Control Theory Of Partial Differential Equations

1st Edition Guenter Leugering download

https://ebookbell.com/product/control-theory-of-partial-
differential-equations-1st-edition-guenter-leugering-1007988

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Mathematical Control Theory Of Coupled Systems Of Partial Differential

Equations Irena Lasiecka Author

https://ebookbell.com/product/mathematical-control-theory-of-coupled-
systems-of-partial-differential-equations-irena-lasiecka-
author-2464612

Partial Differential Equations Theory Control And Approximation In

Honor Of The Scientific Heritage Of Jacqueslouis Lions Alain
Bensoussan

https://ebookbell.com/product/partial-differential-equations-theory-
control-and-approximation-in-honor-of-the-scientific-heritage-of-
jacqueslouis-lions-alain-bensoussan-4607134

Control And Estimation Of Dynamical Nonlinear And Partial Differential

Equation Systems Theory And Applications Gerasimos Rigatos

https://ebookbell.com/product/control-and-estimation-of-dynamical-
nonlinear-and-partial-differential-equation-systems-theory-and-
applications-gerasimos-rigatos-44888218

Control Theory Of Digitally Networked Dynamic Systems Jan Lunze

https://ebookbell.com/product/control-theory-of-digitally-networked-
dynamic-systems-jan-lunze-52606424
Control Theory Of Multifingered Hands A Modelling And
Analyticalmechanics Approach For Dexterity And Intelligence 1st
Edition Suguru Arimoto Auth

https://ebookbell.com/product/control-theory-of-multifingered-hands-a-
modelling-and-analyticalmechanics-approach-for-dexterity-and-
intelligence-1st-edition-suguru-arimoto-auth-4240682

Control Theory Of Infinitedimensional Systems 1st Edition Kerner

https://ebookbell.com/product/control-theory-of-infinitedimensional-
systems-1st-edition-kerner-11174678

Risk Control Theory Of Online Transactions Changjun Jiang Wangyang Yu

https://ebookbell.com/product/risk-control-theory-of-online-
transactions-changjun-jiang-wangyang-yu-44546920

The Emotional Mind A Control Theory Of Affective States Cochrane

https://ebookbell.com/product/the-emotional-mind-a-control-theory-of-
affective-states-cochrane-7380408

The Emotional Mind A Control Theory Of Affective States Hardcover Tom

Cochrane

https://ebookbell.com/product/the-emotional-mind-a-control-theory-of-
affective-states-hardcover-tom-cochrane-7429270
Control Theory of
Partial Differential
Equations
 PURE AND APPLIED MATHEMATICS

A Program of Monographs, Textbooks, and Lecture Notes

EXECUTIVE EDITORS

Earl J. Taft Zuhair Nashed

Rutgers University University of Central Florida
New Brunswick, New Jersey Orlando, Florida

EDITORIAL BOARD

M. S. Baouendi Anil Nerode

University of California, Cornell University
San Diego Donald Passman
Jane Cronin University of Wisconsin,
Rutgers University Madison
Jack K. Hale Fred S. Roberts
Georgia Institute of Technology Rutgers University
S. Kobayashi David L. Russell
University of California, Virginia Polytechnic Institute
Berkeley and State University
Marvin Marcus Walter Schempp
University of California, Universität Siegen
Santa Barbara Mark Teply
W. S. Masse University of Wisconsin,
Yale University Milwaukee
 LECTURE NOTES IN
PURE AND APPLIED MATHEMATICS

Recent Titles

R. Costa et al., Nonassociative Algebra and Its Applications

T.-X. He, Wavelet Analysis and Multiresolution Methods
H. Hudzik and L. Skrzypczak, Function Spaces: The Fifth Conference
J. Kajiwara et al., Finite or Infinite Dimensional Complex Analysis
G. Lumer and L. Weis, Evolution Equations and Their Applications in Physical and Life
Sciences
J. Cagnol et al., Shape Optimization and Optimal Design
J. Herzog and G. Restuccia, Geometric and Combinatorial Aspects of Commutative Algebra
G. Chen et al., Control of Nonlinear Distributed Parameter Systems
F. Ali Mehmeti et al., Partial Differential Equations on Multistructures
D. D. Anderson and I. J. Papick, Ideal Theoretic Methods in Commutative Algebra
Á. Granja et al., Ring Theory and Algebraic Geometry
A. K. Katsaras et al., p-adic Functional Analysis
R. Salvi, The Navier-Stokes Equations
F. U. Coelho and H. A. Merklen, Representations of Algebras
S. Aizicovici and N. H. Pavel, Differential Equations and Control Theory
G. Lyubeznik, Local Cohomology and Its Applications
G. Da Prato and L. Tubaro, Stochastic Partial Differential Equations and Applications
W. A. Carnielli et al., Paraconsistency
A. Benkirane and A. Touzani, Partial Differential Equations
A. Illanes et al., Continuum Theory
M. Fontana et al., Commutative Ring Theory and Applications
D. Mond and M. J. Saia, Real and Complex Singularities
V. Ancona and J. Vaillant, Hyperbolic Differential Operators and Related Problems
G. R. Goldstein et al., Evolution Equations
A. Giambruno et al., Polynomial Identities and Combinatorial Methods
A. Facchini et al., Rings, Modules, Algebras, and Abelian Groups
J. Bergen et al., Hopf Algebras
A. C. Krinik and R. J. Swift, Stochastic Processes and Functional Analysis: A Volume of
Recent Advances in Honor of M. M. Rao
S. Caenepeel and F. van Oystaeyen, Hopf Algebras in Noncommutative Geometry and Physics
J. Cagnol and J.-P. Zolésio, Control and Boundary Analysis
S. T. Chapman, Arithmetical Properties of Commutative Rings and Monoids
O. Imanuvilov, et al., Control Theory of Partial Differential Equations
April 18, 2005 9:24 3086 FM
Control Theory of
Partial Differential
Equations

Oleg Imanuvilov
Iowa State University
Ames, Iowa, USA

Guenter Leugering
University of Erlangen
Nuremberg, Germany

Roberto Triggiani
University of Virginia
Charlottesville, Virginia, USA

Bing-Yu Zhang
University of Cincinnati
Cincinnati, Ohio, USA

Boca Raton London New York Singapore

DK2961_Discl.fm Page 1 Thursday, April 21, 2005 10:16 AM

Published in 2005 by
Chapman & Hall/CRC
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2005 by Taylor & Francis Group, LLC

Chapman & Hall/CRC is an imprint of Taylor & Francis Group

No claim to original U.S. Government works

Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1

International Standard Book Number-10: 0-8247-2546-8 (Softcover)

International Standard Book Number-13: 978-0-8247-2546-4 (Softcover)

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with
permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish
reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials
or for the consequences of their use.

No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or
other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Catalog record is available from the Library of Congress

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
Taylor & Francis Group and the CRC Press Web site at
is the Academic Division of T&F Informa plc. http://www.crcpress.com
April 25, 2005 11:56 3086 FM

Preface

The present volume contains contributions by participants in the “Conference on Control Theory for
Partial Differential Equations,” which was held over a two-and-a-half day period, May 30 to June 1,
2003, at Georgetown University, Washington, D.C. The conference was dedicated to the occasion
of the retirement of Professor Jack Lagnese from the Mathematics Department of Georgetown
University.
It seemed most appropriate to honor the productive and successful scientific career of Jack Lagnese
by convening a conference that would bring together a select group of international specialists in the
theory of partial differential equations and their control. Over the years, many of the invitees have
enjoyed a personal and professional association with Jack. The lasting impact of Jack’s contributions
to control theory of partial differential equations and applied mathematics is well documented by over
80 research articles and three books. In addition, Jack served the scientific community for many years
in his capacity, at various times, as a program director in the Applied Mathematics Program within
the National Science Foundation, as an editor on the boards of several journals, as editor-in-chief of
the SIAM Journal on Control and Optimization, and as president of the SIAM Activity Group on
Control and Systems Theory. He was also a consultant to The National Institute for Standards and
Technology for a number of years.
Control theory for distributed parameter systems, and specifically for systems governed by partial
differential equations, has been a research field of its own for more than three decades. Although
having a distinctive identity and philosophy within the theory of dynamical systems, this field has
also contributed to the general theory of partial differential equations. Optimal interior and boundary
regularity of mixed problems, global uniqueness issues for over-determined problems and related
Carleman estimates, various types of a priori inequalities, and stability and long-time behavior are
just some examples of important developments in the theory of partial differential equations arising
from control theoretic considerations. In recent years, the field has broadened considerably as more
realistic models have been introduced and investigated in areas such as elasticity, thermoelasticity,
and aeroelasticity; in problems involving interactions between fluids and elastic structures; and in
other problems of fluid dynamics, to name but a few. These new models present fresh mathematical
challenges. For example, the mathematical foundations of fundamental theoretical issues have to be
developed, and conceptual insights that are useful to the designer and the practitioner need to be
provided. This process leads to novel numerical challenges that must also be addressed. The papers
contained in this volume provide a broad range of significant recent developments, new discoveries,
and mathematical tools in the field and further point to challenging open problems.
The conference was made possible through generous financial support by the National Science
Foundation and Georgetown University, whose sponsorship is greatly appreciated.
We wish to thank Marcel Dekker for agreeing to include this volume in its well-known and highly
regarded series “Lecture Notes in Pure and Applied Mathematics” and for its high professional
standards in handling this volume.
The Scientific Committee:
Oleg Imanuvilov
Guenter Leugering
Roberto Triggiani (Chair)
Bing-Yu Zhang

vii
April 25, 2005 11:56 3086 FM
April 25, 2005 11:56 3086 FM

Contributors

George Avalos Jack E. Lagnese

Department of Mathematics Department of Mathematics
University of Nebraska-Lincoln Georgetown University
Lincoln, Nebraska Washington, D.C.

Viorel Barbu Irena Lasiecka

Department of Mathematics Department of Mathematics
Alexandru Ioan Cuza University University of Virginia
Iasi, Romania Charlottesville, Virginia

Mikhail I. Belishev
Saint Petersburg Department Guenter Leugering
Steklov Institute of Mathematics Institute of Applied Mathematics
Saint Petersburg, Russia University of Erlangen-Nuremberg
Erlangen, Germany

Igor Chueshov
Department of Mathematics and Mechanics Wei Li
Kharkov University School of Mathematics
Kharkov, Ukraine Sichuan University
Chengdu, China
Michel C. Delfour
Department de Mathematiques et de Statistique Walter Littman
Universite de Montreal School of Mathematics
Montreal, Quebec University of Minnesota
Canada Minneapolis, Minnesota

Nicolas Doyon Zhuangyi Liu

Department de Mathematiques et de Statistique Department of Mathematics and Statistics
Universite Laval University of Minnesota, Duluth
Quebec City, Quebec Duluth, Minnesota
Canada
Michael Renardy
Matthias M. Eller Department of Mathematics
Department of Mathematics Virginia Polytechnic Institute and State University
Georgetown University Blacksburg, Virginia
Washington, D.C.
David L. Russell
Oleg Imanuvilov Department of Mathematics
Department of Mathematics Virginia Polytechnic Institute and State University
Iowa State University Blacksburg, Virginia
Ames, Iowa
E.J.P. Georg Schmidt
Victor Isakov Department of Mathematics and Statistics
Department of Mathematics and Statistics McGill University
Wichita State University Montreal, Quebec
Wichita, Kansas Canada

ix
April 25, 2005 11:56 3086 FM

x Contributors

Thomas I. Seidman Masahiro Yamamoto

Department of Mathematics and Statistics Department of Mathematical Sciences
University of Maryland University of Tokyo
Baltimore, Maryland Tokyo, Japan

Ralph E. Showalter Yiming Yang

Department of Mathematics Department of Mathematics and Physics
Oregon State University Beijing Technology and Business University
Corvallis, Oregon Beijing, China

Jiongmin Yong
Marianna A. Shubov Department of Mathematics
Department of Mathematics and Statistics Fudan University
Texas Tech University Shanghai, China
Lubbock, Texas

Bing-Yu Zhang
Jürgen Sprekels Department of Mathematical Sciences
Weierstrass Institute for Applied University of Cincinnati
Analysis and Stochastics Cincinnati, Ohio
Berlin, Germany
Xu Zhang
Dan Tiba School of Mathematics
Weierstrass Institute for Applied Analysis Sichuan University
and Stochastics Chengdu, China
Berlin, Germany
Jean-Paul Zolésio
Roberto Triggiani Centre National de Recherche Scientifique (CNRS) and
Department of Mathematics Institut National de Recherche en Informatique et en
University of Virginia Automatique (INRIA)
Charlottesville, Virginia Sophia Antipolis, France
April 25, 2005 11:56 3086 FM

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

1 Asymptotic Rates of Blowup for the Minimal Energy Function for the
Null Controllability of Thermoelastic Plates: The Free Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
George Avalos and Irena Lasiecka

2 Interior and Boundary Stabilization of Navier-Stokes Equations. . . . . . . . . . . . . . . . . . . . . . . 29

Viorel Barbu

3 On Approximating Properties of Solutions of the Heat Equation . . . . . . . . . . . . . . . . . . . . . . . 43

Mikhail I. Belishev

4 Kolmogorov’s ε-Entropy for a Class of Invariant Sets and Dimension

of Global Attractors for Second-Order Evolution Equations
with Nonlinear Damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Igor Chueshov and Irena Lasiecka

5 Extension of the Uniform Cusp Property in Shape Optimization . . . . . . . . . . . . . . . . . . . . . . . 71

Michel C. Delfour, Nicolas Doyon, and Jean-Paul Zolésio

6 Gårding’s Inequality on Manifolds with Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Matthias M. Eller

7 An Inverse Problem for the Dynamical Lame System with Two Sets
of Local Boundary Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Victor Isakov

8 On Singular Perturbations in Problems of Exact Controllability

of Second-Order Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Jack E. Lagnese

xi
April 25, 2005 11:56 3086 FM

xii Contents

9 Domain Decomposition in Optimal Control Problems for

Partial Differential Equations Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Guenter Leugering

10 Controllability of Parabolic and Hyperbolic Equations: Toward a Unified Theory . . . . 157

Wei Li and Xu Zhang

11 A Remark on Boundary Control on Manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

Walter Littman

12 Model Structure and Boundary Stabilization of an Axially

Moving Elastic Tape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Zhuangyi Liu and David L. Russell

13 Nonlinear Perturbations of Partially Controllable Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

Michael Renardy

14 On Junctions in a Network of Canals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

E.J.P. Georg Schmidt

15 On Uniform Null Controllability and Blowup Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

Thomas I. Seidman

16 Poroelastic Filtration Coupled to Stokes Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

Ralph E. Showalter

17 Operator-Valued Analytic Functions Generated

by Aircraft Wing Model (Subsonic Case) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Marianna A. Shubov

18 Optimal Design of Mechanical Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

Jürgen Sprekels and Dan Tiba

19 Global Exact Controllability on H 1Γ0 (Ω) × L2 (Ω) of Semilinear Wave Equations

with Neumann L2 (0,T;L2 (Γ1 ))-Boundary Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
Roberto Triggiani

20 Carleman Estimates for the Three-Dimensional Nonstationary Lamé

System and Application to an Inverse Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
Oleg Imanuvilov and Masahiro Yamamoto
April 25, 2005 11:56 3086 FM

Contents xiii

21 Forced Oscillations of a Damped Benjamin-Bona-Mahony

Equation in a Quarter Plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Yiming Yang and Bing-Yu Zhang

22 Exact Controllability of the Heat Equation with Hyperbolic Memory Kernel . . . . . . . . . 387
Jiongmin Yong and Xu Zhang
April 25, 2005 11:56 3086 FM
April 4, 2005 10:3 3086 DK2961˙C001

Chapter 1
Asymptotic Rates of Blowup for the Minimal Energy
Function for the Null Controllability of
Thermoelastic Plates: The Free Case
George Avalos1
University of Nebraska-Lincoln

Irena Lasiecka2
University of Virginia

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.2 Description of the PDE Model and Statement of the Problem . . . . . . . . . . . . . . . . . . . . . . 4
1.1.3 Main Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2 The Necessary Observability Inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3 Some Preliminary Machinery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4 A Singular Trace Estimate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5 Proof of Theorem 1.1(1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5.1 Estimating the Mechanical Velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5.2 Estimating the Mechanical Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.3 Conclusion of the Proof of Theorem 1.1(1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.6 Proof of Theorem 1.1(2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.6.1 A First Supporting Estimate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.6.2 Conclusion of the Proof of Theorem 1.1(2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.7 Proof of Theorem 1.1(3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Abstract Continuing the analysis undertaken in References 8 and 9, we consider the null-
controllability problem for thermoelastic plate partial differential equations (PDEs) models in the
absence of rotational inertia, defined on a two-dimensional domain , and subject to the free mechan-
ical boundary conditions of second and third order. It is now known that such uncontrolled systems
generate analytic semigroups on finite energy spaces. Consequently, the concept of null controllabil-
ity is indeed an appropriate question for consideration. It is shown that all finite energy states can be
driven to zero by means of L 2 [(0, T ) × ] controls in either the mechanical or thermal component.
However, the main intent of the paper is to quantify the singularity, as T ↓ 0, of the minimal energy
function relative to null controllability. In particular we shall show that in the case of one control
function acting upon the system, the singularity of minimal energy is optimal; it is in fact of order
O(T − 2 ), which is the same rate of blowup as that of any finite dimensional approximation of the
5

problem. The PDE estimates, which are obtained in the process of deriving this sharp numerology,
will have a strong bearing on regularity properties of related stochastic differential equations.
1 The research of George Avalos was partially supported by NSF DMS-0208121.
2 The research of Irena Lasiecka was partially supported by NSF DMS-01043 and ARO DAAD19-02-1-0179.

