, il en veut tous les dévouements. Les
petites créatures qui passent leur vie à essayer des cachemires ou
se font les porte-manteaux de la mode, n’ont pas de dévouement,
elles en exigent et voient dans l’amour le plaisir de commander, non
celui d’obéir. La véritable épouse en cœur, en chair et en os, se
laisse traîner là où va celui en qui réside sa vie, sa force, sa gloire,
son bonheur. Aux hommes supérieurs, il faut des femmes orientales
dont l’unique pensée soit l’étude de leurs besoins: pour eux, le
malheur est dans le désaccord de leurs désirs et des moyens. Moi,
qui me croyais homme de génie, j’aimais précisément ces petites-
maîtresses! Nourrissant des idées si contraires aux idées reçues,
ayant la prétention d’escalader le ciel sans échelle, possédant des
trésors qui n’avaient pas cours, armé de connaissances étendues
qui surchargeaient ma mémoire et que je n’avais pas encore
classées, que je ne m’étais point assimilées; me trouvant sans
parents, sans amis, seul au milieu du plus affreux désert, un désert
pavé, un désert animé, pensant, vivant, où tout vous est bien plus
qu’ennemi, indifférent! la résolution que je pris était naturelle,
quoique folle; elle comportait je ne sais quoi d’impossible qui me
donna du courage. Ce fut comme un parti fait avec moi-même, et
dont j’étais le joueur et l’enjeu. Voici mon plan. Mes onze cents
francs devaient suffire à ma vie pendant trois ans; je m’accordais ce
temps pour mettre au jour un ouvrage qui pût attirer l’attention
publique sur moi, me faire une fortune ou un nom. Je me réjouissais
en pensant que j’allais vivre de pain et de lait, comme un solitaire de
la Thébaïde, plongé dans le monde des livres et des idées, dans une
sphère inaccessible, au milieu de ce Paris si tumultueux, sphère de
travail et de silence, où, comme les chrysalides, je me bâtissais une
tombe pour renaître brillant et glorieux. J’allais risquer de mourir
pour vivre. En réduisant l’existence à ses vrais besoins, au strict
nécessaire, je trouvais que trois cent soixante-cinq francs par an
devaient suffire à ma pauvreté. En effet, cette maigre somme a
satisfait à ma vie, tant que j’ai voulu subir ma propre discipline
claustrale.
—C’est impossible, s’écria Émile.
—J’ai vécu près de trois ans ainsi, répondit Raphaël avec une
sorte de fierté. Comptons! reprit-il. Trois sous de pain, deux sous de
lait, trois sous de charcuterie m’empêchaient de mourir de faim et
tenaient mon esprit dans un état de lucidité singulière. J’ai observé,
tu le sais, de merveilleux effets produits par la diète sur l’imagination.
Mon logement me coûtait trois sous par jour, je brûlais pour trois
sous d’huile par nuit, je faisais moi-même ma chambre, je portais
des chemises de flanelle pour ne dépenser que deux sous de
blanchissage par jour. Je me chauffais avec du charbon de terre,
dont le prix divisé par les jours de l’année n’a jamais donné plus de
deux sous jour chacun; j’avais des habits, du linge, des chaussures
pour trois années, je ne voulais m’habiller que pour aller à certains
cours publics et aux bibliothèques. Ces dépenses réunies ne
faisaient que dix-huit sous, il me restait deux sous pour les choses
imprévues. Je ne me souviens pas d’avoir, pendant cette longue
période de travail, passé le Pont-des-Arts, ni d’avoir jamais acheté
d’eau; j’allais en chercher le matin, à la fontaine de la place Saint-
Michel, au coin de la rue des Grès. Oh! je portais ma pauvreté
fièrement. Un homme qui pressent un bel avenir marche dans sa vie
de misère comme un innocent conduit au supplice, il n’a point honte.
Je n’avais pas voulu prévoir la maladie: comme Aquilina,
j’envisageais l’hôpital sans terreur. Je n’ai pas douté un moment de
ma bonne santé. D’ailleurs, le pauvre ne doit se coucher que pour
mourir. Je me coupai les cheveux, jusqu’au moment où un ange
d’amour ou de bonté... Mais je ne veux pas anticiper sur la situation
à laquelle j’arrive. Apprends seulement, mon cher ami, qu’à défaut
de maîtresse, je vécus avec une grande pensée, avec un rêve, un
mensonge auquel nous commençons tous par croire plus ou moins.
