Full download test bank at ebook ebookmass.

com

Numerical Solutions of Three

Classes of Nonlinear Parabolic
Integro-Differential Equations 1st
Edition Jangveladze
CLICK LINK TO DOWLOAD

https://ebookmass.com/product/numerical-
solutions-of-three-classes-of-nonlinear-
parabolic-integro-differential-equations-1st-
edition-jangveladze/

ebookmass.com
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Asymptotic Perturbation Methods: For Nonlinear

Differential Equations in Physics Attilio Maccari

https://ebookmass.com/product/asymptotic-perturbation-methods-
for-nonlinear-differential-equations-in-physics-attilio-maccari/

Schaum's Outline of Differential Equations, Fifth

Edition Bronson

https://ebookmass.com/product/schaums-outline-of-differential-
equations-fifth-edition-bronson/

Boundary value problems for systems of differential,

difference and fractional equations : positive
solutions 1st Edition Johnny Henderson

https://ebookmass.com/product/boundary-value-problems-for-
systems-of-differential-difference-and-fractional-equations-
positive-solutions-1st-edition-johnny-henderson/

First Course in differential equations (11ed) /

Differential Equations and Boundary Value Problems
(9ed) Solutions manual Dennis Zill Roberto Martinez

https://ebookmass.com/product/first-course-in-differential-
equations-11ed-differential-equations-and-boundary-value-
problems-9ed-solutions-manual-dennis-zill-roberto-martinez/
Advanced Differential Equations 1st Edition Youssef
Raffoul

https://ebookmass.com/product/advanced-differential-
equations-1st-edition-youssef-raffoul/

Nonlinear Differential Equations in Micro/Nano

Mechanics: Application in Micro/Nano Structures and
Electromechanical Systems 1st Edition Ali Koochi

https://ebookmass.com/product/nonlinear-differential-equations-
in-micro-nano-mechanics-application-in-micro-nano-structures-and-
electromechanical-systems-1st-edition-ali-koochi/

Techniques of Functional Analysis for Differential and

Integral Equations 1st Edition Paul Sacks

https://ebookmass.com/product/techniques-of-functional-analysis-
for-differential-and-integral-equations-1st-edition-paul-sacks/

Mathematical Physics with Differential Equations

Yisong. Yang

https://ebookmass.com/product/mathematical-physics-with-
differential-equations-yisong-yang/

Transforms and Partial Differential Equations T.

Veerarajan

https://ebookmass.com/product/transforms-and-partial-
differential-equations-t-veerarajan/
Numerical Solutions of Three Classes of Nonlinear
Parabolic Integro-Differential Equations
Numerical Solutions of Three Classes
of Nonlinear Parabolic Integro-Differential
Equations

Temur Jangveladze
Ilia Vekua Institute of Applied Mathematics of
Ivane Javakhishvili Tbilisi State University
Tbilisi, Georgia
&
Georgian Technical University
Tbilisi, Georgia

Zurab Kiguradze
Ilia Vekua Institute of Applied Mathematics of
Ivane Javakhishvili Tbilisi State University
Tbilisi, Georgia

Beny Neta
Naval Postgraduate School
Department of Applied Mathematics
Monterey, CA, U.S.A.
Academic Press is an imprint of Elsevier
32 Jamestown Road, London NW1 7BY, UK
225 Wyman Street, Waltham, MA 02451, USA
525 B Street, Suite 1800, San Diego, CA 92101-4495, USA
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

Copyright © 2016 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any

means, electronic or mechanical, including photocopying, recording, or any information
storage and retrieval system, without permission in writing from the publisher. Details
on how to seek permission, further information about the Publisher’s permissions
policies and our arrangements with organizations such as the Copyright Clearance
Center and the Copyright Licensing Agency, can be found at our website:
www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright
by the Publisher (other than as may be noted herein).

Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional
practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge
in evaluating and using any information, methods, compounds, or experiments
described herein. In using such information or methods they should be mindful of their
own safety and the safety of others, including parties for whom they have a professional
responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of
any methods, products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-804628-9

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

For information on all Academic Press Publications

visit our website at http://store.elsevier.com/
PREFACE

This book is concerned with the numerical solutions of some

classes of nonlinear integro-di erential models. Some properties
of the solutions for investigated equations are also given. Three
types of nonlinear integro-di erential models are considered. Al-
gorithms of nding approximate solutions are constructed and
investigated. Results of numerical experiments with graphical
illustrations and their analysis are given. The book consists of
four Chapters.
In the rst Chapter three models (Model I, Model II and
Model III, that will be detailed further) to be discussed are
introduced and a brief history of integro-di erential equations
is given.
In the second Chapter, mathematical modeling of a process
of penetration of an electromagnetic eld into a substance by
integro-di erential models is described. A short description of
the integro-di erential equation that is a special model for one-
dimensional heat ow in materials with memory is also given
in the second Chapter. This model arises in the theory of one-
dimensional viscoelasticity as well. This Chapter closes with
some concluding remarks of the three investigated models.
The third Chapter is devoted to the numerical solution of
the initial-boundary value problems for models stated in the
previous Chapter. Semi-discrete schemes and nite-di erence
approximations, as well as nite elements are discussed. The
mathematical substantiation of all these questions for initial-
boundary value problems is given.
The questions of the realizations of algorithms investigated
in the third Chapter are discussed in the fourth Chapter. Re-
sults of the many numerical experiments with graphical illustra-
tions and their analysis are also given in this Chapter.
viii PREFACE

At the end of the book a list of the quoted literature and

indexes are given. The list of references is not intended to be
an exhaustive bibliography on the subject, but it is nevertheless
detailed enough to enable further independent work.
Each Chapter is concluded with a detailed section, entitled
"Comments and bibliographical notes," containing references to
the principal results treated, as well as information on important
topics related to, but sometimes not included in the body of the
text.
The authors believe that the book will be useful to scientists
working in the eld of nonlinear integro-di erential models. In
the opinion of the authors, the material presented in the book
is helpful for a wide range of readers engaged in mathematical
physics, in problems of applied and numerical mathematics, and
also MS and PhD students of the appropriate specializations.

Temur Jangveladze,
Zurab Kiguradze,
Beny Neta
ACKNOWLEDGMENTS

The rst author thanks Fulbright Visiting Scholar Program for

giving him the opportunity to visit U.S.A. and the Naval Post-
graduate School in Monterey, CA, U.S.A. for hosting him during
the nine months of his tenure in 2012-2013. The second author
thanks Shota Rustaveli National Scienti c Foundation of Re-
public of Georgia for giving him opportunity to visit U.S.A.
and the Naval Postgraduate School in Monterey, CA, U.S.A. for
hosting him during the four months of his tenure in 2013.
Abstract

This book is concerned with the numerical solutions of some

classes of nonlinear integro-di erential models. Some proper-
ties of the solutions of the corresponding initial-boundary value
problems studied in the monograph equations are given. Three
types of nonlinear integro-di erential models are considered. Al-
gorithms of nding approximate solutions are constructed and
investigated. Results of numerical experiments with tables and
graphical illustrations and their analysis are given. The book
consists of four chapters. At the end of the book a list of the
quoted literature and indexes are given. Each chapter is con-
cluded with a detailed section, entitled "Comments and biblio-
graphical notes," containing references to the principal results
treated, as well as information on important topics related to,
but sometimes not included in the body of the text.

Numerical Solutions of Three Classes of Nonlinear Parabolic Integro-Differential Equation.

http://dx.doi.org/10.1016/B978-0-12-804628-9.50008-9
© 2016 Elsevier Ltd. All rights reserved. 1
2 ABSTRACT

Key words: Electromagnetic eld penetration, Maxwell’s

equations, integro-di erential models, existence and uniqueness,
asymptotic behavior, semi-discrete and nite di erence schemes,
Galerkin’s method, nite element approximation, error estimate,
stability and convergence.
Chapter 1

Introduction

Abstract
The description of various kinds of integro-di erential equa-
tions and a brief history of their origin and applications are
given. The importance of investigations of integro-di erential
models is pointed out as well. Classi cation of integro-di eren-
tial equation is given. The main attention is paid on parabolic
type integro-di erential models. In particular, three types of
integro-di erential equations are considered. Two of them are
based on Maxwell’s equations describing electromagnetic eld
penetration into a substance. The third one is obtained by
simulation of heat ow. At the end of the chapter, as at the
end of each chapter, the comments and bibliographical notes is
given, which consists of description of references concerning to
the topic considered.

Key words: Electromagnetic eld penetration, Maxwell’s

system, heat ow equation, integro-di erential models.