1
April 4, 2005 10:3 3086 DK2961˙C001

2 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

1.1 Introduction
In this chapter we address specific questions related to the null controllability of thermoelastic
plates subject to free mechanical boundary conditions, these being represented by shear forces and
moments. These particular boundary conditions are of particular interest in the control theory of
plates [22, 24, 23]. As we shall see below, the model under consideration is one which corresponds
to an infinite speed of propagation; accordingly, null controllability—in arbitrarily short time—is
an appropriate topic for study in regard to these plates. We will give at length a full and precise
description of our thermoelastic control problem; but for the benefit of the reader and in order to
motivate the specific problem under study, we will first provide a few opening remarks.

1.1.1 Motivation
There are several ways of controlling a given plate dynamic. This control can be accomplished by
using: 1. internal controls, 2. boundary controls, or 3. controls localized on an open subset of . In
addition, one may use either one control action (be it thermal or mechanical) or simultaneous me-
chanical and thermal controls (i.e., controls located on both the mechanical and thermal components
of the system). Depending on the objective to be achieved, one framework of control might be more
advantageous than another. For instance, if the particular issue at hand is to guarantee the minimal
support of control functions, then boundary control would be the most appropriate control situation.
However, if one is concerned with the cost of control—or equivalently, with quantifying the associ-
ated “minimal energy”—then internal controls should be considered. In this connection, a question
of both practical and mathematical relevance is the question of finding the optimal asymptotics that
describe the singularity of the associated minimal energy, as T ↓ 0. Since the work of T. Seidman in
Reference 34, the optimal asymptotics are well defined and well known for finite dimensional control
systems. In fact, these asymptotics are given by the sharp formula T −k− 2 where index k corresponds
1

to the Kalman rank condition and measures the defect of controllability (see below). The above for-
mula actually gives a lower bound for the singularity of the minimal energy associated with any PDE
system.
Given then the existence of formula in Reference 34 for controlled finite dimensional systems,
we are in a position to loosely define the “optimal” singularity for any controlled PDE. In fact,
for a given infinite dimensional system, the “optimal” rate of singularity of its associated minimal
energy will be the rate of singularity enjoyed by approximating (or truncated) finite dimensional
systems (assuming of course that each finite dimensional truncation has the same Kalman rank).
For example, scalar first order (in time) models will have its optimal rate of blowup of minimal
energy as being O(T − 2 ); in general, the optimal singularity for vectorial coupled structures will
3

depend on the number of controls used with respect to number of interactions. Thus, in the case
of thermal plates with one control only, the optimal singularity of any finite dimensional truncation
is T − 2 (this is seen below). In the case of two controls used (both thermal and mechanical) the
5

optimal singularity is T − 2 . Whether, however, the minimal energy asymptotics actually obeys the
3

optimal rate of singularity (predicted from finite dimensions) is an altogether different matter. Indeed,
in References 34 and 36 (highly nontrivial) finite-dimensional estimates are derived and can be
subsequently applied to finite-dimensional truncations of infinite-dimensional systems; however, the
delicate estimates are controlled by a constant Cn , say, where n stands for the dimensionality of the
respective approximation. These constants may well tend to infinity as n goes to infinity. In such an
event (as seen in References 14, 6, and 40) the optimal asymptotics for the original PDE are lost.
This brings us to the key question asked in this chapter: Is it possible to achieve the optimal rate of
singularity for a (fully infinite dimensional) controlled PDE model?
April 4, 2005 10:3 3086 DK2961˙C001

1.1 Introduction 3

The answer to the above question—in the negative—has been known for many years in the case of
the heat equation with either boundary or localized controls. Indeed, the rate for boundary control of
the heat equation is the exponential blowup rate eO( T ) ; see References 35 and 37. This rate is known
1

to be sharp [20]. A similar negative answer has been provided in the case of thermoelastic systems
under the influence of boundary controls—in fact, such boundary controls likewise lead to eO( T )
1

exponential blowup [25]. Therefore, in light of the rational rates of minimal energy blowup exhibited
by finite-dimensional controlled systems (as shown in Reference 34) and of the definition given above
for optimal rates of minimal energy blowup for controlled PDEs, it is manifest that thermoelastic
plates under the influence of boundary or localized controls will not give rise to minimal energies
that exhibit an optimal (finite dimensional) singularity. Thus, in searching for PDE control situations,
which will yield up the optimal algebraic singularity enjoyed by finite dimensional truncations, the
only reasonable choice left is the implementation of internal controls. In the specific context of
our thermoelastic PDE, the relevant question then becomes: Do the minimal energies of internally
controlled (fully infinite dimensional) thermoelastic plates exhibit the optimal rate of blowup O(T − 2 )
5

by either mechanical or thermal control?

The relevance of this question should not be underestimated from both a practical and mathematical
point of view. Indeed, from a practical point of view one would like to know whether a given
finite-dimensional approximation of the system contains critical information and moreover reflects
controllability properties of the original PDE model. From a mathematical point of view, the solution
to the null controllability problem is not only of interest in its own right as an issue in control theory,
but this solution can also give rise to deep and significant connections between the algebraic optimal
singularity of minimal energy and other fields of analysis, including stochastic analysis. In point
of fact, within the field of stochastic differential equations, there is an acute need to know of those
PDE control environments that will yield up optimal (and algebraic) rates of singularity of minimal
energy. These particular rates are critical in finding the regularity and solvability of certain stochastic
differential equations [14, 15, 19], as well as in setting conditions for the hypoellipticity of certain
degenerate infinite dimensional elliptic problems [32]. It is shown in Reference 32 that Hormander’s
hypoellipticity condition is strongly linked to the singularity of the minimal energy function. Null
controllability is also related to the analysis of regularity of the Bellman’s function, which is associated
with the minimal time control problem. Indeed, as eloquently described in References 14 and 15,
this property bears a close relation to the regularity of some Markov semigroups, including Orstein–
Uhlenbeck processes and related Kolmogorov equations. For some of these semigroups (see, e.g.,
Reference 15—Theorem 8.3.3) the minimal energy singularity associated with null controllability
describes differentiability properties and regularizing effects of the Orstein–Uhlenbeck process.
Moreover, the regularity of solutions to the Kolmogorov equation depends on the singularity of the
minimal energy as T ↓ 0. Also, as shown in Reference 14, optimal estimates for the norms of
controls are critical in being able to prove Liouville’s property for harmonic functions of Markov
processes (see p. 108 in Reference 15). In sum, there is an abundance of examples from the literature
that clearly illustrate that, in the context of computing optimal minimal energy asymptotics as T ↓ 0,
the tools of controllability can potentially enable a mathematical control theorist to transcend his or
her deterministic realm so as to solve fundamental problems in other areas of analysis, including
stochastic PDEs.
In addition, the procurement of optimal algebraic estimates for the minimal energy allows one to
clearly explain the role of the hyperbolic-parabolic coupling within the PDE structure (in Eq. (1.1)
below). In particular, it has been shown recently in Reference 25 that, owing to optimal algebraic
singularities of minimal energy, it is possible to offset the singularity of minimal energy by introducing
a very strong coupling within the system. Thus, in some sense, the lack of a second control in the
system may be quantitatively compensated for by taking large values of the coupling parameter “α.”
From our remarks above, it is clear that this compensatory phenomenon will not be observed with
boundary or partially supported controls, which, as we have said, lead to blowups of exponential type.
April 4, 2005 10:3 3086 DK2961˙C001

4 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

Having decribed the goal and motivation for the problem considered, we shall describe the main
contribution of this chapter within the context of recent work in that area. The problem of controlla-
bility/reachability for thermal plates has attracted considerable attention in recent years with many
contributions available in the literature [22, 23, 24, 1, 2, 3, 10, 18, 16, 17, 11], but we shall focus
particularly on works related directly to singular behavior, as T ↓ 0, of the minimal energy relative
to null controllability.
The study of optimal singularity for thermoelastic plates with internal controls started in Refer-
ences 8, 9, and 40, where for the first time the optimal rates T − 2 were established for the “commu-
5

tative” case (i.e., plates with hinged mechanical boundary conditions). The proof given in Reference
40 is based on a spectral method that exploits the commutativity in an essential way, whereas the
proof given in Reference 8 is based on weighted energy estimates, thereby giving one the chance
to extend this method to other noncommutative models (e.g., clamped or free mechanical bound-
ary conditions). The “commutative” case (hinged boundary conditions) has been also treated in
Reference 11, where null controllability with thermal controls of partially localized support was
proved. For this commutative model under boundary control (either thermal or boundary), the ex-
ponentially blowing up and sharp asymptotics eO( T ) have been shown [25]. The techniques used in
1

these papers rely critically on spectral analysis and commutativity.

It turns out that a proper and necessarily more technical extension of the method introduced in
Reference 8 will allow the consideration of noncommutative models. (By “noncommutative models”
we mean those models wherein the domains of the respective spatial differential operators of the
plate and heat dynamics do not necessarily enjoy any sort of compatibility.) In particular, the optimal
singularity of the (null control) minimal energy is proved in Reference 9 for clamped plates with one
control only. It should be stressed that the proof in the noncommutative case depends in an essential
manner upon estimates provided by the analyticity of the underlying thermoelastic semigroup; this
property of analyticity was discovered for the clamped case in References 31 and 28 and for free
case in Reference 27. The most challenging case is, of course, that of the free mechanical bound-
ary conditions (introduced in the context of control theory in Reference 22), in which a coupling
between thermal and mechanical variables also occurs on the boundary. This additional coupling
compels us to develop below a delicate string of trace estimates that measure the singularity at the
boundary.
The main aim of this chapter is to provide a complete analysis of the free case. We shall show
that in the case of mechanical control one still obtains the optimal singularity. Instead, in the case
of thermal control the estimate is “off” by 3/4. A question whether this estimate can be improved,
thereby leading to the optimal singularity T − 2 , still remains an open question.
5

1.1.2 Description of the PDE Model and Statement of the Problem

Having given our general remarks above, we now proceed to precisely describe the present prob-
lem under consideration; this work will continue and extend the analysis that has been previously
undertaken in References 6, 7, 8, 9 through and 40. We will consider throughout the two-dimensional
PDE system of thermoelasticity in the absence of rotational inertia. As we have already stated, it
is now known that for all possible mechanical boundary conditions, the thermoelastic PDE model
is associated with the generator of an analytic C0 -semigroup (see References 31, 28, 39, and 27).
Given then that the underlying PDE dynamics are “parabolic-like,” it is natural to consider the
null controllability problem for the thermoelastic system, namely, can one find L 2 (Q) controls
(mechanical or thermal) that steer the solution of the PDE from the initial data to the zero state?
(We shall make this control theoretic notion more precise below. As usual, Q here denotes the
cylinder  × (0, T ).) Having established L 2 (Q)-null controllability for the PDE, and moreover
assuming that the controllability time is arbitrary, we can subsequently proceed to measure the rate
of blowup, as T ↓ 0, for the minimal energy function that is associated to null controllability. As
is well known, and as we shall see below, this work is very much tied up with obtaining the sharp
April 4, 2005 10:3 3086 DK2961˙C001

1.1 Introduction 5

observability inequality associated with null controllability; moreover, this analysis is rather sensi-
tive to the mechanical boundary conditions imposed. In Reference 8—as well as in Reference 40
via a very different methodology—the problem of blowup for the minimal energy function was
undertaken in the canonical case of hinged mechanical boundary conditions; in Reference 9, we
revisit this problem for the more difficult clamped case. In this paper, we complete the picture
by analyzing the singularity of minimal energy for the case of the thermoelastic PDE under the
so-called free boundary conditions. In general, the analyses involved in the attainment of (null and
exact control) observability inequalities for thermoelastic systems are profoundly sensitive to the
particular set of boundary conditions are being imposed. But the free case, presently under consid-
eration, will give rise to the most problematic scenario of all. This situation is due to the high degree
of coupling between the mechanical and the thermal variables, with the coupling taking place in the
PDE itself and in the free mechanical boundary conditions.
We describe the problem in detail. Let  be a bounded open set of R2 , with smooth boundary . For
the free case, following [22, 23] the corresponding model PDE is as follows: the (mechanical) vari-
ables [ω(t, x), ωt (t, x)] and the (thermal) variable θ (t, x) solve, for given data {[ω0 , ω1 , θ0 ], u 1 , u 2 },
the PDE system


 ωtt + 2 ω + α θ = a1 u 1

 on (0, T ) × 

 θt − θ − α ω t = a 2 u 2







 

  ω + (1 − µ)B1 ω + αθ = 0


 ∂ ω ∂ B2 ω ∂θ on (0, T ) ×
 + (1 − µ) −ω+α =0 (1.1)

 ∂ν ∂τ ∂ν





 ∂θ

 + λθ = 0 on (0, T ) × , where λ > 0



 ∂ν





ω(t = 0) = ω0 ; ωt (t = 0) = ω1 ; θ (t = 0) = θ0 on .

Here, α > 0 is the parameter that couples the disparate dynamics (i.e., the heat equation vs. the
Euler plate equation); the constant µ ∈ (0, 1) is Poisson’s ratio. Also, the (control) parameters a1
and a2 satisfy a1 ≥ 0, a2 ≥ 0 and a1 + a2 > 0 (in other words, at least one of the controls, be it
thermal or mechanical, is always present.) The (free) boundary operators Bi are given by

∂ 2w ∂ 2w ∂ 2w
B1 w ≡ 2ν1 ν2 − ν12 2 − ν22 2 ;
∂ x∂ y ∂y ∂x
 2 (1.2)
 2  ∂ 2
w ∂ w ∂ 2w
B2 w ≡ ν1 − ν22 + ν 1 ν2 − .
∂ x∂ y ∂y 2 ∂x2
The PDE Eq. (1.1) is the model explicitly derived and analyzed in References 24 and 22 in the
“limit case.” That is to say, we are considering the two-dimensional thermoelastic system in the
absence of rotational forces; the small and nonnegative, classical parameter γ is taken here to be
zero. As we stated at the outset, it is now well known that the lack of rotational inertia in the model
Eq. (1.1) will result in the corresponding dynamics having their evolution described by the generator
of an analytic semigroup on the associated basic space of finite energy. In short, the present case
γ = 0 corresponds to parabolic-like dynamics; this is in stark contrast to the case γ > 0—as analyzed
in the control papers [22], [23], [3] and myriad others—for which the corresponding PDE manifests
hyperbolic-like dynamics.
In fact, if we define

H ≡ H 2 () × L 2 () × L 2 (), (1.3)

April 4, 2005 10:3 3086 DK2961˙C001

6 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

then one can proceed to show by the Lumer Phillips theorem that the thermoelastic plate model can
be associated with the generator of a C0 -semigroup of contractions on H. That is to say, there exists
A : D(A) ⊂ H → H, and {eAt }t≥0 ⊂ L(H) such that [ω, ωt , θ ] satisfies the PDE (1.1) if and only
if [ω, ωt , θ ] satisfies the abstract ODE
         
ω(t) ω(t) 0 ω(0) ω0
d         
 ωt (t)  = A ωt (t)  +  a1 u 1 (t) ;  ωt (0)  =  ω1  .
dt
θ (t) θ (t) a2 u 2 (t) θ (0) θ0

In consequence of this relation, we have immediately from classical semigroup theory that

{[ω0 , ω1 , θ0 ], [u 1 , u 2 ]} ∈ H × [L 2 (Q)]2 ⇒ [ω, ωt , θ ] ∈ C([0, T ]; H). (1.4)

Because of the underlying analyticity, which will ultimately mean that there are smoothing effects
associated with the application of the semigroup {eAt }t≥0 , the null controllability problem for the
controlled PDE Eq. (1.1)—with respect to internal L 2 -controls—is an appropriate one to study.
Moreover, one might speculate that, as in the case of the canonical heat equation [12], should the
PDE Eq. (1.1) in fact be null controllable, it will be so in arbitrary small time (because of the
underlying infinite speed of propagation). It is this speculation that motivates our working definition
of null controllability for the present paper.