Aujourd’hui je ris de moi, de ce moi, peut-être saint et sublime, qui
n’existe plus. La société, le monde, nos usages, nos mœurs, vus de
près, m’ont révélé le danger de ma croyance innocente et la
superfluité de mes fervents travaux. Ces approvisionnements sont
inutiles à l’ambitieux: que léger soit le bagage de qui poursuit la
fortune. La faute des hommes supérieurs est de dépenser leurs
jeunes années à se rendre dignes de la faveur. Pendant qu’ils
thésaurisent, leur force est la science pour porter sans effort le poids
d’une puissance qui les fuit; les intrigants, riches de mots et
dépourvus d’idées, vont et viennent, surprennent les sots, et se
logent dans la confiance des demi-niais: les uns étudient, les autres
marchent; les uns sont modestes, les autres hardis; l’homme de
génie tait son orgueil, l’intrigant arbore le sien et doit arriver
nécessairement. Les hommes du pouvoir ont si fort besoin de croire
au mérite tout fait, au talent effronté, qu’il y a chez le vrai savant de
l’enfantillage à espérer les récompenses humaines. Je ne cherche
certes pas à paraphraser les lieux communs de la vertu, le cantique
des cantiques éternellement chanté par les génies méconnus; je
veux déduire logiquement la raison des fréquents succès obtenus
par les hommes médiocres. Hélas! l’étude est si maternellement
bonne, qu’il y a peut-être crime à lui demander des récompenses
autres que les pures et douces joies dont elle nourrit ses enfants. Je
me souviens d’avoir quelquefois trempé gaiement mon pain dans
mon lait, assis auprès de ma fenêtre en y respirant l’air, en laissant
planer mes yeux sur un paysage de toits bruns, grisâtres, rouges, en
ardoises, en tuiles, couverts de mousses jaunes ou vertes. Si
d’abord cette vue me parut monotone, j’y découvris bientôt de
singulières beautés: tantôt le soir des raies lumineuses, parties des
volets mal fermés, nuançaient et animaient les noires profondeurs
de ce pays original; tantôt les lueurs pâles des réverbères projetaient
d’en bas des reflets jaunâtres à travers le brouillard, et accusaient
faiblement dans les rues les ondulations de ces toits pressés, océan
de vagues immobiles; parfois de rares figures apparaissaient au
milieu de ce morne désert. Parmi les fleurs de quelque jardin aérien,
j’entrevoyais le profil anguleux et crochu d’une vieille femme
arrosant des capucines, ou dans le cadre d’une lucarne pourrie
quelque jeune fille faisant sa toilette, se croyant seule, et dont je ne
pouvais apercevoir que le beau front et les longs cheveux élevés en
l’air par un joli bras blanc. J’admirais dans les gouttières quelques
végétations éphémères, pauvres herbes bientôt emportées par un
orage! J’étudiais les mousses, leurs couleurs ravivées par la pluie, et
qui sous le soleil se changeaient en un velours sec et brun à reflets
capricieux. Enfin les poétiques et fugitifs effets du jour, les tristesses
du brouillard, les soudains scintillements du soleil, le silence et les
magies de la nuit, les mystères de l’aurore, les fumées de chaque
cheminée, tous les accidents de cette singulière nature m’étaient
devenus familiers et me divertissaient. J’aimais ma prison, elle était
volontaire. Ces savanes de Paris formées par des toits nivelés
comme une plaine, mais qui couvraient des abîmes peuplés, allaient
à mon âme et s’harmoniaient avec mes pensées. Il est fatiguant de
retrouver brusquement le monde quand nous descendons des
hauteurs célestes où nous entraînent les méditations scientifiques.
Aussi ai-je alors parfaitement conçu la nudité des monastères.
Quand je fus bien résolu à suivre mon nouveau plan de vie, je
cherchai mon logis dans les quartiers les plus déserts de Paris. Un
soir, en revenant de l’Estrapade, je passais par la rue des Cordiers
pour retourner chez moi. A l’angle de la rue de Cluny, je vis une
petite fille d’environ quatorze ans, qui jouait au volant avec une de
ses camarades, et dont les rires et les espiègleries amusaient les
voisins. Il faisait beau, la soirée était chaude, le mois de septembre
durait encore. Devant chaque porte, des femmes étaient assises et
devisaient comme dans une ville de province par un jour de fête.