Numerical Solutions of Three Classes of Nonlinear Parabolic Integro-Differential Equation.

http://dx.doi.org/10.1016/B978-0-12-804628-9.50001-6
© 2016 Elsevier Inc. All rights reserved. 3
4 CHAPTER 1. INTRODUCTION

In mathematical modeling of applied tasks di erential, inte-

gral, and integro-di erential (I-D, for short) equations appear
very often. There are numerous scienti c works devoted to the
investigation of di erential equations. There is a vast literature
in the eld of integral and integro-di erential models as well.
The di erential equations are connecting unknown functions,
their derivatives, and independent variables. On the other hand,
integral equations contain the unknown functions under an in-
tegral as well.
The term integro-di erential equation in the literature is
used in the case when the equation contains unknown function
together with its derivatives and when either unknown function,
or its derivatives, or both appear under an integral.
Let us recall the general classi cation of integro-di erential
equations. If the equation contains derivatives of unknown func-
tion of one variable then the integro-di erential equation is called
ordinary integro-di erential equation. The order of an equation
is the same as the highest-order derivative of the unknown func-
tion in the equation.
The integro-di erential equations often encountered in math-
ematics and physics contain derivatives of various variables;
therefore, these equations are called integro-di erential equa-
tions with partial derivatives or partial integro-di erential equa-
tions.
In the applications very often there are integro-di erential
equations with partial derivatives and multiple integrals as well,
for example, Boltzmann equation [66] and Kolmogorov-Feller
equation [288].
Volterra is one of the founders of the theory of integral and
integro-di erential equations. His works, especially in the inte-
gral and integro-di erential equations, are often cited till today.
The classical book by Volterra [469] is widely quoted in the liter-
ature. In 1884 Volterra [465] began his research in the theory of
CHAPTER 1. INTRODUCTION 5

integral equations devoted to distribution of an electrical charge

on a spherical patch. This work led to the equation, which in
the modern literature is called the integral equation of the rst
kind with symmetric kernel.
The work on the theory of elasticity became the beginning
research of Volterra leading to the theory of partial integro-
di erential equations. In 1909 Volterra [466] has studied a par-
ticular type of such equations and has shown that this integro-
di erential equation is equivalent to a system consisting of three
linear integral equations and a second order partial di erential
equations.
The rst examples of integro-di erential equations with par-
tial derivatives investigated in the beginning of the twentieth
century were in Schlesinger’s works [417], [418], where the fol-
lowing equation is investigated:
Z b
@U (x; y)
= f (x; y; s)U (x; s)ds:
@x a
Numerous works in the beginning of the twentieth century
were devoted to research of integro-di erential equations of vari-
ous kinds. The excellent bibliography in this case is given in the
classical book by Volterra [469]. In addition, Kerimov [271], the
editor of the Russian translation of this book, has updated (up
to 1970s) the list of references on integral and integro-di erential
equations.
Let us describe some classes of mathematical models of sec-
ond order promoting intensive research on partial integro-di er-
ential equations.
When we take into account hereditary phenomena, the ques-
tions of physics and mechanics lead to integro-di erential equa-
tions. A hereditary phenomenon occurs in a system when the
phenomenon does not depend only on the actual state of the
system but on all the preceding states through which the sys-
6 CHAPTER 1. INTRODUCTION

tem has passed; that is to say, it depends on the history of the

system and may therefore be called hereditary.
One of the important representatives of an integro-di er-
ential equations of elliptic type is the following equation con-
nected with the hereditary phenomenon [467]
Z tX3
@ 2 U (x; )
U (x; t) + fi (t; )d = 0;
0 i=1 @x2i

where x = (x1 ; x2 ; x3 ); is a classical three-dimensional Laplace

operator

@ 2 U (x; t) @ 2 U (x; t) @ 2 U (x; t)

U (x; t) = + + ;
@x21 @x22 @x23
and fi are known functions of their arguments. Let us note that
here and below everywhere instead of x; y; z we use x1 ; x2 ; x3 for
the designation of Cartesian coordinates.
The mathematical modeling of the vibrations of an elastic
chord in the case of linear hereditary process gives a hyperbolic
type integro-di erential equation [467]

t
@ 2 U (x; t) @ 2 U (x; t) @ 2 U (x; )
Z
= + (t; )d ; (1.1)
@t2 @x2 0 @x2

where is a known function of its arguments.