DEFINITION 1.1 The PDE (1.1) is said to be null controllable if, for any T > 0 and arbitrary
initial data x ≡ [ω0 , ω1 , θ0 ] ∈ H, there exists a control function [u 1 , u 2 ] ∈ [L 2 (Q)]2 such that the
corresponding solution [ω, ωt , θ] ∈ C([0, T ]; H) satisfies [ω(T ), ωt (T ), θ (T )] = [0, 0, 0].

However, the issue of null controllability, although certainly an important part of this paper, is
subordinate to our main objective, which is to measure the rate of singularity of the associated
minimal energy function.
We develop this notion of “minimal energy.” Assume for the time being that the Eq. (1.1) is
null controllable within the class of [L 2 (Q)]2 -controls, in the sense of the Definition 1.1. Subse-
quently, one can then speak of the associated minimal norm control, relative to given initial data
x ≡ [ω0 , ω1 , θ0 ] ∈ H and given terminal time T . That is to say, we can consider the problem of finding
a control u0T (x) that steers the solution [ω, ωt , θ ] of Eq. (1.1) (with [u 1 , u 2 ] = u0T (x) therein) from
initial data x to zero in arbitrary time T and minimizes the L 2 norm. In fact, by standard convex
optimization arguments (see, e.g., Reference 13), given any x ∈ H and fixed T , one can find a control
u0T (x) which solves the problem
 0 
u (x) 2 2 = min u[L 2 (Q)]2 ,
T [L (Q)]

where, above, the minimum is taken with respect to all possible null controllers u = [u 1 , u 2 ] ∈
[L 2 (Q)]2 of the PDE (1.1) (which steer initial data x to rest at time t = T ). Subsequently, we can
define the minimal energy function Emin (T ) as
 
Emin (T ) ≡ sup u0T (x)[L 2 (Q)]2 . (1.5)
xH =1

Under the assumption of null controllability, as defined in Definition 1.1, we have that Emin (T ) is
bounded away from zero. A natural follow-up question is “how does Emin (T ) behave as terminal time
T ↓ 0, or equivalently (by Eq. (1.5)), for given time T , how exactly does the quantity u0T (x)[L 2 (Q)]2
grow as T ↓ 0?”
April 4, 2005 10:3 3086 DK2961˙C001

1.1 Introduction 7

The problem of studying the rate of blowup for minimal norm controls is a classical one and has
its origins from the finite dimensional setting. In fact, a very complete and satisfactory solution has
been given in Reference 34 for the following controlled ODE in Rn :

d
y (t) = Ay (t) + B u (t), y0 ∈ Rn (1.6)
dt

where u ∈ L 2 (0, T ; Rm ) and A (resp., B) is an n × n (resp. n × m) matrix, with m ≤ n (so

consequently the solution y ∈ C([0, T ]; Rn ). The problem in this finite dimensional milieu, like that
for our controlled PDE (1.1), is to ascertain the rate of singularity for the associated minimal energy
function, which is defined in the same way as in Eq. (1.5). The solution to this problem is tied up
with the classical Kalman’s rank condition. Namely, a beautifully simple (though highly nontrivial)
formula in Reference 34—an alternative constructive proof of this formula is given in Reference 40;
see also Reference 36—yields that the minimal energy function associated to the null controllability
of Eq. (1.6) is O(T −k− 2 ), where k is the Kalman’s rank of the system Eq. (1.6) (that is, k is the
1

smallest integer such that rank ([B, AB, . . . , Ak B]) = n; see Reference 41).
By a formal application of Seidman’s finite dimensional result, one can get an inkling of the
numerology involved in the computation of the minimal energy Emin (T ) for the PDE system Eq. (1.1).
For example, let us consider the thermoelastic Eq. (1.1) but with now ω satisfying the canonical hinged
mechanical/Dirichlet thermal boundary conditions

ω| = ω| = θ| = 0 on . (1.7)

In this case, it is shown in Reference 27 that when, say, thermal control only is implemented (i.e.,
a1 = 0 in Eq. (1.1)), the thermoelastic PDE under the hinged boundary conditions Eq. (1.7) may be
associated with the ordinary differential equation (ODE) (1.6), with
   
0 1 0 0
   
A =  −1 0 1 , and B =  0 . (1.8)
0 −1 −1 a2

This ODE in three space dimensions is a direct consequence of the analysis undertaken for the
canonical hinged case in Reference 26. By way of obtaining the ODE (1.6), we have formally
“factored out” the Laplacian from the (rearranged) infinitesimal generator of the thermoelastic
semigroup, which is given in (Section 1.2.2) of Reference 27 (see also Reference 28, p. 311).
Considering now finite dimensional truncations of (by making use of the spectral resolution of
the Laplacian under Dirichlet boundary conditions) and applying the algorithm of Seidman to the
given controllability pair [A, B] in Eq. (1.8), we compute readily that the minimal energy func-
tion associated with the null controllability of the finite dimensional Eq. (1.6)—an approximation
in some sense of the thermoelastic system under the hinged boundary conditions—blows up at a
rate on the order of T − 2 . These numerics lead to the following question: Does the minimal en-
5

ergy Eq. (1.5) (i.e., the minimal energy for the full-fledged infinite dimensional system) obey the
law Emin (T ) = O(T − 2 )? Of course, Seidman’s formula for matrices gives no conclusive proof as
5

to what is actually happening for the fully infinite dimensional model. In fact, it is well known
that the minimal energy of a given infinite dimensional system may bear no relation to the limit
of minimal energies of any given sequence of finite dimensional approximations. For example, it
was shown in Reference 14 that the growth of the minimal energy function for a given infinite
dimensional system may be arbitrarily large, even when Kalman’s rank k = 1 and spectral diag-
onal systems are being considered. Moreover, in Reference 35 it is shown that for the case of the
boundary controlled heat equation, the sharp observability inequality corresponding to the (null)
minimal energy of a given heat operator’s finite dimensional truncation obeys rational rates of
April 4, 2005 10:3 3086 DK2961˙C001

8 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

singularity. On the other hand, the asymptotics of the minimal energy, which are obtained for the
(infinite dimensional) heat equation, are of exponential type. A similar phenomenon is observed in
References 40 and 6, wherein strongly damped wave equations under internal control are considered.
In this situation,
β
with the damping operator given by Aβ , the asymptotics of minimal energy be-
− 2(1−β)
have as T for any β > 34 . Thus, when the damping operator approaches Kelvin’s Voight
β
damping, the singularity loses its algebraic character with 2(1−β) ↑ ∞. Instead, for β ≤ 34 , the
singularity is optimal (i.e., the same as that for finite dimensional truncations) and is equal
to T − 2 .
3

But as formal as the application of Seidman’s finite dimensional algorithm may seem in the present
context, there is in fact a relevance here to the thermoelastic PDE, which is approximately described
by controllability pair [A, B]. The minimal energy function with respect to null controllability of the
thermoelastic PDE, under the hinged boundary conditions Eq. (1.7), does indeed obey the singular
rate Emin (T ) = O(T − 2 ). This minimal energy analysis for the hinged case was shown independently
5

in References 8 and 40 (and most recently in Reference 25 where the asymptotics with respect to the
coupling α are also provided). In Reference 40, it is of prime importance that the hinged boundary
conditions Eq. (1.7) be in play, for these mechanical boundary conditions allow a fortuitous spectral
resolution of the underlying thermoelastic generator. With the eigenfunctions of the thermoelastic
dynamics in hand, it is shown in Reference 40 via a constructive class of suboptimal steering controls
that the delicate observability estimates for solutions for the spectrally truncated adjoint problem—
adjoint with respect to null controllability—are preserved; as a consequence, a rational rate of
singularity for the infinite dimensional null minimal energy is obtained in the limit. However, for
other sets of mechanical boundary conditions, including the physically relevant clamped and (above
all) free boundary conditions under consideration at present, there will be no such available spectral
decomposition.
On the other hand, the methodology employed in References 8 and 9, and the present work, is
“eigenfunction independent”; in particular, we blend a weighted multiplier method of Carleman’s
type with boundary trace estimates exhibiting singular behavior of the boundary traces. This rather
special behavior is a consequence of the underlying analyticity. In principle, our work in Reference 8
to estimate the blowup of the “minimal norm control” as T ↓ 0 is applicable to a variety of
dynamics. (In fact, our method of proof in Reference 8 and in the present work is used in
Reference 7 to estimate the minimal norm control of the abstract wave equation under Kelvin–
Voight damping.) Moreover, the robustness of our method allows us in Reference 9 to analyze
the rate of singularity of the minimal energy function for the null controllability of thermoelastic
plates in the case of clamped boundary conditions. As we said above, there is no spectral de-
composition or factorization of the thermoelastic generator in the case of mechanical boundary
conditions other than the canonical hinged case and thus no rigorous association with the abstract
ODE (1.8). Still, we show in Reference 9 that for the clamped case, the minimal energy obeys the
singular rate “predicted” in Reference 34, namely, Emin (T ) = O(T − 2 ). Our intent in this paper
5

is to bring the story to a close by investigating the minimal energy function for the null control-
lability of thermoelastic systems under the high-order free boundary conditions that are present in
Eq. (1.1).

1.1.3 Main Result

In regards to our stated problem, the main result is as follows:

THEOREM 1.1
Let terminal time T > 0 be arbitrary and a1 , a2 ≥ 0 with a1 + a2 > 0. Then, given initial data
[ω0 , ω1 , θ0 ] ∈ H, there exist control(s) [u 1 , u 2 ] ∈ [L 2 (Q)]2 such that the corresponding solution
[ω, ωt , θ ] of (1.1) satisfies [ω(T ), ωt (T ), θ(T )] = [0, 0, 0]. (That is to say, the PDE model Eq. (1.1)
April 4, 2005 10:3 3086 DK2961˙C001

1.2 The Necessary Observability Inequality 9

is null controllable within the class of [L 2 (Q)]2 —controls in arbitrary short time.) Moreover, We
have the following rates of blowup for the minimal energy function:
1. (thermal control) If a1 = 0, then Emin (T ) = O(T − 4 − ) for all  > 0;
13

2. (mechanical control) If a2 = 0, then Emin (T ) = O(T − 2 );

5

3. If a1 > 0 and a2 > 0, then Emin (T ) = O(T − 2 ).

3

REMARK 1.1 The null controllability of thermoelastic plates with free boundary conditions and
under one internal control (be it mechanical or thermal) appears to be, as far as we know, a new
result in the literature. The Theorem 1.1 above, in addition to asserting the said null controllability
property, provides the asymptotics for the singularity of the associated minimal energy function.
These asymptotics are optimal in the case of a single mechanical control and in the case of two
controls acting upon the system. In the case of a single thermal control the estimate is “off” by 3/4
with respect to the desired “finite dimensional prediction” in Reference 34. Whether this estimate
can be improved upon is an open question.

Our method of proof of Theorem 1.1 is based on weighted energy estimates that are flexible
enough to accomodate analytic estimates and the resulting singularity. The proof has the following
main technical ingredients:
1. special weighted nonlocal multipliers introduced in Reference 4 and subsequently invoked in
References 3, 5, 29, and 6, and elsewhere;

2. the analyticity of semigroups associated with thermoelastic PDE models in the absence of
rotational forces, as demonstrated in References 31, 26, 27, and 28;

3. new singular estimates for boundary traces of solutions of Eq. (1.9), which are of their own
intrinsic interest and which are needed to handle the boundary terms resulting from the
weighted estimates employed.

1.2 The Necessary Observability Inequality

The proof of Theorem 1.1 is based on the derivation of the observability inequality associated
with the null controllability of the PDE (1.1) with respect to thermal or mechanical control or both.
This inequality is formulated in terms of the solution of the homogeneous PDE, which is “dual”
or “adjoint” to that in Eq. (1.1). Namely, we shall consider solutions [φ, φt , ϑ] to the following
system:


 φtt + 2 φ + α ϑ = 0 on (0, T ) × 





 ϑt − ϑ − α φt = 0 on (0, T ) × 



 




  φ + (1 − µ)B1 φ + αϑ = 0
 on 
 ∂ φ + (1 − µ) ∂ B2 φ − φ + α ∂ϑ = 0 (1.9)

 ∂ν ∂τ ∂ν





 ∂ϑ

 + λϑ = 0 on , λ > 0



 ∂ν




[φ(0), −φt (0), ϑ(0)] = [φ0 , φ1 , ϑ0 ] ∈ H.
April 4, 2005 10:3 3086 DK2961˙C001

10 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

If we define the bilinear form a(·, ·) : H 2 () × H 2 () → R by


a(w, w̃) ≡ [w x x w̃ x x + w yy w̃ yy + µ(w x x w̃ yy + w yy w̃ x x ) + 2(1 − µ)w x y w̃ x y ] d,


then we can state the “Green’s formula,” which involves this bilinear form (see Reference 22) and
which is valid for functions w, w̃ (smooth enough):
   
∂ w ∂ B2 w
( 2 w)w̃d = a(w, w̃) + + (1 − µ) w̃d
 ∂ν ∂τ

∂ w̃
− [ w + (1 − µ)B1 w] d . (1.10)
∂ν
Let E(t) denote the energy of the adjoint system Eq. (1.9), where
 
1 1 1
E(t) ≡ a(φ(t), φ(t))d + |φ(t)|2 d + |ϑ(t)|2 d. (1.11)
2 2 2 
In terms of this energy, then one can show by classical functional analytical arguments (see, e.g.,
References 41 and 3) that the PDE (1.1) is null controllable, in the sense of 1, if and only if the
adjoint variables [φ, φt , ϑ] of Eq. (1.9) satisfy the following continuous observability inequality, for
some constant C T :
[φ(T ), φt (T ), ϑ(T )]H ≤ C T (a1 φt  L 2 (Q) + a2 ϑ L 2 (Q) ). (1.12)
Having worked to establish the sharp constant C T in the observability inequality Eq. (1.12), one
can proceed through an algorithmic procedure—using an explicit representation of the minimal norm
control, by convex optimization—so as to have that for all terminal time T > 0,
Emin (T ) = O(C T ).
Because the details of this argument are known and have been previously spelled out (see, e.g.,
References 9 and 8), we defer from repeating them here.
Because of this characterization of the behavior of Emin (T ) with the constant C T in Eq. (1.12), our
work will accordingly be geared toward establishing this inequality (where, again, control parameters
ai satisfy a1 , a2 ≥ 0, and a1 + a2 > 0).

1.3 Some Preliminary Machinery

In this section, we explicitly define the underlying generator A : D(A) ⊂ H → H, which describes
the thermoelastic flow. Subsequently, a proposition is derived with which to associate powers of this
generator with specific Sobolev spaces. This characterization of the powers will be critical in work.
r To start, we define the linear operator A D : D(A D ) ⊂ L 2 () → L 2 () by
AD ≡ − ;
(1.13)
D(A D ) = H 2 () ∩ H01 ().
r We will also need the following (Dirichlet) map D : L 2 ( ) → L 2 ():
Df = g ⇔ g = 0 on  and g| = f on . (1.14)
s+ 12
By the classical elliptic regularity, we have that D ∈ L(H s ( ), H ()) for all s (see
Reference 30).
April 4, 2005 10:3 3086 DK2961˙C001

1.3 Some Preliminary Machinery 11

r We also define the linear operator Å : D(A D ) ⊂ L 2 () → L 2 () by setting Å = 2
,
for  ∈ D(Å), where

D(Å) =  ∈ H 4 () : [  + (1 − µ)B1  ] = 0
  
∂  ∂ B2 
and + (1 − µ) − =0 ,
∂ν ∂τ

where the boundary operators Bi are as defined in Eq. 1.3.