J’observai d’abord la jeune fille, dont la physionomie était d’une
admirable expression, et le corps tout posé pour un peintre. C’était
une scène ravissante. Je cherchai la cause de cette bonhomie au
milieu de Paris, je remarquai que la rue n’aboutissait à rien, et ne
devait pas être très-passante. En me rappelant le séjour de J.-J.
Rousseau dans ce lieu, je trouvai l’hôtel Saint-Quentin, et le
délabrement dans lequel il était me fit espérer d’y rencontrer un gîte
peu coûteux. Je voulus le visiter. En entrant dans une chambre
basse, je vis les classiques flambeaux de cuivre garnis de leurs
chandelles, méthodiquement rangés au-dessus de chaque clef, et
fus frappé de la propreté qui régnait dans cette salle, ordinairement
assez mal tenue dans les autres hôtels. Elle était peignée comme un
tableau de genre: son lit bleu, les ustensiles, les meubles, avaient la
coquetterie d’une nature de convention. La maîtresse de l’hôtel,
femme de quarante ans environ, dont les traits exprimaient des
malheurs, dont le regard était comme terni par des pleurs, se leva,
vint à moi; je lui soumis humblement le tarif de mon loyer. Sans en
paraître étonnée, elle chercha une clef parmi toutes les autres, et me
conduisit dans les mansardes, où elle me montra une chambre qui
avait vue sur les toits, sur les cours des maisons voisines, par les
fenêtres desquelles passaient de longues perches chargées de
linge. Rien n’était plus horrible que cette mansarde aux murs jaunes
et sales, qui sentait la misère et appelait son savant. La toiture s’y
abaissait régulièrement et les tuiles disjointes laissaient voir le ciel. Il
y avait place pour un lit, une table, quelques chaises, et sous l’angle
aigu du toit je pouvais loger mon piano. N’étant pas assez riche pour
meubler cette cage digne des plombs de Venise, la pauvre femme
n’avait jamais pu la louer. Ayant précisément excepté de la vente
mobilière que je venais de faire les objets qui m’étaient en quelque
sorte personnels, je fus bientôt d’accord avec mon hôtesse, et
m’installai le lendemain chez elle. Je vécus dans ce sépulcre aérien
pendant près de trois ans, travaillant nuit et jour sans relâche, avec
tant de plaisir, que l’étude me semblait être le plus beau thème, la
plus heureuse solution de la vie humaine. Le calme et le silence
nécessaires au savant ont je ne sais quoi de doux, d’enivrant
comme l’amour. L’exercice de la pensée, la recherche des idées, les
contemplations tranquilles de la science nous prodiguent d’ineffables
délices, indescriptibles comme tout ce qui participe de l’intelligence,
dont les phénomènes sont invisibles à nos sens extérieurs. Aussi
sommes-nous toujours forcés d’expliquer les mystères de l’esprit par
des comparaisons matérielles. Le plaisir de nager dans un lac d’eau
pure, au milieu des rochers, des bois et des fleurs, seul et caressé
par une brise tiède, donnerait aux ignorants une bien faible image du
bonheur que j’éprouvais quand mon âme était baignée dans les
lueurs de je ne sais quelle lumière, quand j’écoutais les voix terribles
et confuses de l’inspiration, quand d’une source inconnue les images
ruisselaient dans mon cerveau palpitant. Voir une idée qui pointe
dans le champ des abstractions humaines comme le lever du soleil
au matin et s’élève comme lui, qui, mieux encore, grandit comme un
enfant, arrive à la puberté, se fait lentement virile, est une joie
supérieure aux autres joies terrestres, ou plutôt c’est un divin plaisir.
L’étude prête une sorte de magie à tout ce qui nous environne. Le
bureau chétif sur lequel j’écrivais, et la basane brune qui le couvrait,
mon piano, mon lit, mon fauteuil, les bizarreries de mon papier de
tenture, mes meubles, toutes ces choses s’animèrent, et devinrent
pour moi d’humbles amis, les complices silencieux de mon avenir.