One of the most important representatives of integro-di er-
ential models is the following nonlinear equation describing string
vibration obtained by Kirchho [286] in 1876

" #
2
@ 2 U (x; t) @ 2 U (x; t)
Z
2 @U (x; t)
+ dx = 0; (1.2)
@t2 0 @x @x2
CHAPTER 1. INTRODUCTION 7

where is a known constant. Many authors investigated equa-

tion (1.2) and its natural generalizations:

"Z #
2
@ 2 U (x; t) @U (x; t) @ 2 U (x; t)
a dx =0 (1.3)
@t2 0 @x @x2

and

@ 2 U (x; t) 1
a A 2 U (t) AU (t) = 0; (1.4)
@t2
where a(S) a0 = const > 0 is a known function of its argu-
ment and A is a self-adjoint positive operator, i.e., A = A > 0.
The norm used in (1.4) is the one de ned on the range of the
operator A.
In investigating (1.3) and (1.4) type models it is su cient
to mention the following publications: [9], [24], [36], [37], [55],
[61], [65], [130], [309], [327], [328], [349], [355], [356], [357], [373],
[380], [383], [390], [394], [395] though this list is not complete.
Let us note that equations (1.3) and (1.4) are also called Kirch-
ho equations. They, along with some similar equations, de-
scribe important physical processes, among which are linear and
nonlinear dynamics of di erent dimensional bodies (see, for ex-
ample, [36], [355], [373] and [471]).
In other questions connected with hereditary phenomena,
one nds the integro-di erential equations of a parabolic type,
which were investigated by Evans [160]. These equations look
like
Z t 2
@U (x; t) @ 2 U (x; t) @ U (x; )
2
+ A(t; )d = 0; (1.5)
@t @x t0 @x2

where A is a known function of its arguments.

8 CHAPTER 1. INTRODUCTION

Integro-di erential equations of parabolic type arise in the

study of various problems in physics, chemistry, technology,
economics, etc. One very important problem of applied type
is generated by mathematical modeling of processes of electro-
magnetic eld penetration into a substance and is described by
the well-known Maxwell’s equations [300]. In the works [187]
and [188], complex system corresponding to nonlinear partial-
di erential equations was reduced to integro-di erential form. If
the coe cient of thermal heat capacity and electroconductivity
of the substance are highly dependent on temperature, then the
Maxwell’s system can be rewritten in the following form (see
[187] and [188]):

@W (x; t)
+r [a(S(x; t))r W (x; t)] = 0;
@t (1.6)
r W (x; t) = 0;
where Z t
S(x; t) = jr W (x; )j2 d : (1.7)
0
In system (1.6), r W and r W are the usual vector oper-
ators with respect to the variables x = (x1 ; x2 ; x3 ): Even one-
dimensional scalar version of this model is very complicated and
its investigation has been possible yet only for special cases. The
one-dimensional scalar case of the model (1.6), (1.7) has the fol-
lowing form

" Z ! #
t 2
@U (x; t) @ @U (x; ) @U (x; t)
= a d ; (1.8)
@t @x 0 @x @x

where a(S) a0 = const > 0 is again a known function of its

argument. Investigation of (1.6), (1.7), and (1.8) type models
CHAPTER 1. INTRODUCTION 9

began in the works [138], [187], and [188]. Since then many
scienti c publications were devoted to the investigation of exis-
tence and uniqueness of their solutions under various types of
initial and boundary conditions. In this respect, especially sig-
ni cant are the works [49], [50], [141], [146], [147], [219], [220],
[238], [254], [256], [302], [303], [304], [305], [306], [307], [322],
[333], [334], and reference therein. Authors of this book have
also made contribution in this direction, for example, see [137],
[138], [139], [140], [141], [145], [146], [147], [186], [187], [188],
[219], [220], [223], [238], [247], [248], [253], [254], [256], [261],
and reference therein.
Making certain physical assumptions in mathematical de-
scription of the above-mentioned process of penetration of elec-
tromagnetic eld into a substance, Laptev [306] has constructed
a new integro-di erential model, which represents a general-
ization of the system introduced in [187] and [188]. Founded
on Maxwell’s system the following parabolic integro-di erential
model is obtained