This operator is densely defined, positive definite, and self-adjoint. Consequently by
Reference 21, one has the characterization

 1 2
D(Å 2 ) ≈ H 2 (); with moreover Å 2 φ  L 2 () = a(φ, φ) +
1
φ2d .

r Moreover, we define the elliptic operators G i by




2
v=0 on 



G1h = v ⇔  v + (1 − µ)B1 v = h ;



 ∂ v ∂ B2 v on
 + (1 − µ) −v =0
∂ν ∂τ



2
v = 0 on 



G2h = v ⇔  v + (1 − µ)B1 v = 0 . (1.15)



 ∂ v ∂ B2 v on
 + (1 − µ) −v =h
∂ν ∂τ

By elliptic regularity (see, e.g., Reference 30) one has that for all real s,
5 7
G 1 ∈ L(H s ( ), H s+ 2 ()); G 2 ∈ L(H s ( ), H s+ 2 ()). (1.16)

With these operators defined above, we have that the generator A : D(A) ⊂ H → H of the
thermoelastic semigroup may be given the explicit representation
 
0 I 0
 
A =  −Å 0 α(A D (I − Dγ0 ) − ÅG 1 γ0 + λÅG 2 γ0 )  ;
0 −α A D (I − Dγ0 ) −α A D (I − Dγ0 )

D(A) = [ω0 , ω1 , θ0 ] ∈ H 2 () × H 2 () × H 2 () : Å [ω0 + α (G 1 γ0 − λG 2 γ0 ) θ0 ] ∈ L 2 ()
  
∂θ0
and + λθ0 = 0 (1.17)
∂ν

(here, γ0 ∈ L(H 1 (), H 2 ( )) is the classical Sobolev trace map; i.e., γ0 f = f | for f ∈ C ∞ ()).
1

As we have said, it is now known that the generator A : D(A) ⊂ H → H for the thermoelas-
tic plate, with free mechanical boundary conditions, is associated with an analytic C0 -semigroup
{eAt }t≥0 of contractions on H (see Reference 27 and references therein), with moreover A−1 being
bounded on H.
April 4, 2005 10:3 3086 DK2961˙C001

12 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

For this realization of the generator, we now proceed to show the following:

PROPOSITION 1.1
Let integer k = 1, 2, . . . . Then D(Ak ) ⊂ H 2k+2 () × H 2k () × H 2k ().

PROOF OF PROPOSITION 1.1 Let first [ω0 , ω1 , θ0 ] ∈ D(A). Then by definition, ω1 , θ0 ∈

H 2 (). Moreover, from the abstract representation in Eq. (1.17), we have

Åω0 + α ÅG 1 θ0 | − αλÅG 2 θ0 | + α θ0 = f ∈ L 2 ().

3
Because θ0 | ∈ H 2 ( ), then consequently from the elliptic regularity results posted in Eq. (1.16),

ω0 = Å−1 f − αG 1 θ0 | + αλG 2 θ0 | − α Å−1 θ0 ∈ H 4 ().

So the assertion is true for k = 1.

Proceeding now by induction, suppose that the result holds true for integer k − 1, k ≥ 2, and let
[ω0 , ω1 , θ0 ] ∈ D(Ak ). Then, because
 
ω0
 
A ω1  ∈ D(Ak−1 ),
θ0

we have

ω1 ∈ H 2k ();
Åω0 + α ÅG 1 θ0 | − αλÅG 2 θ0 | + α θ0 = f ∈ H 2k−2 ();
A R θ0 − α ω1 = g ∈ H 2k−2 (). (1.18)

Here, A R : D(A R ) ⊂ L 2 () → L 2 () is the elliptic operator defined by

 
∂f
AR f = − f ; D(A R ) = f ∈ H 2 () : + λ f = 0, λ > 0 . (1.19)
∂ν

Reading off the third equation in Eq. (1.18), we obtain, after using elliptic regularity,

θ0 = A−1
R (g + Dγ0 θ0 − α ω1 ) ∈ H ().
2k

In turn, we can use again the result in Eq. (1.16) to have that

ω0 = Å−1 f − αG 1 γ0 θ0 + αλG 2 γ0 θ0 − α Å−1 θ0 ∈ H 2k+2 ().

This concludes the proof of Proposition 1.1.

1.4 A Singular Trace Estimate

In this section, we exploit the underlying analyticity of the thermoelastic semigroup so as to
generate pointwise (in time) estimate of boundary traces of the adjoint variables φt (t) and ϑ(t)
of Eq. (1.9). These estimates will be of use to us in the proof of Theorem 1.1, inasmuch as they
April 4, 2005 10:3 3086 DK2961˙C001

1.4 A Singular Trace Estimate 13

each reflect a proper “distribution” between the measurement E(t) of the energy and the observation
term—be it φt or ϑ. The price to pay for these benefical estimates is the appearance therein of singular
weights of the form t1s , where parameter s will depend on the order of derivatives present.

LEMMA 1.1
Let x (t) ≡ [φ(t), φt (t), ϑ(t)] denote the solution of the adjoint system Eq. (1.9), subject to the initial
condition x (0) = [φ0 , −φ1 , ϑ0 ] ∈ H. Let,moreover, Dm be a differential operator of order m ≥ 0
with respect to the interior variables. Then for integers k = 1, 2, . . . , and all t > 0 we have
Ck  t  1k 1− 1k
1. Dm ϑ(t) L( ) ≤ m 1 eA 2 x0 H2 ϑ(t) L 2 () 2
;
+
t2 4
Ck  1 1
eA 2t x0  2k φt (t)1−2 2k ;
2. Dm φt (t) L( ) ≤ H L ()
2 +4
m 1
t
Ck  
eA 2t x0  .
3. D1 φtt (t) L( ) ≤ 7 H
t4

PROOF OF LEMMA 1.1 By a trace interpolation result (see, e.g., Reference 38) and the iterative
use of a classical PDE moment inequality, we have the following string of estimates, which is valid
for any g ∈ H 2 (m+1) ():
k+1

1 1 1 1
Dm g L( ) ≤ C Dm g L()
2
Dm g H2 1 () ≤ C g H2 m () g H2 m+1 ()
1 1 1 3 1 1
≤ C g L()
2
g H4 2m () g H4 2(m+1) () ≤ C g L()
4
g H8 4m () g H8 4(m+1) ()
1 1 1
1−
≤ . . . ≤ C g L() g H2 2k m () g H2 2k (m+1) () .
2 k k+1 k+1
(1.20)

Now by virtue of the analyticity of the thermoelastic semigroup {eAt }t≥0 and Proposition 1.1, we
have for all t > 0,
k−1
[φ(t), φt (t), ϑ(t)] ∈ D(A2 m
) ⇒ [φt (t), ϑ(t)] ∈ [H 2km ()]2 . (1.21)

Setting now g ≡ ϑ(t) (resp., φt (t)) in Eq. (1.20), we obtain

1 1 1
1−
Dm ϑ(t) L( ≤ C ϑ(t) L() ϑ(t) H2 2k m () ϑ(t) H2 2k (m+1) ()
2 k k+1 k+1
)
 k−1  k+1
1   k+1 1  1− 21k
≤ C A2 m x (t)H2 A2 (m+1) x (t)H2 ϑ(t) L()
k−1

 k−1  k+1
1   k+1
1  1− 21k
= C A2 m eA 2 eA 2 x0 H2 A2 (m+1) eA 2 eA 2 x0 H2 ϑ(t) L()
t t k−1 t t
. (1.22)

At this point, we can invoke the well known pointwise estimate that is valid for any generator of
an analytic semigroup: for all time t > 0 and integer m = 1, 2, . . . ,
Cm
Am eAt L(H) ≤ , (1.23)
tm
where constant C is independent of m (see, e.g., Reference 33, p. 70). Applying this estimate to the
chain Eq. (1.22), we have
C  1 1
eA 2t x0  2k ϑ(t)1− 2k .
Dm ϑ(t) L( ) ≤ H L()
2 +4
m 1
t
April 4, 2005 10:3 3086 DK2961˙C001

14 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

This gives (Lemma 1.1, Step 1) (Step 2) is obtained in the very same way, by setting g = φt in
Eq. (1.22) and then invoking the containment Eq. (1.21). For (Step 3), we have along the same lines,
by means of the trace interpolation inequality in Reference 38 and the containment Eq. (1.21),
1 1  1 1 1
D1 φtt (t) L 2 ( ) ≤ C φtt (t) H2 1 () φtt (t) H2 2 () ≤ C A 2 xt (t)H2 Axt (t)H2
1 C  t 1
≤ CA2 x (t)H2 ≤ 7 eA 2 x0 H2 ,
t4
which completes the proof.

1.5 Proof of Theorem 1.1(1)

1.5.1 Estimating the Mechanical Velocity
In what follows, we will have need of the polynomial weight h(t), defined by
h(t) ≡ t s (T − t)s . (1.24)
For the proof of Theorem 1.1(1), we will take s ≡ 6.
In terms of the the solution [φ, φt , ϑ] of Eq. (1.9) and its corresponding energy E(t), the necessary
inequality for the case of thermal control is

E(T ) ≤ C T ϑ L 2 (Q) . (1.25)
It is the derivation of this inequality that will drive the proof of Theorem 1.1.
We will start by applying the Laplacian to both sides of the heat equation in Eq. (1.9). This gives
ϑt − 2
ϑ −α 2
φt = 0 in .
From this expression and the free boundary conditions in Eq. (1.9), we have that the velocity term
φt satisfies the following elliptic problem for all t > 0:


 2
φt (t) =
1
ϑt (t) −
1 2
ϑ(t) in 



 α α

 (1.26)

  φt (t) + (1 − µ)B1 φt (t) = −αϑt

 on .

 ∂ φt (t) ∂ B2 φt (t)
 + (1 − µ) − φt (t) = αλϑt
∂ν ∂τ
Using this Green’s map defined in Eq. (1.15), we have from Eq. (1.26) that the velocity φt may
be written explicitly as
1 −1  
φt (t) = Å ϑt (t) − 2
ϑ(t) − αG 1 γ0 (ϑt (t)) + αλG 2 γ0 (ϑt (t)). (1.27)
α
From this, we have
 T
h φt 2L 2 () dt
0
 T 
1 −1
= h Å [ ϑt (t) − 2
ϑ(t)] − αG 1 γ0 (ϑt (t)) + αλG 2 γ0 (ϑt (t)), φt dt, (1.28)
0 α L 2 ()

where h(t) is the polynomial weight described in Eq. (1.24).

April 4, 2005 10:3 3086 DK2961˙C001

1.5 Proof of Theorem 1.1(1) 15

Analysis of the right-hand side of Eq. (1.28).

1.
 T  T
h ([G 1 − λG 2 ] γ0 ϑt , φt ) L 2 () dt = − h  ([G 1 − λG 2 ] γ0 ϑ, φt ) L 2 () dt
0 0
 T
− h([G 1 − λG 2 ] γ0 ϑ, φtt ) L 2 () dt. (1.29)
0

a. By the regularity posted in Eq. (1.16) and an application of Lemma 1.1 (with m = 0 and
k = 2, say) we have
  
 T  T
 h  ([G 1 − λG 2 ] γ0 ϑ, φt ) L 2 () dt  ≤ C |h  | ϑ L 2 ( ) φt  L 2 () dt

0 0
  5
T
|h  | h(t) 8 3  t 5
≤C 1 ϑ(t) L4 2 () eA 2 x0 H4 dt.
0 t 4 h(t)

Invoking Hölder’s inequality to this right hand side, with Hölder conjugates ( 83 , 85 ), we
obtain now the estimate
 T 
 
 h 
([G − λG ] γ ϑ, φ ) dt 
 1 2 0 t L ()
2

0
 T  T 
26 t
≤ C T 3 h(t) ϑ L 2 () dt + 
2
h(t)E dt. (1.30)
0 0 2

b. Proceeding as above, with m = 0 and k = 2 in Lemma 1.1, we have

 
 T 
 h([G 1 − λG 2 ] γ0 ϑ, φtt ) L 2 () dt 

0
 T
h(t) 3  t 1
≤C 1 ϑ(t) L4 2 () eA 2 x0 H4 Axt (t) L 2 () dt
0 t4
 T
h(t) 3  t 5
≤C 5 ϑ(t) L4 2 () eA 2 x0 H4 dt
0 t4
 T  T 
26 t
≤ C T 3 h(t) ϑ2L 2 () dt +  h(t)E dt. (1.31)
0 0 2

Combining Eq. (1.30) and Eq. (1.31) now gives

    
 T  T T
t
 
h([G 1 − λG 2 ] γ0 ϑt , φt ) L 2 () dt  ≤ C T
20
ϑ2L 2 () dt +  h(t)E dt.

3

0 0 0 2
(1.32)

2. Next,
 
T   T  
h Å−1 ( ϑt ), φt L 2 () dt = − h  Å−1 ( ϑ), φt L 2 () dt
0 0
 T  
− h Å−1 ( ϑ), φtt L 2 () dt. (1.33)
0
April 4, 2005 10:3 3086 DK2961˙C001

16 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

a. By Green’s theorem and the Lemma 1.1, with m = 0 and k = 2, we have

 
 T   
 h Å ( ϑ), φt L 2 () dt 
 −1

0
 
 T  ∂  T   
 
= h  ϑ, + I Å−1 φt dt − h ϑ, −1
Å φt L 2 () dt 
 0 ∂ν L2( ) 0 
 T  T 
26 t
≤ C T 3 ϑ2L 2 () dt +  h(t)E dt. (1.34)
0 0 2

b. Likewise, by Green’s Theorem, the analyticity of the semigroup and Lemma 1.1, with
m = 0 and k = 2, we have
    
 T    T T
t
 h(t) −1
ϑ, Å φtt dt  ≤ C T
26
ϑ2L 2 () dt +  h(t)E dt.
 L 2 ()   3

0 0 0 2
(1.35)

Applying the estimates Eq. (1.34) and Eq. (1.35) to Eq. (1.33) now yields

    
 T  T T
t
 
h( ϑt , φt ) L 2 () dt  ≤ C T
26
ϑ2L 2 () dt +  h(t)E dt. (1.36)

3

0 0 0 2

3. By the Green’s identity posted in Eq. (1.10), we have

 
T  −1
 T
h 2
ϑ, Å φt L 2 ()
dt = h(t)(ϑ, φt ) L 2 () dt
0
 T
0
 
∂ ∂ B2
= h(t) + (1 − µ) ϑ, Å−1 φt
0 ∂ν ∂τ L2( )
 
∂ −1
− [ + (1 − µ)B1 ] ϑ, Å φt dt.
∂ν L2( )

Applying once more the Lemma 1.1 (e.g., with m = 3, k = 3) we have

    
 T    T T
t
 h 2
ϑ, Å−1 φt dt  ≤ C T 8
ϑ2L 2 () dt +  h(t)E dt. (1.37)
 L 2 ()  
2
0 0 0

Combining the expression Eq. (1.28) with the estimates of Eqs. (1.32 ), (1.36), and (1.37) gives
us the following estimate for the mechanical velocity:

LEMMA 1.2
With s = 6 in Eq. (1.24), the solution [φ, φt , ϑ] of (1.9) satisfies the following estimate for all  > 0:

 T  T  T 
t
h(t) φt 2L 2 () dt ≤ C T 8 ϑ2L 2 () dt +  E dt.
0 0 0 2
April 4, 2005 10:3 3086 DK2961˙C001

1.5 Proof of Theorem 1.1(1) 17

1.5.2 Estimating the Mechanical Displacement

Here, we shall show the following:

LEMMA 1.3
The solution [φ, φt , ϑ] of Eq. (1.9) satisfies the following estimate for all , δ > 0:
 T  T   T
 1 2 t
 2 
h(t) Å φ L 2 () dt ≤ C T
13
2 −δ
ϑ L 2 () dt + 
2
h(t)E dt +  h(t)E(t)dt.
0 0 2 0

PROOF OF LEMMA 1.3 We start by applying the multiplier h(t)φ(t) to the mechanical
component in Eq. (1.9). We arrive at the relation
 T
 1 2
h(t)Å 2 φ  L 2 () dt
0
 T  T
= h  (t)(φt , φ) L 2 () dt + h(t) φt 2L 2 () dt
0 0
  T
T  1 1   1 1 
+ αλ h(t) Å G 2 γ0 ϑ, Å 2 φ
2
L 2 ()
dt − αλ h(t) Å 2 G 2 γ0 ϑ, Å 2 φ L 2 () dt
0 0
  T 
T
∂φ
−α h(t)(ϑ, φ) L 2 () dt + α h(t) (ϑ, λφ + dt. (1.38)
0 0 ∂ν L 2 ( )

Now, using the elliptic regularity posted in Eq. (1.16) and the usage of Lemma 1.1, with m = 0
and k = 3, we obtain
  T
 T  1   1 
 1 1
 h(t) Å G 2 γ0 ϑ, Å φ L 2 () dt − αλ
2 2 h(t) Å 2 G 2 γ0 ϑ, Å 2 φ L 2 () dt
 0 0
 T  T  
∂φ 