Combien de fois ne leur ai-je pas communiqué mon âme, en les
regardant? Souvent, en laissant voyager mes yeux sur une moulure
déjetée, je rencontrais des développements nouveaux, une preuve
frappante de mon système ou des mots que je croyais heureux pour
rendre des pensées presque intraduisibles. A force de contempler
les objets qui m’entouraient, je trouvais à chacun sa physionomie,
son caractère; souvent ils me parlaient: si, par-dessus les toits, le
soleil couchant jetait à travers mon étroite fenêtre quelque lueur
furtive, ils se coloraient, pâlissaient, brillaient, s’attristaient ou
s’égayaient, en me surprenant toujours par des effets nouveaux.
Ces menus accidents de la vie solitaire, qui échappent aux
préoccupations du monde, sont la consolation des prisonniers.
N’étais-je pas captivé par une idée, emprisonné dans un système;
mais soutenu par la perspective d’une vie glorieuse! A chaque
difficulté vaincue, je baisais les mains douces de la femme aux
beaux yeux, élégante et riche, qui devait un jour caresser mes
cheveux en me disant avec attendrissement: Tu as bien souffert,
pauvre ange! J’avais entrepris deux grandes œuvres. Une comédie
devait en peu de jours me donner une renommée, une fortune, et
l’entrée de ce monde, où je voulais reparaître en y exerçant les
droits régaliens de l’homme de génie. Vous avez tous vu dans ce
chef-d’œuvre la première erreur d’un jeune homme qui sort du
collége, une véritable niaiserie d’enfant. Vos plaisanteries ont détruit
de fécondes illusions, qui depuis ne se sont plus réveillées. Toi seul,
mon cher Émile, as calmé la plaie profonde que d’autres firent à mon
cœur! Toi seul admiras ma Théorie de la volonté, ce long ouvrage
pour lequel j’avais appris les langues orientales, l’anatomie, la
physiologie, auquel j’avais consacré la plus grande partie de mon
temps; œuvre qui, si je ne me trompe, complétera les travaux de
Mesmer, de Lavater, de Gall, de Bichat, en ouvrant une nouvelle
route à la science humaine. Là s’arrête ma belle vie, ce sacrifice de
tous les jours, ce travail de ver-à-soie inconnu au monde et dont la
seule récompense est peut-être dans le travail même. Depuis l’âge
de raison jusqu’au jour où j’eus terminé ma théorie, j’ai observé,
appris, écrit, lu sans relâche, et ma vie fut comme un long pensum.
Amant efféminé de la paresse orientale, amoureux de mes rêves,
sensuel, j’ai toujours travaillé, me refusant à goûter les jouissances
de la vie parisienne. Gourmand, j’ai été sobre; aimant et la marche
et les voyages maritimes, désirant visiter plusieurs pays, trouvant
encore du plaisir à faire, comme un enfant, ricocher des cailloux sur
l’eau, je suis resté constamment assis, une plume à la main; bavard,
j’allais écouter en silence les professeurs aux Cours publics de la
Bibliothèque et du Muséum; j’ai dormi sur mon grabat solitaire
comme un religieux de l’ordre de Saint-Benoît, et la femme était
cependant ma seule chimère, une chimère que je caressais et qui
me fuyait toujours! Enfin ma vie a été une cruelle antithèse, un
perpétuel mensonge. Puis jugez donc les hommes! Parfois mes
goûts naturels se réveillaient comme un incendie long-temps couvé.