Z tZ
@W (x; t)
=a jr W (x; )j2 dxd W (x; t): (1.9)
@t 0

In the works [303], [305], and [306] for conditionally closed

operators, an operator scheme is constructed. This scheme is
used for (1.6), (1.7) type models to prove existence and unique-
ness of solution of initial-boundary value problems. In the above-
mentioned work [306] Laptev points out that for these so-called
averaged integro-di erential models (1.9), it is necessary to de-
velop a di erent, special approach.
In systems (1.6), (1.7), and (1.9), W = (W1 ; W2 ; W3 ) denotes
a vector, which is connected by a vector of a magnetic eld H =
(H1 ; H2 ; H3 ) and is the function of the variables (x1 ; x2 ; x3 ; t);
which we will shorten to (x; t).
10 CHAPTER 1. INTRODUCTION

The multidimensional scalar analogues of systems (1.6), (1.7),

and (1.9) have the following forms:

Z t
@U (x; t)
=r a jrU (x; )j2 d rU (x; t) (1.10)
@t 0

and
Z tZ
@U (x; t)
=a jrU (x; )j2 dxd U (x; t); (1.11)
@t 0

respectively.
In equations (1.10) and (1.11) we have x = (x1 ; : : : ; xn ) and
the vector operator rU is given by

@U (x; t) @U (x; t)
rU = gradU = ;:::; = (D1 U; : : : ; Dn U ) :
@x1 @xn
Some generalizations of the models (1.6)-(1.11) are given in
the works [137], [139], [145], [219], [305], and [306]. One kind of
these models has the forms:

n
@U (x; t) X
Di a(S(x; t)) jrU (x; t)jq 2
Di U (x; t)
@t i=1 (1.12)

= f (x; t)
and

n
@U (x; t) X
a(S(t)) Di jrU (x; t)jq 2
Di U (x; t)
@t i=1 (1.13)

= f (x; t);
CHAPTER 1. INTRODUCTION 11

where Z t
S(x; t) = jrU (x; )jq d
0
and Z tZ
S(t) = jrU (x; )jq d :
0
The models of type (1.6)-(1.13) are complex and have been
intensively studied by many authors. The existence and unique-
ness of global solutions of initial-boundary value problems for
equations and systems of type (1.6) - (1.13) were studied in
[49], [50], [137], [138], [139], [140], [141], [145], [186], [187],[188],
[219], [223], [238], [247], [248], [253], [261], [302], [303], [304],
[305], [306], [307], [322], [333], [334], and in a number of other
works as well.
The existence theorems that are proved in [137], [138], [139],
[140], [141], [145], [187], [188], and [219] are based on a-priori
estimates, modi ed Galerkin’s method and compactness argu-
ments as in [327], [328], [461], and [462] for nonlinear elliptic
and parabolic equations.
For equation (1.8) with nonhomogeneous right-hand side and
a(S) = 1 + S, or for equation (1.13) in the one-dimensional
case with q = 2, such theorem for rst initial-boundary value
problem is proved in section 3.6.
The asymptotic behavior as t ! 1 of the solutions of such
models has been the object of intensive research in recent years,
see [28], [29], [31], [32], [33], [34], [35], [139], [141], [145], [146],
[147], [217], [218], [219], [220], [223], [227], [228], [229], [231],
[232], [233], [234], [235], [236], [237], [238], [239], [240], [241],
[242], [243], [245], [246], [247], [248], [251], [252], [253], [254],
[256], [257], [261], [276], [277], [278], [280], [282], [283], and ref-
erence therein.
Another model considered in this book and studied by one
of the authors of this monograph [376] is
12 CHAPTER 1. INTRODUCTION

Z t
@U (x; t) @ @U (x; )
= a(t ) d + f (x; t): (1.14)
@t 0 @x @x

This equation serves as a very special model for one-dimens-

ional heat ow in material with memory [196], [345]. It also
arises in the theory of viscoelasticity [109], [344], [346]. It is an
example in the general theory of equations of the form
Z t
U_ (t) + a(t )g(U ( ))d = f (t) (1.15)
0

on a Hilbert space H with g a nonlinear bounded operator [51].