−α h(t)(ϑ, φ) L 2 () dt + α h(t) (ϑ, λφ + dt 
0 0 ∂ν L 2 ( ) 
 T  T 
26 t
≤ C T 3 ϑ2L 2 () dt +  h(t)E dt. (1.39)
0 0 2
Combining this estimate with that in Lemma 1.2 then gives the preliminary estimate

  
T  1 2  T 
h(t)Å 2 φ  L 2 () dt ≤  h  (t)(φt , φ) L 2 () dt 
0 0
 T  T 
t
+ CT 8
ϑ2L 2 () dt +  h(t)E dt. (1.40)
0 0 2
Apparently, we must estimate the first term on the right-hand side of Eq. (1.40). To this end, we
use the pointwise expression for φt in Eq. (1.27):

 T  T 
  1 −1
h (φt , φ) L 2 () dt = h Å ( ϑt − 2
ϑ) − αG 1 γ0 ϑt + αλG 2 γ0 ϑt , φ dt
0 0 α L 2 ()
(1.41)
April 4, 2005 10:3 3086 DK2961˙C001

18 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

1. The abstract Green’s Theorem gives

    
T
 −1
T
 ∂ ∂ B2
h( 2
ϑ, Å φ) L 2 () dt = h (ϑ, φ) L 2 () + + (1 − µ) ϑ, Å−1 φ
0 0 ∂ν ∂τ L2( )
 
∂ −1
− [ + (1 − µ)B1 ] ϑ, Å φ dt.
∂ν L2( )

Applying now the Lemma 1.1 with m = 3, 2 yields

  
 T  T
|h  | 1− 1k  A t 1+ 1k
 h(  2
ϑ, Å φ) L 2 () dt  ≤ C
−1 2 
ϑ L 2 () e 2 x0 H 2 dt.
 7
0 0 t 4

Let k ≥ 4. Then, because h  (t) = 6t 5 (T − 2t)(T − t)5 , we can apply now Hölder’s inequality
k
with Hölder conjugates (2 2k2−1 , 1 +21−k−1 ) so as to have
2
   T k−1
 T  −6 2k+1 2k+2 −6
 ϑ, Å−1 φ) L 2 () dt  ≤ C
2

 h( 2
t 2k −1 |T − 2t| 2k −1 (T − t) 2k −1 ϑ2L 2 () dt
0 0
 T   T  T 
t 13×2k−1 −12 t
+ h(t)E dt ≤ C T 2 k −1
ϑ L 2 () dt + 
2
h(t)E dt
0 2 0 0 2

(again this inequality being valid for k ≥ 4). Now for any δ > 0, we can rechoose integer k
large enough so as to have 13×22k −1−12 ≥ 13
k−1

2
− δ. This gives, then, for T < 1,
    
 T  T T
t
 h(  2
ϑ, Å φ) L 2 () dt  ≤ C T 2 −δ
−1 13
ϑ2L 2 () dt +  h(t)E dt. (1.42)
 2
0 0 0

(This is the term which ultimately dictates the singularity.)

2. Next,
 T  T

h [(G 1 − λG 2 )γ0 ϑt , φ] L 2 () dt = − h  [(G 1 − λG 2 )γ0 ϑ, φ] L 2 () dt
0 0
 T
− h  [(G 1 − λG 2 )γ0 ϑ, φt ] L 2 () dt. (1.43)
0

a. By the regularity posted in Eq. (1.16) and Lemma 1.1,

  
 T  T
|h  | 1− 1k  A t  1
  
h [(G 1 − λG 2 )γ0 ϑ, φ] L 2 () dt  ≤ C ϑ L 2 ()
2 e 2 x0 1+ 2k dt.
 1 H
0 0 t 4

k
Applying Hölder’s inequality to the right-hand side, with Hölder conjugates (2 2k2−1 ,
1
1
+2−k−1
) now yields
2

    −k−1
|h  | h(t) 2 +2
1
 T  T
 h  [(G 1 − λG 2 )γ0 ϑ, φ] L 2 () dt  ≤ C
 1
0 0 t 4 h(t)
 T  
 t 1+ 1k
1− 1 k−1
3 5×2 −4
T
t
× ϑ L 2 () eA 2 x0 H 2 dt ≤ C T 2k −1
2k
ϑ L 2 () + 
2
h(t)E dt.
0 0 2
April 4, 2005 10:3 3086 DK2961˙C001

1.5 Proof of Theorem 1.1(1) 19

Because for any δ > 0, we can choose integer k large enough so that 3 5×22k −1−4 ≥ 15
k−1

2
− δ,
we then get
 T   T  T 
  t
   15
h [(G 1 − λG 2 )γ0 ϑ, φ] L 2 () dt  ≤ C T 2 −
ϑ L 2 () + 
2
h(t)E dt.
 2
0 0 0
(1.44)

b. In the same way as above, we have for integer k large enough in Lemma 1.1,
 T   T  T 
  t
   15
h [(G 1 − λG 2 )γ0 ϑ, φt ] L 2 () dt  ≤ C T 2 −δ
ϑ L 2 () + 
2
h(t)E dt.
 2
0 0 0
(1.45)

The estimates Eqs. (1.44) and (1.45), applied to the relation Eq. (1.43) now give
 T   T  T 
  t
   15
h [(G 1 − λG 2 )γ0 ϑt , φ] L 2 () dt  ≤ C T 2 −
ϑ L 2 () + 
2
h(t)E dt.
 2
0 0 0
(1.46)

for integer k large enough.

3.
  
T

 −1
 T
 −1
T  
h Å ϑt , φ L 2 ()
dt = − h ( ϑ, Å φ) L 2 () dt − h ϑ, Å−1 φt L 2 ()
dt
0 0 0
(1.47)

a. By Green’s Theorem and Lemma 1.5, we have in a fashon similar to that in (1.a.),
 T    T
  
 h ( ϑ, Å φ) L 2 () dt  =  −
 −1
h  (θ, Å−1 φ) L 2 ()
 
0 0

      T
T
∂ 
 −1  k−1
3 5×2 −4
+ h θ, + λ Å φ  ≤ C T 2k −1 ϑ2L 2 ()
0 ∂ν L2( ) 0

 T   T  T 
t t
dt ≤ C T 2 −δ
15
+ h(t)E ϑ2L 2 () dt +  h(t)E dt,
0 2 0 0 2
(1.48)

for integer k large enough.

b. In the same way,
 T    
    T T
t
 h 
ϑ, Å −1
φ dt  ≤ C T 152 −δ ϑ2L 2 () dt +  h(t)E dt.
 t L ()
2  2
0 0 0
(1.49)

Eqs. (1.47), (1.48), and (1.49) together give the estimate

 T   T  
    T
t
 h 
ϑ , Å −1
φ dt  ≤ C T 152 −δ ϑ 2
dt +  h(t)E dt.
 t L ()
2  L ()
2
2
0 0 0
(1.50)

Combining Eqs. (1.40), (1.41), (1.46), (1.50), and (1.42) will complete the proof of Lemma 1.3.
April 4, 2005 10:3 3086 DK2961˙C001

20 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

1.5.3 Conclusion of the Proof of Theorem 1.1(1)

Combining Lemmas 1.2 and 1.3 gives the following estimate for the energy:
 T  T  T 
13
−δ t
E(t) dt ≤ C T 2 ϑ L 2 () dt + 
2
h(t)E dt;
0 0 0 2
or after changing limits of integration,
 T  T  T
2
2 −δ
13
[(1 − )h(t) − 2h(2t)] E(t) dt + (1 − ) h(t)E(t)dt ≤ C T ϑ2L 2 () dt.
T
0 2 0

For  > 0 small enough, this yields then

 T  T
h(t)E(t) dt ≤ C T 2 −δ
13
ϑ2L 2 () dt.
T
2 0

Using the dissipation inherent in the thermoelastic system (i.e., E(t) ≤ E(s) for s ≤ t), we finally
obtain
 T
E(T ) ≤ C T 2 −δ−13
13
ϑ2L 2 () dt.
0

This establishes the inequality Eq. (1.25), with C T = C T −q , where q = 13

4
− 2δ , for any δ > 0. This
concludes the proof of Theorem 1.1(1).

1.6 Proof of Theorem 1.1(2)

1.6.1 A First Supporting Estimate
In what follows, we will again make use of the polynomial weight h(t) in Eq. (1.24), with s = 4
therein.
In the present case of mechanical control, the necessary inequality (Eq. (1.12)) becomes

E(T ) ≤ C T φt  L 2 (Q) . (1.51)

to be valid for all finite energy solutions to Eq. (1.9).

We start by establishing the following estimate:

PROPOSITION 1.2
The solution [φ, φt , ϑ] of Eq. (1.9) satisfies the relation
 T   T  
  −1   T
t
 
h(t) A R ϑt , ϑ L 2 () dt  ≤ C T 4
φt  L 2 () dt + 
2
h(t)E dt.
 2
0 0 0

PROOF OF PROPOSITION 1.2 From the mechanical component of Eq. (1.9) we have, after
an extra differentiation in time, the expression −α ϑt = ∂t∂ 3 φ + 2 φt ; whence we obtain
3

1 −2 ∂ 3 1
A−1
R ϑt = A φ + A−2 2
φt ,
α R ∂t 3 α R
April 4, 2005 10:3 3086 DK2961˙C001

1.6 Proof of Theorem 1.1(2) 21

where the positive definite, self-adjoint operator A R : D(A R ) : L 2 () → L 2 () is as defined in
Eq. (1.19). Subsequently, we will have the following relation:

  
T   1 T
∂3
h(t) A−1
R ϑt , ϑ L 2 () dt = h(t) φ, A−2
R ϑ dt
0 α 0 ∂t 3 L 2 ()

1 T  
+ h(t) 2
φt , A−2
R ϑ L 2 ()
dt. (1.52)
α 0

We need to estimate the right-hand side of this expression.

1. For the first term on the right-hand side of Eq. (1.52), integration by parts gives

  3 
T
∂ T  
h(t) φ, A−2
R ϑ dt = h(t) φt , A−2
R ϑtt L 2 () dt
0 ∂t 3 L 2 () 0

 
T   T  
+2 h  (t) φt , A−2
R ϑt L 2 () dt + h  (t) φt , A−2
R ϑ L 2 () dt. (1.53)
0 0

We proceed to scrutinize each term on the right-hand side. To this end, we introduce the
(Robin) map R ∈ L[L 2 ( ), L 2 ()], defined by

∂g
Rf = g ⇔ g = 0 on  and + λg = f on (1.54)
∂ν

3
(by elliptic regularity, we have in fact that R ∈ L[H s ( ), H s+ 2 ()] for all real s). Using this
quantity with the heat equation in Eq. (1.9), we will then have the relations

  
∂
A−2 ϑ t = −A −1
ϑ − α A −1
I − R λγ 0 + φt ;
R R R
∂ν
      (1.55)
−2 ∂ −1 ∂
A R ϑtt = ϑ + α I − R λγ0 + φt − α A R I − R λγ0 + φtt .
∂ν ∂ν

a. From Eq. (1.56), we have

  
 T    T
 h(t) φt , A−2 dt  ≤
 R ϑtt L 2 () h(t)
0 0
     
 ∂ 
 
×  φt , ϑ + α I − R λγ0 + φt  dt
 ∂ν L 2 () 

      
T  ∂ 
 
+ h(t)  φt , α A−1 I − R λγ0 + φtt  dt. (1.56)
0  R
∂ν L 2 () 
April 4, 2005 10:3 3086 DK2961˙C001

22 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

To handle the most problematic term on the right-hand side of this expression (with again
x (t) = [φ(t), φt (t), ϑ(t)]), we use the singular trace estimate in Lemma 1.1(3):
 T    T  
 ∂  ∂ 
 
−1
h(t)  φt , A R R φtt  dt ≤ C h(t) t L 2 ()  φtt 
φ   dt
0  ∂ν L 2 ()  0 ∂ν  L 2 ( )
 T
h(t)  t 
≤C 3 φt  L 2 () eA 2 x (0)H dt
0 t4t
 T
h(t)
≤ C 7 φt 2L 2 () dt
0 t2
 T 
t
+ h(t)E dt
0 2
 T  T 
9 t
≤ C T 2 φt  L 2 () dt + 
2
h(t)E dt.
0 0 2
Applying this estimate to Eq. (1.56) and treating in like fashion the other terms on the
right-hand side thereof, we have
 T   T  T 
    t
 h(t) φ , A −2
ϑ dt  ≤ C T
9
φ  2
dt +  h(t)E dt.
   2
t R tt L 2 () t L 2 ()
0 0 0 2
(1.57)

b. Using the first relation in Eq. (1.56), we have, analogously to what was obtained in (1.a),
 T  T 
     
 h 
(t) φ , A −2
ϑ dt + 2 h 
(t) φ , A −2
ϑ dt 
 t R L 2 () t R t L 2 () 
0 0
 T 
|h  (t)|  t 
≤ C |h  (t)| + 3 φt  L 2 () eA 2 x (0)H dt
0 t4
 T  T 
t
≤ C T 4
φt  L 2 () dt + 
2
h(t)E dt. (1.58)
0 0 2
Combining Eqs. (1.56) and (1.58) now gives
  3   T  T 
 T ∂ 
 −2  t
 h(t) φ, A R ϑ dt  ≤ C T 4
φt  L 2 () dt + 
2
h(t)E dt.
 0 ∂t 3
L 2 ()  0 0 2
(1.59)

2. By the “Green’s” formula in Eq. (1.10), we have

 T  T
   
h(t) 2 φt , A−2
R ϑ L 2 ()
dt = h(t)a φt , A−2R ϑ L 2 () dt
0 0
 T  T  
 −2
 ∂ −2
+ h(t) αλϑt + φt , A R ϑ L 2 ( ) dt − h(t) φt + (1−µ)B1 φt , AR ϑ dt
0 0 ∂ν L2( )
 T   
∂
=− h(t) φt + (1 − µ)B1 φt , λI + A−2 ϑ dt
0 ∂ν R
L2( )
 T  T  
 2 −2
 ∂ −2
+ h(t) φt , A R ϑ L 2 () dt + h(t) φt , [ + (1 − µ)B1 ] A R ϑ dt
0 0 ∂ν L2( )
 T    
∂ ∂ B2
− h(t) φt , + (1 − µ) − I A−2 ϑ dt. (1.60)
0 ∂ν ∂τ R
L2( )
April 4, 2005 10:3 3086 DK2961˙C001

1.6 Proof of Theorem 1.1(2) 23

For the first term on the right-hand side of Eq. (1.60), we apply the Lemma 1.1(1) (with
m = 2 and D2 ≡ + (1 − µ)B1 therein) so as to have
    
 T ∂ 
 −2 
 h(t) φt + (1 − µ)B1 φt , λI + AR ϑ dt 
 0 ∂ν L ( )
2 
 T 
1− 1k  A t 1 1− 1k  A t  1
T
h(t) e 2 x0  2k ϑ L 2 () dt ≤ C h(t)  2 x0 1+ 2k .
≤ C 5 φt  H 2 ()
2
H 5 φt  H 2 () e
2
H
0 t4 0 t4
 
Now letting k = 2, say, we can invoke Hölder’s inequality, with Hölder conjugates 83 , 53 , to
obtain the estimate
    
 T ∂ 
 −2 
 h(t) φt + (1 − µ)B1 φt , λI + AR ϑ dt 
 0 ∂ν L ( )
2 
 T  T
14
≤ C T 3 h(t) φt 2L 2 () dt +  h(t)E(t/2) dt. (1.61)
0 0

Applying this estimate to the right-hand side of Eq. (1.60) and subsequently handling the
other terms thereof in a similar way—via the use of Lemma 1.1—we will have
 T 
  2  
 h(t) φt , A R ϑ L 2 () dt 
−1

0
 T  T
14
≤ C T 3 h(t) φt 2L 2 () dt +  h(t)E(t) dt. (1.62)
0 0

Combining Eqs. (1.52), (1.59), and (1.62) concludes the proof of Proposition 1.2.