Par une sorte de mirage ou de calenture, moi, veuf de toutes les
femmes que je désirais, dénué de tout et logé dans une mansarde
d’artiste, je me voyais alors entouré de maîtresses ravissantes! Je
courais à travers les rues de Paris, couché sur les moelleux
coussins d’un brillant équipage! J’étais rongé de vices, plongé dans
la débauche, voulant tout, ayant tout; enfin ivre à jeun, comme saint
Antoine dans sa tentation. Heureusement le sommeil finissait par
éteindre ces visions dévorantes; le lendemain la science m’appelait
en souriant, et je lui étais fidèle. J’imagine que les femmes dites
vertueuses doivent être souvent la proie de ces tourbillons de folie,
de désirs et de passions, qui s’élèvent en nous, malgré nous. De tels
rêves ne sont pas sans charmes: ne ressemblent-ils pas à ces
causeries du soir, en hiver, où l’on part de son foyer pour aller en
Chine. Mais que devient la vertu, pendant ces délicieux voyages où
la pensée a franchi tous les obstacles? Pendant les dix premiers
mois de ma réclusion, je menai la vie pauvre et solitaire que je t’ai
dépeinte: j’allais chercher moi-même, dès le matin et sans être vu,
mes provisions pour la journée; je faisais ma chambre, j’étais tout
ensemble le maître et le serviteur, je diogénisais avec une incroyable
fierté. Mais après ce temps, pendant lequel l’hôtesse et sa fille
espionnèrent mes mœurs et mes habitudes, examinèrent ma
personne et comprirent ma misère, peut-être parce qu’elles étaient
elles-mêmes fort malheureuses, il s’établit d’inévitables liens entre
elles et moi. Pauline, cette charmante créature dont les grâces
naïves et secrètes m’avaient en quelque sorte amené là, me rendit
plusieurs services qu’il me fut impossible de refuser. Toutes les
infortunes sont sœurs: elles ont le même langage, la même
générosité, la générosité de ceux qui ne possédant rien sont
prodigues de sentiment; paient de leur temps et de leur personne.
Insensiblement Pauline s’impatronisa chez moi, voulut me servir et
sa mère ne s’y opposa point. Je vis la mère elle-même
raccommodant mon linge et rougissant d’être surprise à cette
charitable occupation. Devenu malgré moi leur protégé, j’acceptai
leurs services. Pour comprendre cette singulière affection, il faut
connaître l’emportement du travail, la tyrannie des idées de cette
répugnance instinctive qu’éprouve pour les détails de la vie
matérielle l’homme qui vit par la pensée. Pouvais-je résister à la
délicate attention avec laquelle Pauline m’apportait à pas muets mon
repas frugal, quand elle s’apercevait que, depuis sept ou huit
heures, je n’avais rien pris? Avec les grâces de la femme et
l’ingénuité de l’enfance, elle me souriait en faisant un signe pour me
dire que je ne devais pas la voir. C’était Ariel se glissant comme un
sylphe sous mon toit, et prévoyant mes besoins. Un soir, Pauline me
raconta son histoire avec une touchante ingénuité. Son père était
chef d’escadron dans les grenadiers à cheval de la garde impériale.
Au passage de la Bérésina, il avait été fait prisonnier par les
Cosaques. Plus tard, quand Napoléon proposa de l’échanger, les
autorités russes le firent vainement chercher en Sibérie. Au dire des
autres prisonniers, il s’était échappé avec le projet d’aller aux Indes.
Depuis ce temps, madame Gaudin, mon hôtesse, n’avait pu obtenir
aucune nouvelle de son mari. Les désastres de 1814 et 1815 étaient
arrivés. Seule, sans ressources et sans secours, elle avait pris le
parti de tenir un hôtel garni pour faire vivre sa fille. Elle espérait
toujours revoir son mari. Son plus cruel chagrin était de laisser
Pauline sans éducation, sa Pauline, filleule de la princesse
Borghèse, et qui n’aurait pas dû mentir aux belles destinées
promises par son impériale protectrice. Quand madame Gaudin me
confia cette amère douleur qui la tuait, et me dit avec un accent
déchirant: «Je donnerais bien et le chiffon de papier qui crée Gaudin
baron de l’empire, et le droit que nous avons à la dotation de
Witschnau, pour savoir Pauline élevée à Saint-Denis!» tout à coup je
tressaillis, et pour reconnaître les soins que me prodiguaient ces
deux femmes, j’eus l’idée de m’offrir à finir l’éducation de Pauline. La
candeur avec laquelle ces deux femmes acceptèrent ma proposition
fut égale à la naïveté qui la dictait. J’eus ainsi des heures de