1.6.2 Conclusion of the Proof of Theorem 1.1(2)

1. Estimating the Thermal Component. Applying the multiplier h(t)A−1 R ϑ(t) to the heat compo-
nent of the system Eq. (1.9) and subsequently invoking Proposition 1.2, we have

 
T T 
h(t) ϑ2L 2 () = − h(t) A−1R ϑt , ϑ dt
0 0
 T    
∂
−α h(t) I − R λγ0 + φt , ϑ dt
0 ∂ν
 T  T 
t
≤ C T 4 φt 2L 2 () dt +  h(t)E dt
0 0 2
 T  
 ∂ 
+ 
h(t) λφt + φt  ϑ L 2 () dt. (1.63)
0 ∂ν  L 2 ( )
∂
Via the Lemma 1.1 (with m = 1, D1 = λI + ∂ν , and k = 1, say), we can estimate the third
term on the right-hand side of Eq. (1.63) as
 T    T  T 
 ∂  t
h(t) λφ + φ
 t ∂ν t  2
 ϑ L ()
2 dt ≤ C T 5
φ  2
t L ()
2 dt +  h(t)E dt.
0 L ( ) 0 0 2
Combining this estimate with Eq. (1.63), we now obtain
 T  T  T 
t
h(t) ϑ2L 2 () ≤ C T 4 φt 2L 2 () dt +  h(t)E dt. (1.64)
0 0 0 2
April 4, 2005 10:3 3086 DK2961˙C001

24 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

2. Estimating the Mechanical Component. Here, we apply the multiplier intrinsic to uncoupled
plates and beams. To wit, from the mechanical component of Eq. (1.9), we have via h(t)φ(t)
and an invocation of the Green’s Theorem Eq. (1.10) the expression
 T  T  T  T
 1 2
Å 2 φ  2 dt = −α h(t)(ϑ, φ) dt + h 
(t)(φ t , φ) dt + h(t) φt 2L 2 () dt.
L ()
0 0 0 0
(1.65)
Applying the estimate of Eq. (1.64) (available for the thermal component) now gives
 T  T  T 
 1 2 t
Å 2 φ  2 dt ≤ C T 4 φ  2
t L 2 () dt +  h(t)E dt. (1.66)
L () 2
0 0 0

Combining the estimates of Eqs. (1.64) and (1.66) now give the estimate for the energy
 T  T  T 
t
h(t)E(t) dt ≤ C T 4 φt 2L 2 () dt +  h(t)E dt.
0 0 0 2
With this in hand, we can proceed as in the previous case so as to have the observability
inequality Eq. (1.51), with C T = T − 2 . Subsequently, we will determine that in the present case
5

of mechanical control, one has Emin (T ) = O(T − 2 ). This concludes the proof of Theorem 1.1(2)
5

with free boundary conditions and one control.

1.7 Proof of Theorem 1.1(3)

Here we set the index s = 2 in Eq. (1.24). In this present case of dual—mechanical and thermal—
control, the necessary inequality is

E(T ) ≤ C T (φt  L 2 (Q) + ϑ L 2 (Q) ), (1.67)
where again [φ, φt , ϑ] solve the homogeneous system Eq. (1.9). Using the relation Eq. (1.65),
we have
 T  T
 1 2 [h  (t)]2  
(1 − ) h(t)Å 2 φ  L 2 () dt ≤ C h(t) + φt 2L 2 () + ϑ2L 2 () dt
0 0 h(t)
 T
 
≤ CT 2 φt 2L 2 () + ϑ2L 2 () dt.
0
This then gives
  T
T 
h(t)E(t) dt ≤ C T 2 φt 2L 2 () + ϑ2L 2 () dt,
0 0
whence we obtain the inequality Eq. (1.67). From here, we can use the usual algorithmic argument
so as to have Emin (T ) = O(T − 2 ). This concludes the proof of Theorem 1.1(3).
3

References
[1] P. Albano and D. Tataru, Carleman estimates and boundary observability for a coupled
parabolic-hyperbolic system, Electron. J. Differential Equations, 2000, No. 22 (2000),
1–15.
April 4, 2005 10:3 3086 DK2961˙C001

References 25

[2] G. Avalos, Exact controllability of a thermoelastic system with control in the thermal compo-
nent only, Differential Integral Equations, 13, (2000), 613–630.

[3] G. Avalos and I. Lasiecka, Boundary controllability of thermoelastic plates via the free bound-
ary conditions, SIAM J. Control Optim., 38, No. 2 (2000), 337–383.

[4] G. Avalos and I. Lasiecka, Exponential Stability of a Thermoelastic System without Mechanical
Dissipation, Rendiconti dell’Istituto di Matematica dll’Università di Trieste, Vol. XXVIII,
Supplemento (1996), 1–28.

[5] G. Avalos and I. Lasiecka, Exponential stability of a thermoelastic system with free boun-
dary conditions without mechanical dissipation, SIAM J. Math. Anal., 29, No. 1 (1998), 155–
182.

[6] G. Avalos and I. Lasiecka, A Note on the Null Controllability of a Thermoelastic Plates and
Singularity of the Associated Minimal Energy Function, Scuola Normale Superiore (Pisa),
Preprints di Matematica, n. 10 (Giugno 2002).

[7] G. Avalos and I. Lasiecka, Optimal blowup rates for the minimal energy null control of the
strongly damped abstract wave equation, Ann. Scuola Norm. Sup. Pisa Cl. Sci. (5), II (2003),
601–616.

[8] G. Avalos and I. Lasiecka, Mechanical and thermal null controllability of thermoelastic plates
and singularity of the associated minimal energy function, Control Cybern. special volume
dedicated to K. Malanowski. 32, No. 3 (2003), 473–491.

[9] G. Avalos and I. Lasiecka, The null controllability of thermoelastic plates and singularity of
the associated minimal energy function, J. Math. Anal. Appl., 294, (2004), 34–61.

[10] G. Avalos and I. Lasiecka, Boundary controllability of thermoelastic plates via the free bound-
ary conditions, SIAM J. Control Optim., 38, No. 2 (2000), 337–383.

[11] A. Benabdallah and M. G. Naso, Null controllability of a thermoelastic plate, Abstr. Appl.
Anal., in press.

[12] A. Bensoussan, G. Da Prato, M. Delfour, and S. Mitter, Representation and Control of Infinte
Dimensional Systems, Vol. II, Birkhäuser, Boston, (1993).

[13] J. Cea, Lectures on Optimization—Theory and Algorithms Published for the Tata Institute of
Fundamental Research, Springer-Verlag, New York, (1978).

[14] G. Da Prato and J. Zabczyk, Stochastic Equations in Infinite Dimensions, Cambridge University
Press, New York, New York, (1992).

[15] G. Da Prato, An Introduction to Infinite Dimensional Analysis, Scuola Normale Superiore,

(2001).

[16] M. Eller, I. Lasiecka, and R. Triggiani. Simultaneous exact-approximate boundary controlla-

bility of thermoelastic plates with variabale thermal coefficients and moment control. J. Math.
Anal. Appl., 251, (2000), 452–478.

[17] M. Eller, I. Lasiecka and R. Triggiabi. Simultaneous exact controllability of thermoelastic

plates with variable thermal coefficients and clamped/Dirichlet boundary controls, Discrete
Continuous Dynamical Syst., 7, No. 2 (2001), 283–302.
April 4, 2005 10:3 3086 DK2961˙C001

26 Asymptotic Rates of Blowup for the Minimal Energy Function for the Null Controllability

[18] S. Hansen and B. Zhang, Boundary control of a thermoelastic beam, J. Math. Anal. Appl., 210,
(1997), 182–205.

[19] F. Gozzi and P. Loreti, Regularity of the minimum time function and minimum energy problems,
SIAM J. Control Optim., in press.

[20] B. Guichal, A lower bound of the norm of the control operator for the heat equation. J. Math.
Anal. Appl. 110, No. 2 (1985), 519–527.

[21] P. Grisvard, Caracterization de quelques espaces d’interpolation, Arch. Rational Mech. Anal.,
25, (1967), 40–63.

[22] J. Lagnese, Boundary Stabilization of Thin Plates, SIAM Stud. Appl. Math., 10, SIAM,
Philadelphia, (1989).

[23] J. Lagnese, The reachability problem for thermoelastic plates, Arch. Rational Mech. Anal.,
112, (1990), 223–267.

[24] J. Lagnese and J.L. Lions, Modelling, Analysis and Control of Thin Plates, Masson Paris,
(1988).

[25] I. Lasiecka and T. Seidman, Thermal boundary control of a thermoelastic system,

preprint (2004).

[26] I. Lasiecka and R. Triggiani, Analyticity of thermo-elastic semigroups with clamped/hinged

B.C., Abst. Appl. Anal., 3, No. 2 (1998), 153–169.

[27] I. Lasiecka and R. Triggiani, Analyticity of thermo-elastic semigroups with coupled BC. Part II:
The case of free BC, Annali Scuola Normale di Pisa, Classes Scienze (Serie IV, Fasicolo 3–4),
Vol. XXVII (1998[c]), 457–497.

[28] I. Lasiecka and R. Triggiani, Control Theory for Partial Differential Equations: Continuous
and Approximation Theories. I: Abstract Parabolic Systems, Cambridge University Press,
New York, (2000).

[29] I. Lasiecka, Uniform decay rates for full von Karman system of dynamic thermoelasticity
with free boundary conditions and partial boundary dissipation, Commun. Partial Differential
Equations, 24, 9&10 (1999), 1801–1847.

[30] J.L. Lions and E. Magenes, Non-Homogeneous Boundary Value Problems and Applications,
Vol. 1, Springer-Verlag, New York, (1972).

[31] Z. Liu and M. Renardy, A note on the equations of a thermoelastic plate, Appl. Math. Lett., 8,
No. 3 (1995), 1–6.

[32] A. Lunardi, Schauder’s estimates for a class of degenerate elliptic and parabolic operators with
unbounded coefficients in Rn , Ann. Scuola Sup. Pisa (IV), XXIV (1997), 133–164.

[33] T.I. Seidman, How fast are violent controls? Math. Controls Signals Syst., (1988), 89–95.

[34] T.I. Seidman, Two results on exact boundary control of parabolic equations, Appl. Math.
Optim., 11, (1984), 145–152.

[35] T.I. Seidman and J. Yong, How fast are violent controls?, II, Math. Controls Signals Syst., 9,
(1997), 327–340.
April 4, 2005 10:3 3086 DK2961˙C001

References 27

[36] T. Seidman, S. Avdonian, and S. Ivanov, The window problem for series of complex exponen-
tials, J. Fourier’s Anal., 6, (2000), 235–254.

[37] V. Thomée, Galerkin Finite Element Methods for Parabolic Problems, Springer-Verlag,
New York, (1984).

[38] R. Triggiani, Analyticity, and lack thereof, of semigroups arising from thermo-elastic plates, in
the special volume entitled Computational Science for the 21st Century, Chapter on Control:
Theory and Numerics, Wiley 1997; Proceedings in honor of R. Glowinski, (May, 1997).

[39] R. Triggiani, Optimal estimates of norms of fast controls in exact null controllability of two
non-classical abstract parabolic systems, Adv. Differential Equations, 8, No. 2 (2003), 189–229.

[40] J. Zabczyk, Mathematical Control Theory, Birkhäuser, Boston, (1992).

April 4, 2005 10:3 3086 DK2961˙C001
April 5, 2005 13:52 3086 DK2961˙C002

Chapter 2
Interior and Boundary Stabilization
of Navier-Stokes Equations

Viorel Barbu
Alexandru Ioan Cuza University

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.2 Part I: Interior Control [4] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.2 Main Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3 Part II: Boundary Control [3] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.2 Main Results (Case d = 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Abstract We report on very recent work on the stabilization of the steady-state solutions to
Navier-Stokes equations on an open bounded domain  ⊂ R d , d = 2, 3, by either interior or else
boundary control.
More precisely, as to the interior case, we obtain that the steady-state solutions to Navier-Stokes
equations on  ⊂ R d , d = 2, 3, with no-slip boundary conditions, are locally exponentially stabi-
lizable by a finite-dimensional feedback controller with support in an arbitrary open subset ω ⊂ 
of positive measure. The (finite) dimension of the feedback controller is minimal and is related to
the largest algebraic multiplicity of the unstable eigenvalues of the linearized equation.
Second, as to the boundary case, we obtain that the steady-state solutions to Navier-Stokes
equations on a bounded domain  ⊂ R d , d = 2, 3, are locally exponentially stabilizable by a
boundary closed-loop feedback controller, acting tangentially on the boundary ∂, in the Dirichlet
boundary conditions. If d = 3, the nonlinearity imposes and dictates the requirement that stabilization
must occur in the space [H 2 +  ()]3 ,  > 0, a high topological level. A first implication thereof is
3

that, for d = 3, the boundary feedback stabilizing controller must be infinite dimensional. Moreover, it
generally acts on the entire boundary ∂. Instead, for d = 2, where the topological level for stabilizat-
ion is [H 2 − ()]2 , the boundary feedback stabilizing controller can be chosen to act on an arbitrarily
3

small portion of the boundary. Moreover, still for d = 2, it may even be finite dimensional, and this
occurs if the linearized operator is diagonalizable over its finite-dimensional unstable subspace.

2.1 Introduction
We hereby report on recent joint work on the stabilization of steady-state solutions to Navier-
Stokes equations on an open bounded domain  ⊂ R d , d = 2, 3, by either interior feedback control
or else boundary feedback control. The case of interior control is taken from the joint work with
Triggiani in Reference 4. The case of boundary control is taken from the joint work with Lasiecka
and Triggiani in Reference 3. To enhance readability, we provide independent accounts of each case.

29
April 5, 2005 13:52 3086 DK2961˙C002

30 Interior and Boundary Stabilization of Navier-Stokes Equations

2.2 Part I: Interior Control [4]

2.2.1 Introduction
The controlled Navier-Stokes equations
Consider the controlled Navier-Stokes equations (see Reference 6, p. 45, and Reference 13, p. 253
for the uncontrolled case u ≡ 0) with the non-slip Dirichlet B.C.:
yt (x, t) − ν y(x, t) + (y · ∇)y(x, t) = m(x)u(x, t) + f e (x)
+ ∇ p(x, t) in Q =  × (0, ∞), (2.1)

∇ · y=0 in Q;
y=0 on = ∂ × (0, ∞);
y(x, 0)=y0 (x) in .

Here,  is an open, smooth, bounded domain of R d , d = 2, 3; m is the characteristic function of an

open smooth subset ω ⊂  of positive measure; u is the control input; and y = (y1 , y2 , . . . , yd ) is
the state (velocity) of the system. The function v = mu can be viewed itself as an internal controller
with support in Q ω = ω × (0, ∞). The functions y0 , f e ∈ [L 2 ()]d are given, the latter being a body
force, whereas p is the unknown pressure.
Let (ye , pe ) ∈ [(H 2 ()]d ∩ V ) × H 1 () be a steady-state solution to Eq. (2.1), that is,
−ν ye + (ye · ∇)ye = f e + ∇ pe in ; (2.2)

∇ · ye = 0 in ;
ye = 0 on ∂.
The steady-state solution is known to exist for d = 2, 3, (see Reference 6, Theorem 7.3, p. 59). Here
[6, p. 9], [13, p. 18]

 d 
V = y ∈ H01 () ; ∇ · y = 0 , with norm yV ≡ y
  12
= |∇ y(x)|2 d . (2.3)


Literature
According to some recent results of Imanuvilov [9] (see also Reference 1) any such solution ye
is locally exactly controllable on every interval [0, T ] with controller u with support in Q ω . More
precisely, if the distance ye − y0  H 2 () is sufficiently small, then there is a solution (y, p, u) to
Eq. (2.1) of appropriate regularity such that y(T ) ≡ ye . The steering control is open-loop and
depends on the initial condition. Subsequently, Reference 2 proved that any steady-state solution
ye is locally exponentially stabilizable by means of an infinite-dimensional feedback controller by
using the controllability of the linear Stokes equation. In contrast, here we shall prove, via the state
decomposition technique of References 14 and 15 and the first-order stabilization Riccati equation
method developed in our previous work (Reference 2; see also Reference 5 still in the parabolic
case, as well as Reference 11 in the hyperbolic case), that any steady-state solution ye is locally
exponentially stabilizable by a finite-dimensional closed-loop feedback controller of the form

2K
u=− (R N (y − ye ), ψi )ω ψi , (2.4)
i=1
April 5, 2005 13:52 3086 DK2961˙C002

2.2 Part I: Interior Control [4] 31

1 1
where R N ∈ L(D(A 4 )]∩L[D(A 2 ); H ) is the solution of the algebraic Riccati Eq. (2.19) below asso-
ciated with the linearized system of Eq. (2.14) below and {ψi }i=1 2K
is an explicitly constructed (in
(3.3.5) of Reference 4) system of functions related to the space of eigenfunctions corresponding to the
unstable eigenvalues of such linearized system. Here A is the Stokes operator defined by Eq. (2.6);
H the space in Eq. (2.5); and ( · , · )ω is the scalar product in [L 2 (ω)]d . The present closed-loop
feedback stabilization result has two main features, besides being finite dimensional:
1. It is more precise and less restrictive concerning the vectors y0 and ye than the open-loop
version provided by the local exact controllability result established in Reference 9 or the
closed-loop stabilization in Reference 2 (in that smallness of the distance between y0 and ye
1 1
is measured in the D(A 4 )-norm, i.e., the [H 2 ()]d -norm, see the set Vρ in Eq. (2.22) below
rather than in the [H ()] -norm, as recalled above, where A is defined in Eq. (2.6)).
2 d

2. It is independent of the Carleman inequality for the Stokes equation, which is necessary for
the proof of local controllability.
There is a large body of literature on the stabilization problem of steady-state solutions to Navier-
Stokes equations. Here we confine ourselves to mention only a few of the papers (References 2 and 7),
which are more related to this present work. We also refer to the recent paper of Fursikov [8] for a
study of a boundary—rather than interior—problem for the Navier-Stokes equations, which, however,
does not pertain to the topic of feedback stabilization in the established sense, as in the present paper.

Notation
Here we shall use the standard notation for the spaces of summable functions and Sobolev spaces
on . In particular, H s () is the Sobolev space of order s with the norm denoted by  · s . The
following notation will also be used:

∇ · y = div y, (y · ∇)y = yi Di y j = y · ∇ y j ,
∂
j = 1, . . . , d, Di = ;
∂ xi

H = {y ∈ [L 2 ()]d ; ∇ · y = 0, y · n = 0 on ∂} (see Reference 6, p. 7);

H ⊥ = {y ∈ [L 2 ()]d : y = ∇ p, p ∈ H 1 ()}, [L 2 ()]d = H + H ⊥ , (2.5)

H ⊥ being the orthogonal complement of H in [L 2 ()]d (see Reference 13, p. 15) with summation
convention to be used throughout the paper, presently in i = 1, . . . , d, where n is the outward
normal to the boundary ∂ of . We shall denote by P : [L 2 ()]d → H the orthogonal Leray
projector (Reference 6, p. 9), and moreover (Reference 6, p. 31),
1
Ay = −P y, ∀ y ∈ D(A) = [H 2 ()]d ∩ V , V = D(A 2 ), (2.6)

which is a self-adjoint positive definite operator in H with compact (resolvent) A−1 on H (Refer-
ence 6, p. 32). Accordingly, the fractional powers As , 0 < s < 1, are well defined (Reference 6, p. 33).
We have V = D(A 2 ) (Reference 6, p. 33). Furthermore, we define B : V → V  by (Reference 6,
1

p. 47, p. 54), (Reference 13, p. 162),

By = P[(y · ∇)y], (By, w) = b(y, y, w), ∀ y, w ∈ V, (2.7)

where the trilinear form is defined by (Reference 6, p. 49), (Reference 13, p. 161)
 
b(y, z, w) = yi (Di z j )w j d x = y · ∇z, w R d d,
 
y, w ∈ H, z ∈ V. (2.8)
April 5, 2005 13:52 3086 DK2961˙C002

32 Interior and Boundary Stabilization of Navier-Stokes Equations

We shall denote by (·, ·) the scalar product in both H and [L 2 ()]d . Similarly, we shall denote
by the same symbol | · | the norm of both [L 2 ()]d and H and by  ·  the norm of the space V as
defined in Eq. (2.3).

Preliminaries
In the notation introduced above, Eq. (2.1) can be equivalently rewritten in abstract form as
dy
+ ν Ay + By = P(mu + f e ); y(0) = y0 ∈ H, (2.9)
dt
because the procedure of applying P to Eq. (2.1) eliminates the pressure from the equations in
Reference 6, p. 47, the orthogonal space H ⊥ to H being made up of [L 2 ()]d -functions which are
the gradients of H 1 ()-functions by Eq. (2.5). Moreover, y ∈ H for Eq. (2.1) implies P yt = yt .

2.2.2 Main Results

Assumptions
1. The boundary ∂ of  is a finite union of d − 1 dimensional C 2 -connected manifolds.
Moreover, the boundary ∂ω of ω is of class C 2 .

2. The steady-state solution (ye , pe ) defined in Eq. (2.2) belongs to ([H 2 ()]d ∩ V ) × H 1 (),
where we recall from Eq. (2.6) that then ye ∈ D(A). (For d = 2, 3, this property is guaranteed
by Reference 6, Theorem 7.3, p. 59) on ye , for f e ∈ H , followed by Reference 6, Theorem 3.11,
p. 30 on pe , for sufficiently smooth ∂.)

Preliminaries. The translated problem

By the substitutions y → ye + y, p → pe + p, we are readily led via Eq. (2.1), Eq. (2.2) to the
study of null stabilization of the equation
yt − ν y + (y · ∇)y + (ye · ∇)y + (y · ∇)ye = mu + ∇ p in Q; (2.10)

∇ · y = 0 in Q;
y = 0 on ;
y(x, 0) = y 0 (x) = y0 (x) − ye (x).
By use of Eq. (2.5) on H , Eq. (2.6) and Eq. (2.7) on A and B, we see that Eq. (2.10), after application
of P, can be rewritten abstractly as
dy
+ ν Ay + A0 y + By = P(mu), t > 0; y(0) = y 0 (2.11)
dt
[compare with Eq. (2.9), again P yt = yt , because y ∈ H by Eq. (2.10)], where we have now
introduced the operator A0 ∈ L(V ; H ),
1
A0 y = P[(ye · ∇)y + (y · ∇)ye ], D(A0 ) = V = D(A 2 ), (2.12a)
or equivalently, recalling Eq. (2.7),
(A0 y, z) = b(ye , y, z) + b(y, ye , z), ∀ y ∈ V, z ∈ H. (2.12b)
The operator A0 in Eq. (2.12) is well-defined H ⊃ V = D(A0 ) → H . This follows from the
estimate
1
|A0 y| ≤ C1 ye 2 y, ∀ y ∈ V = D(A0 ) = D(A 2 ), (2.13)
which is obtained directly by use of the definition of Eq. (2.12b).
April 5, 2005 13:52 3086 DK2961˙C002

2.2 Part I: Interior Control [4] 33

The linearized problem

Next, we consider the following linearized system of the translated model in Eq. (2.10) or
Eq. (2.11):
dy
+ ν Ay + A0 y = P(mu), t > 0; y(0) = y 0 ∈ H. (2.14)
dt
We have already noted in Eq. (2.6) that the operator −ν A (ν > 0, the viscosity coefficient) is neg-
ative self-adjoint and has compact resolvent on H . Thus, −ν A generates an analytic (self-adjoint)
1
C0 -semigroup on H . It then follows from here and from D(A0 ) = V = D(A 2 ), as noted in Eq. (2.6)
and in Eq. (2.13), that: the perturbed operator
A = −(ν A + A0 ), with domain D(A) = D(A) = [H 2 ()]d ∩ V (2.15)
likewise has compact resolvent and generates an analytic C0 -semigroup on H . This is well known.
It follows from the above claim that the operator A has a finite number N of eigenvalues λ j with Re
λ j ≥ 0 (the unstable eigenvalues). The eigenvalues are repeated according to their algebraic mul-
tiplicity  j . Let {ϕ j } Nj=1 be a corresponding system of generalized eigenfunctions, ϕ j = ϕ 1j + iϕ 2j ,
j = 1, . . . , N of A. (See Reference 10, p. 41, p. 181.) More precisely, we shall denote by M the
number of distinct unstable eigenvalues, so that 1 + 2 + · · · +  M = N . In order to state our first
result, we finally need to introduce the following finite-dimensional real spaces X αN , α = 1, 2 as well
as the following natural number K :

N
X αN = span ϕ αj j=1 ; K = max{ j ; 1 ≤ j ≤ M}. (2.16)

Main results. Linearized problem Eq. (2.14)

We first state the following feedback stabilization result for the linearized system in Eq. (2.14).

THEOREM 2.1
Let  > 0 be arbitrary but fixed, and let γ0 = |Reλ N +1 | − . Then, for each λ, 0 ≤ λ ≤ γ0 , there are
functions {ψi }i=1
K
⊂ X 1N , {ψi }i=K
2K
+1 ⊂ X N and a linear self-adjoint operator R N : D(R N ) ⊂ H →
2

H such that for some constants 0 < a1 < a2 < ∞ and C1 > 0, we have:
1.
1 1 1
a1 |A 4 y|2 ≤ (R N y, y) ≤ a2 |A 4 y|2 , ∀ y ∈ D(A 4 ), (2.17)
1
1
so that D(A 4 ) ⊂ D(R N2 );

2.
1
|R N y| ≤ C1 y, ∀ y ∈ V = D(A 2 ); (2.18)
3. R N satisfies the following algebraic Riccati equation:

1
2K
1 3
−((A + λ)y, R N y) + (ψi , R N y)2ω = |A 4 y|2 , ∀ y ∈ D(A). (2.19)
2 i=1 2

The vectors {ψi }i=1

2K
are explicitly constructed in (3.3.5) of Lemma 4 of Reference 4. Moreover, with
2K ≤ N , the feedback controller,

2K
u=− (R N y, ψi )ω ψi (2.20a)
i=1
April 5, 2005 13:52 3086 DK2961˙C002

34 Interior and Boundary Stabilization of Navier-Stokes Equations

once inserted in Eq. (2.14), exponentially stabilizes the corresponding closed-loop system of
Eq. (2.14). The margin of stability for such a closed-loop system is λ. [See Remark 3.3.1 of Reference 4
for the effective number of controls 2K ≤ N .] More specifically, this means that the solution of

dy  2K
1
+ ν Ay + A0 y + P m (R N y, ψi )ω ψi = 0, y(0) = y 0 ∈ D(A 4 ) (2.20b)
dt i=1

satisfies

|A 4 y(t)| ≤ Cλ e−λt |A 4 y 0 |, t ≥ 0.
1 1
(2.20c)

Nonlinear system Eq. (2.9)

We next use the stabilizer in Theorem 2.1 to the linearized system of Eq. (2.14) of the translated
problem (2.1), or (2.2), to obtain the sought-after closed loop, local, feedback stabilization of the
steady-state solution ye to the Navier-Stokes Eq. (2.9).

THEOREM 2.2
With reference to Theorem 2.1, the feedback controller


2K
u=− (R N (y − ye ), ψi )ω ψi (2.21)
i=1

(where the vectors ψi are defined in (3.3.5) of Lemma 4 in Reference 4), once inserted in the
Navier-Stokes system Eq. (2.9), exponentially stabilizes the steady-state solution ye to Eq. (2.1) in
a neighborhood
1 1
Vρ = {y0 ∈ D(A 4 ); |A 4 (y0 − ye )| < ρ} (2.22)

of ye , for suitable ρ > 0. More precisely, if ρ > 0 is sufficiently small, then for each y0 ∈ Vρ there
exists a weak solution y ∈ L ∞ (0, T ; H ) ∩ L 2 (0, T ; V ), dy ∈ L 3 (0, T ; V  ) for d = 3, and dy
4

dt dt
∈
L 2 (0, T ; V  ) for d = 2, ∀ T > 0, to the closed-loop system

dy  2K
+ ν Ay + By + P m (R N (y − ye ), ψi )ω ψi = P f e , t ≥ 0, y(0) = y 0 , (2.23)
dt i=1

obtained from inserting the control of Eq. (2.21) in Eq. (2.9), such that the following two properties
hold:
1.
 ∞
3 2 1 2
e2λt A 4 (y(t) − ye ) dt ≤ C2 A 4 (y0 − ye ) ; (2.24)
0

2.

A 4 (y(t) − ye ) ≤ C3 e−λt A 4 (y0 − ye ) ,

1 1
∀ t ≥ 0. (2.25)

We refer to Reference 6, p. 71 for definition of weak solutions to equations of the form in Eq. (2.23)
and the asserted regularity. If d = 2, the solution to Eq. (2.23) is strong and unique (see Reference 6,
p. 83).
April 5, 2005 13:52 3086 DK2961˙C002

2.3 Part II: Boundary Control [3] 35

The pressure p
Theorem 2.2 implies the following result giving corresponding asymptotic properties of the
pressure p.

THEOREM 2.3
The solution y provided by Theorem 2.2 satisfies also the equation


2K
yt − ν y + (y · ∇)y + m (R N (y − ye ), ψi )ω ψi = ∇ p + f e in Q ≡  × (0, ∞);
i=1
(2.26)

∇·y ≡ 0 in Q;
y ≡0 on = ∂ × (0, ∞);
y(x, 0) = y0 (x) in .

Moreover, the following relations hold true for the pressure p:

1. for d = 2, we have
 ∞
1 2 1 2
t| p(t) − pe |2[H 1 ()]d dt ≤ C A 4 (y0 − ye ) 1 + A 4 (y0 − ye ) ; (2.27)
0

2. for d = 3, we have
 ∞
1 2  1 2
| p(t) − pe |2[L 2 ()]d/R dt ≤ C A 4 (y − ye ) 1 + A 4 (y0 − ye ) . (2.28)
0

2.3 Part II: Boundary Control [3]

2.3.1 Introduction
Boundary controlled Navier-Stokes equations
We consider the controlled Navier-Stokes equations (see Reference 6, p. 45, and Reference 13,
p. 253 for the uncontrolled case u ≡ 0) with boundary control u in the Dirichlet B.C.:

 yt (x, t) − ν0 y(x, t) + (y · ∇)y(x, t) = f e (x) + ∇ p(x, t) in G; (2.29a)


 ∇·y = 0 in G; (2.29b)

 y = u on ; (2.29c)


y(x, 0) = y0 (x) in . (2.29d)

Here, G =  × (0, ∞); = ∂ × (0, ∞) and  is an open smooth bounded domain of R d ,

d = 2, 3; u ∈ L 2 (0, T ; [L 2 (∂)]d ) is the boundary control input; and y = (y1 , y2 , . . . , yd ) is the
state (velocity) of the system. The constant ν0 > 0 is the viscosity coefficient. The functions y0 , f e ∈
[L 2 ()]d are given, the
 latter being abody force, whereas p is the unknown pressure. Because of the
divergence theorem:  ∇ · y d =  y · ν d, [ = ∂, ν = unit outward normal to ∂], we must
require (at least) the integral boundary compatibility condition:  u · ν d = 0 on the control
function u. Actually, a more stringent condition has to be imposed, in our final results: u · ν ≡ 0
on , to sustain the pointwise boundary compatibility condition contained in the definition of the
Exploring the Variety of Random
Documents with Different Content
 Squallide piagge, ahi d'altra morte degni
Gl'Itali prodi....

e sia pure, come il Leopardi cantò; ma ancora ne rimasero tanti da

fare tre rivoluzioni sulla terra nostra, da esser maestri alla
generazione che l'ha composta in un regno. Ond'io posso applicare a
noi ciò ch'egli diceva per tutti: “Qual gioventù lascio io dopo di me!
Ed è opera mia. Essa mi vendicherà abbastanza, con tutto quello che
saprà volere. Veduta l'opera, converrà bene render giustizia
all'artefice.„ Il quale, non lo dimentichiamo, fu colpito dal fato nella
pienezza degli anni virili. Dapprima confinato all'Elba, fra le tre culle
della sua gente, San Miniato, Sarzana ed Aiaccio, tacque, sperando
sempre di fare; poi quasi fuori del mondo, sopra uno scoglio
dell'Oceano, come un eroe antico, disperando di fare, ha detto il
segreto dell'anima sua, epico nell'opera, tragico nella contemplazione
che ne ha fatta, rivivendola intiera. E si assiste a quella evocazione,
come alla recita di un dramma, con tutte le ansie dell'affetto
commosso, con tutti i dubbi della mente turbata, sapendo la
catastrofe, e pure desiderando, quasi sperando il miracolo ch'ella
riesca diversa. E là non più un uomo; è tutto un simbolo; il simbolo
della nuova Europa che sorge, combatte e cade, lasciando ferito a
morte il nemico. Dopo di lui, è bene l'Europa contemporanea; il
passato può dar guizzi di serpe troncato; son guizzi d'agonia, mentre
tutte le cose vitali rinascono; la patria nostra, ad esempio.... Ma io
non farò un inno, qui. So bene che tutto non è andato secondo le
sapienti preparazioni e le giuste speranze dei migliori; che i pericoli
non son tutti stornati da noi, se ancora ne durano in noi; che
bisogna pensare, provvedere, meritare con forza e con senno una
grande fortuna. Sia dunque l'inno riserbato alla Italia futura, se la
dovremo davvero a noi stessi, savia, operosa, concorde, soprattutto
concorde; e Dante e il Machiavello, grandi anime fiorentine,
l'assistano.
 V.