récréation. La petite avait les plus heureuses dispositions: elle apprit
avec tant de facilité, qu’elle devint bientôt plus forte que je ne l’étais
sur le piano. En s’accoutumant à penser tout haut près de moi, elle
déployait les mille gentillesses d’un cœur qui s’ouvre à la vie comme
le calice d’une fleur lentement dépliée par le soleil. Elle m’écoutait
avec recueillement et plaisir, en arrêtant sur moi ses yeux noirs et
veloutés qui semblaient sourire. Elle répétait ses leçons d’un accent
doux et caressant, en témoignant une joie enfantine quand j’étais
content d’elle. Sa mère, chaque jour plus inquiète d’avoir à préserver
de tout danger une jeune fille qui développait en croissant toutes les
promesses faites par les grâces de son enfance, la vit avec plaisir
s’enfermer pendant toute la journée pour étudier. Mon piano étant le
seul dont elle pût se servir, elle profitait de mes absences pour
s’exercer. Quand je rentrais, je la trouvais chez moi, dans la toilette
la plus modeste; mais au moindre mouvement, sa taille souple et les
attraits de sa personne se révélaient sous l’étoffe grossière. Elle
avait un pied mignon dans d’ignobles souliers, comme l’héroïne du
conte de Peau-d’Ane. Mais ses jolis trésors, sa richesse de jeune
fille, tout ce luxe de beauté fut comme perdu pour moi. Je m’étais
ordonné à moi-même de ne voir qu’une sœur en Pauline, j’aurais eu
horreur de tromper la confiance de sa mère, j’admirais cette
charmante fille comme un tableau, comme le portrait d’une
maîtresse morte. Enfin, c’était mon enfant, ma statue. Pygmalion
nouveau, je voulais faire d’une vierge vivante et colorée, sensible et
parlante, un marbre. J’étais très-sévère avec elle, mais plus je lui
faisais éprouver les effets de mon despotisme magistral, plus elle
devenait douce et soumise. Si je fus encouragé dans ma retenue et
dans ma continence par des sentiments nobles, néanmoins les
raisons de procureur ne me manquèrent pas. Je ne comprends point
la probité des écus sans la probité de la pensée. Tromper une
femme ou faire faillite a toujours été même chose pour moi. Aimer
une jeune fille ou se laisser aimer par elle constitue un vrai contrat
dont les conditions doivent être bien entendues. Nous sommes
maîtres d’abandonner la femme qui se vend, mais non pas la jeune
fille qui se donne: elle ignore l’étendue de son sacrifice. J’aurais
donc épousé Pauline, et c’eût été une folie: n’était-ce pas livrer une
âme douce et vierge à d’effroyables malheurs? Mon indigence
parlait son langage égoïste, et venait toujours mettre sa main de fer
entre cette bonne créature et moi. Puis, je l’avoue à ma honte, je ne
conçois pas l’amour dans la misère. Peut-être est-ce en moi une
dépravation due à cette maladie humaine que nous nommons la
civilisation; mais une femme, fût-elle attrayante autant que la belle
Hélène, la Galatée d’Homère, n’a plus aucun pouvoir sur mes sens
pour peu qu’elle soit crottée. Ah! vive l’amour dans la soie, sur le
cachemire, entouré des merveilles du luxe qui le parent
merveilleusement bien, parce que lui-même est un luxe peut-être.
J’aime à froisser sous mes désirs de pimpantes toilettes, à briser
des fleurs, à porter une main dévastatrice dans les élégants édifices
d’une coiffure embaumée. Des yeux brûlants, cachés par un voile de
dentelle que les regards percent comme la flamme déchire la fumée
du canon, m’offrent de fantastiques attraits. Mon amour veut des
échelles de soie escaladées en silence, par une nuit d’hiver. Quel
plaisir d’arriver couvert de neige dans une chambre éclairée par des
parfums, tapissée de soies peintes, et d’y trouver une femme qui,
elle aussi, secoue de la neige: car quel autre nom donner à ces
voiles de voluptueuses mousselines à travers lesquels elle se
dessine vaguement comme un ange dans son nuage, et d’où elle va
sortir? Puis il me faut encore un craintif bonheur, une audacieuse
sécurité. Enfin je veux revoir cette mystérieuse femme, mais
éclatante, mais au milieu du monde, mais vertueuse, environnée
d’hommages, vêtue de dentelles, de diamants, donnant ses ordres à
la ville, et si haut placée et si imposante que nul n’ose lui adresser
des vœux. Au milieu de sa cour, elle me jette un regard à la dérobée,
un regard qui dément ces artifices, un regard qui me sacrifie le

You might also like