È a Milano, nell'atrio di Brera, un Napoleone di bronzo, colossale;

non lui, veramente, ma il suo genio, come l'ha ideato Antonio
Canova. Vedendolo, senza quegli indumenti ond'è caratteristica
l'immagine dell'Imperatore, intendo la verità profonda di ciò che
l'artista di Possagno diceva spesso e volentieri intorno alla
significazione del nudo nell'arte. Così, fatto genio, Napoleone è il
capolavoro del Canova. Altri lodi Ebe, e Psiche e le Grazie
abbracciate; altri si compiaccia di Paolina giacente, o ammiri papa
Rezzonico orante tra i suoi leoni pensosi. Nell'atrio di Brera penso io,
davanti al colosso. Foggiato per andare all'aperto, con la vittoria
alata nel cavo della mano, egli muove allo stretto, là dentro, e si
pensa che in men di due passi avrà finito il suo corso. Perchè là
dentro? All'aperto non lo vollero i potenti, che approfittavano della
sventura sua; e fu ancora lodevol pudore che lo lasciassero intatto,
come fu buon consiglio negli ultimi successori concederlo alla
ammirazione sminuita dei tempi nuovi confinato in quel chiuso, che è
pur sempre un tempio della scienza e dell'arte. E va, senza muoversi
dal suo piedestallo di granito, fremente nella sua grandezza, che ha
ancora ed avrà sempre alcun che di segreto. E più torreggia allo
stretto, e più sembra che cammini, mentre a noi par di comprendere
tutto ciò ch'egli non fu, avendo la potenza di essere, e intorno a lui,
irta di ostacoli, serrandosi la congiura del mondo.
Veduto quello, muovo a cercarne il riscontro, in un altro cortile, di là
dal Naviglio, ove si raccoglievano i Senatori di quello che Ugo
Foscolo chiamò “il bello italo regno„. Non più un genio, là dentro, ma
un uomo, un cavaliere salutante; in atto di muoversi anch'egli,
caracollando, e anch'egli rinchiuso! Ragioni d'ordine vario han fatti
prigionieri i due bronzi, non volendoli in piazza; dove infine, piacendo
meno alle moltitudini, riuscirebbero meno eloquenti. Lampi di genio
inconsapevole; ne siano perdonate (quasi direi benedette) le cause,
per il gaudio estetico che destano in noi. Vedo quell'altro, e penso....
penso una grande giornata e un indimenticabil servizio. Quel
cavaliere che saluta, levandosi sul cavallo di mezzo al fogliame d'una
aiuola fiorita, un po' sfinge nella immobilità dello sguardo, ma
gentiluomo nel sorriso e nel gesto cortese, si associa nel mio
pensiero al rifiorimento di una leggenda di gloria che Roma ha
cantata nei secoli. E rammentando quante ombre circondino certe
figure storiche, sento anche meglio la solenne grandezza onde sono
privilegiate. Nell'atrio del Senato vedo la continuazione ideale
dell'atrio di Brera. Solferino procede da Sant'Elena; il discendente
degli antichi condottieri italiani (piace a me pure di chiamarlo così) è
stato il liberatore della patria schiava, per aver dato il primo crollo, e
come vigoroso! alle mura istesse del suo carcere quindici volte
secolare; onde il mio panegirico, se parrà tale, è sentenza di storia.
Infine, odio la storia che non illumina i fatti con luce viva d'amore.
 Nota del Trascrittore
Ortografia e punteggiatura originali sono state
mantenute, correggendo senza annotazione minimi errori
tipografici.
Copertina creata dal trascrittore e posta nel pubblico
dominio.
 *** END OF THE PROJECT GUTENBERG EBOOK NAPOLEONE: LA
VITA ITALIANA DURANTE LA RIVOLUZIONE FRANCESE E L'IMPERO
***

Updated editions will replace the previous one—the old editions will
be renamed.

Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright in
these works, so the Foundation (and you!) can copy and distribute it
in the United States without permission and without paying
copyright royalties. Special rules, set forth in the General Terms of
Use part of this license, apply to copying and distributing Project
Gutenberg™ electronic works to protect the PROJECT GUTENBERG™
concept and trademark. Project Gutenberg is a registered trademark,
and may not be used if you charge for an eBook, except by following
the terms of the trademark license, including paying royalties for use
of the Project Gutenberg trademark. If you do not charge anything
for copies of this eBook, complying with the trademark license is
very easy. You may use this eBook for nearly any purpose such as
creation of derivative works, reports, performances and research.
Project Gutenberg eBooks may be modified and printed and given
away—you may do practically ANYTHING in the United States with
eBooks not protected by U.S. copyright law. Redistribution is subject
to the trademark license, especially commercial redistribution.

START: FULL LICENSE

THE FULL PROJECT GUTENBERG LICENSE
 PLEASE READ THIS BEFORE YOU DISTRIBUTE OR USE THIS WORK

To protect the Project Gutenberg™ mission of promoting the free

distribution of electronic works, by using or distributing this work (or
any other work associated in any way with the phrase “Project
Gutenberg”), you agree to comply with all the terms of the Full
Project Gutenberg™ License available with this file or online at
www.gutenberg.org/license.

Section 1. General Terms of Use and

Redistributing Project Gutenberg™
electronic works
1.A. By reading or using any part of this Project Gutenberg™
electronic work, you indicate that you have read, understand, agree
to and accept all the terms of this license and intellectual property
(trademark/copyright) agreement. If you do not agree to abide by all
the terms of this agreement, you must cease using and return or
destroy all copies of Project Gutenberg™ electronic works in your
possession. If you paid a fee for obtaining a copy of or access to a
Project Gutenberg™ electronic work and you do not agree to be
bound by the terms of this agreement, you may obtain a refund
from the person or entity to whom you paid the fee as set forth in
paragraph 1.E.8.

1.B. “Project Gutenberg” is a registered trademark. It may only be

used on or associated in any way with an electronic work by people
who agree to be bound by the terms of this agreement. There are a
few things that you can do with most Project Gutenberg™ electronic
works even without complying with the full terms of this agreement.
See paragraph 1.C below. There are a lot of things you can do with
Project Gutenberg™ electronic works if you follow the terms of this
agreement and help preserve free future access to Project
Gutenberg™ electronic works. See paragraph 1.E below.
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright law
in the United States and you are located in the United States, we do
not claim a right to prevent you from copying, distributing,
performing, displaying or creating derivative works based on the
work as long as all references to Project Gutenberg are removed. Of
course, we hope that you will support the Project Gutenberg™
mission of promoting free access to electronic works by freely
sharing Project Gutenberg™ works in compliance with the terms of
this agreement for keeping the Project Gutenberg™ name associated
with the work. You can easily comply with the terms of this
agreement by keeping this work in the same format with its attached
full Project Gutenberg™ License when you share it without charge
with others.

1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the
terms of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.

1.E. Unless you have removed all references to Project Gutenberg:

1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project Gutenberg™
work (any work on which the phrase “Project Gutenberg” appears,
or with which the phrase “Project Gutenberg” is associated) is
accessed, displayed, performed, viewed, copied or distributed:
 This eBook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this eBook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

1.E.2. If an individual Project Gutenberg™ electronic work is derived

from texts not protected by U.S. copyright law (does not contain a
notice indicating that it is posted with permission of the copyright
holder), the work can be copied and distributed to anyone in the
United States without paying any fees or charges. If you are
redistributing or providing access to a work with the phrase “Project
Gutenberg” associated with or appearing on the work, you must
comply either with the requirements of paragraphs 1.E.1 through
1.E.7 or obtain permission for the use of the work and the Project
Gutenberg™ trademark as set forth in paragraphs 1.E.8 or 1.E.9.

1.E.3. If an individual Project Gutenberg™ electronic work is posted

with the permission of the copyright holder, your use and distribution
must comply with both paragraphs 1.E.1 through 1.E.7 and any
additional terms imposed by the copyright holder. Additional terms
will be linked to the Project Gutenberg™ License for all works posted
with the permission of the copyright holder found at the beginning
of this work.

1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files containing a
part of this work or any other work associated with Project
Gutenberg™.

1.E.5. Do not copy, display, perform, distribute or redistribute this

electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1
with active links or immediate access to the full terms of the Project
Gutenberg™ License.

1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if you
provide access to or distribute copies of a Project Gutenberg™ work
in a format other than “Plain Vanilla ASCII” or other format used in
the official version posted on the official Project Gutenberg™ website
(www.gutenberg.org), you must, at no additional cost, fee or
expense to the user, provide a copy, a means of exporting a copy, or
a means of obtaining a copy upon request, of the work in its original
“Plain Vanilla ASCII” or other form. Any alternate format must
include the full Project Gutenberg™ License as specified in
paragraph 1.E.1.

1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™ works
unless you comply with paragraph 1.E.8 or 1.E.9.

1.E.8. You may charge a reasonable fee for copies of or providing

access to or distributing Project Gutenberg™ electronic works
provided that:

• You pay a royalty fee of 20% of the gross profits you derive
from the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
 about donations to the Project Gutenberg Literary Archive
Foundation.”

• You provide a full refund of any money paid by a user who

notifies you in writing (or by e-mail) within 30 days of receipt
that s/he does not agree to the terms of the full Project
Gutenberg™ License. You must require such a user to return or
destroy all copies of the works possessed in a physical medium
and discontinue all use of and all access to other copies of
Project Gutenberg™ works.

• You provide, in accordance with paragraph 1.F.3, a full refund of

any money paid for a work or a replacement copy, if a defect in
the electronic work is discovered and reported to you within 90
days of receipt of the work.

• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.

1.E.9. If you wish to charge a fee or distribute a Project Gutenberg™

electronic work or group of works on different terms than are set
forth in this agreement, you must obtain permission in writing from
the Project Gutenberg Literary Archive Foundation, the manager of
the Project Gutenberg™ trademark. Contact the Foundation as set
forth in Section 3 below.

1.F.

1.F.1. Project Gutenberg volunteers and employees expend

considerable effort to identify, do copyright research on, transcribe
and proofread works not protected by U.S. copyright law in creating
the Project Gutenberg™ collection. Despite these efforts, Project
Gutenberg™ electronic works, and the medium on which they may
be stored, may contain “Defects,” such as, but not limited to,
incomplete, inaccurate or corrupt data, transcription errors, a
copyright or other intellectual property infringement, a defective or
damaged disk or other medium, a computer virus, or computer
codes that damage or cannot be read by your equipment.

1.F.2. LIMITED WARRANTY, DISCLAIMER OF DAMAGES - Except for

the “Right of Replacement or Refund” described in paragraph 1.F.3,
the Project Gutenberg Literary Archive Foundation, the owner of the
Project Gutenberg™ trademark, and any other party distributing a
Project Gutenberg™ electronic work under this agreement, disclaim
all liability to you for damages, costs and expenses, including legal
fees. YOU AGREE THAT YOU HAVE NO REMEDIES FOR
NEGLIGENCE, STRICT LIABILITY, BREACH OF WARRANTY OR
BREACH OF CONTRACT EXCEPT THOSE PROVIDED IN PARAGRAPH
1.F.3. YOU AGREE THAT THE FOUNDATION, THE TRADEMARK
OWNER, AND ANY DISTRIBUTOR UNDER THIS AGREEMENT WILL
NOT BE LIABLE TO YOU FOR ACTUAL, DIRECT, INDIRECT,
CONSEQUENTIAL, PUNITIVE OR INCIDENTAL DAMAGES EVEN IF
YOU GIVE NOTICE OF THE POSSIBILITY OF SUCH DAMAGE.

1.F.3. LIMITED RIGHT OF REPLACEMENT OR REFUND - If you

discover a defect in this electronic work within 90 days of receiving
it, you can receive a refund of the money (if any) you paid for it by
sending a written explanation to the person you received the work
from. If you received the work on a physical medium, you must
return the medium with your written explanation. The person or
entity that provided you with the defective work may elect to provide
a replacement copy in lieu of a refund. If you received the work
electronically, the person or entity providing it to you may choose to
give you a second opportunity to receive the work electronically in
lieu of a refund. If the second copy is also defective, you may
demand a refund in writing without further opportunities to fix the
problem.

1.F.4. Except for the limited right of replacement or refund set forth
in paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.

1.F.5. Some states do not allow disclaimers of certain implied

warranties or the exclusion or limitation of certain types of damages.
If any disclaimer or limitation set forth in this agreement violates the
law of the state applicable to this agreement, the agreement shall be
interpreted to make the maximum disclaimer or limitation permitted
by the applicable state law. The invalidity or unenforceability of any
provision of this agreement shall not void the remaining provisions.

1.F.6. INDEMNITY - You agree to indemnify and hold the Foundation,

the trademark owner, any agent or employee of the Foundation,
anyone providing copies of Project Gutenberg™ electronic works in
accordance with this agreement, and any volunteers associated with
the production, promotion and distribution of Project Gutenberg™
electronic works, harmless from all liability, costs and expenses,
including legal fees, that arise directly or indirectly from any of the
following which you do or cause to occur: (a) distribution of this or
any Project Gutenberg™ work, (b) alteration, modification, or
additions or deletions to any Project Gutenberg™ work, and (c) any
Defect you cause.

Section 2. Information about the Mission

of Project Gutenberg™
Project Gutenberg™ is synonymous with the free distribution of
electronic works in formats readable by the widest variety of
computers including obsolete, old, middle-aged and new computers.
It exists because of the efforts of hundreds of volunteers and
donations from people in all walks of life.

Volunteers and financial support to provide volunteers with the

assistance they need are critical to reaching Project Gutenberg™’s
goals and ensuring that the Project Gutenberg™ collection will
remain freely available for generations to come. In 2001, the Project
Gutenberg Literary Archive Foundation was created to provide a
secure and permanent future for Project Gutenberg™ and future
generations. To learn more about the Project Gutenberg Literary
Archive Foundation and how your efforts and donations can help,
see Sections 3 and 4 and the Foundation information page at
www.gutenberg.org.

Section 3. Information about the Project

Gutenberg Literary Archive Foundation
The Project Gutenberg Literary Archive Foundation is a non-profit
501(c)(3) educational corporation organized under the laws of the
state of Mississippi and granted tax exempt status by the Internal
Revenue Service. The Foundation’s EIN or federal tax identification
number is 64-6221541. Contributions to the Project Gutenberg
Literary Archive Foundation are tax deductible to the full extent
permitted by U.S. federal laws and your state’s laws.

The Foundation’s business office is located at 809 North 1500 West,

Salt Lake City, UT 84116, (801) 596-1887. Email contact links and up
to date contact information can be found at the Foundation’s website
and official page at www.gutenberg.org/contact

Section 4. Information about Donations to

the Project Gutenberg Literary Archive
Foundation
Project Gutenberg™ depends upon and cannot survive without
widespread public support and donations to carry out its mission of
increasing the number of public domain and licensed works that can
be freely distributed in machine-readable form accessible by the
widest array of equipment including outdated equipment. Many
small donations ($1 to $5,000) are particularly important to
maintaining tax exempt status with the IRS.

The Foundation is committed to complying with the laws regulating

charities and charitable donations in all 50 states of the United
States. Compliance requirements are not uniform and it takes a
considerable effort, much paperwork and many fees to meet and
keep up with these requirements. We do not solicit donations in
locations where we have not received written confirmation of
compliance. To SEND DONATIONS or determine the status of
compliance for any particular state visit www.gutenberg.org/donate.

While we cannot and do not solicit contributions from states where

we have not met the solicitation requirements, we know of no
prohibition against accepting unsolicited donations from donors in
such states who approach us with offers to donate.

International donations are gratefully accepted, but we cannot make

any statements concerning tax treatment of donations received from
outside the United States. U.S. laws alone swamp our small staff.

Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.

Section 5. General Information About

Project Gutenberg™ electronic works
Professor Michael S. Hart was the originator of the Project
Gutenberg™ concept of a library of electronic works that could be
freely shared with anyone. For forty years, he produced and
distributed Project Gutenberg™ eBooks with only a loose network of
volunteer support.
Project Gutenberg™ eBooks are often created from several printed
editions, all of which are confirmed as not protected by copyright in
the U.S. unless a copyright notice is included. Thus, we do not
necessarily keep eBooks in compliance with any particular paper
edition.

Most people start at our website which has the main PG search
facility: www.gutenberg.org.

This website includes information about Project Gutenberg™,

including how to make donations to the Project Gutenberg Literary
Archive Foundation, how to help produce our new eBooks, and how
to subscribe to our email newsletter to hear about new eBooks.
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com