The dot above U means time derivative.
MacCamy [345] and later Sta ans [438] have discussed the
existence, uniqueness, boundedness and asymptotic behavior of
solutions of initial-boundary value problems for (1.14).
Many scienti c publications are dedicated to the investiga-
tion of characteristics of (1.14) and (1.15) type equations (see,
for example, [67], [70], [80], [110], [111], [162], [206], [433], and
reference therein).
Numerous scienti c works are devoted to construction of al-
gorithms for the numerical solution of initial-boundary value
problems for the above-stated models (see, for example, [20],
[30], [32], [35], [80], [96], [125], [139], [145], [148], [220], [223],
[227], [239], [241], [243], [244], [247], [248], [249], [250], [251],
[252], [253], [255], [256], [257], [258], [261], [274], [276], [277],
[279], [280], [281], [282], [283], [284], [318], [333], [376], [378],
[423], [424], and [501]).
The purpose of the present monograph is the continuation
of study and uniform description of results developed for the
numerical solution of integro-di erential models (1.6)-(1.14) and
their generalizations. It should be noted that main attention is
CHAPTER 1. INTRODUCTION 13

paid on theorems with proofs for corresponding one-dimensional

equations.
The book consists of four chapters. In the next chapter
mathematical modeling of a process of penetration of an elec-
tromagnetic eld into a substance by integro-di erential models
and the conclusion of investigated models such as (1.6)-(1.13)
is brie y given. We also give the mathematical model for heat
conduction in materials with memory (1.14). Some features of
this model are also given.
The third chapter is devoted to the numerical solution of
the initial-boundary value problems for the models stated in
the previous chapter. The semi-discrete schemes, nite di er-
ence approximations, Galerkin’s method schemes, and nite el-
ements are discussed. The mathematical substantiation of all
these questions for various types of initial-boundary value prob-
lems is given.
Results of the various numerical experiments with graphical
illustrations and their analyses are given in the fourth chapter.
At the end of the book a list of the quoted literature and
indexes are given.
The list of references is not intended to be an exhaustive bib-
liography on the subject, but it is nevertheless detailed enough
to enable further independent work.
Each chapter is concluded with a detailed section, entitled
"Comments and bibliographical notes," containing references to
the principal results treated, as well as information on important
topics related to, but sometimes not included in the body of the
text.
The authors believe that the book will be useful to the ex-
perts working in the eld of nonlinear integro-di erential mod-
els. In the opinion of the authors, the material presented in the
book is helpful for a wide range of readers engaged in mathemat-
ical physics, in problems of applied and numerical mathematics,
14 CHAPTER 1. INTRODUCTION

and also MS and PhD students of the appropriate specializa-

tions.

1.1 Comments and bibliographical

notes
Mathematical models of many natural phenomena and processes
can be described by the initial-boundary value problems posed
for nonstationary partial di erential and integro-di erential equa-
tions and systems of such equations. Investigation and numer-
ical solution of these problems are the actual sphere of mathe-
matical physics and numerical analysis. One such partial integro-
di erential model describes the process of electromagnetic eld
penetration into a substance. In the quasi-stationary approxi-
mation, the corresponding system of Maxwell’s partial di eren-
tial equations can be rewritten in integro-di erential form (1.6),
(1.7) (see [187] and [188]).
Mathematical models describing electromagnetic processes
and many relative phenomena are given in many scienti c pa-
pers and books, see, for example, [92], [136], [300], [359], [360],
[399], [419], [473], and references therein.
It is well known that electromagnetic eld di usion pro-
cesses and many other important practical processes are simu-
lated by Maxwell’s systems of partial di erential equations and
Maxwell’s-type systems as well (see, for example, [195], [197],
[290], [297], [298], [300], [360], [426], [485], [486], [487], [489],
[491], and references therein).
Many scienti c works are devoted to investigation of various
problems for Maxwell’s and Maxwell’s-type systems, see [4], [5],
[56], [60], [101], [106], [120], [121], [124], [131], [139], [144], [145],
[152], [158], [194], [195], [197], [208], [224], [225], [226], [247],
[253], [275], [276], [290], [362], [365], [367], [397], [409], [427],
1.1. COMMNETS AND BIBLIOGRAPHICAL NOTES 15

[452], [453], [454], [470], [475], [476], [477], [485], [486], [487],
[489], [491], and references therein.
As we have already pointed out, by using Maxwell’s system
[300] for mathematical simulation of the process of electromag-
netic eld penetrating into a substance, new class of integro-
di erential models (1.6), (1.7), (1.8), and (1.10) arises (see [187]
and [188]).
Integro-di erential equations arise in many other practical
processes as well, see, for example, [11], [21], [22], [36], [37],
[55], [59], [74], [84], [119], [127], [128], [164], [185], [187], [188],
[295], [298], [306], [310], [311], [312], [355], [356], [357], [363],
[373], [380], [384], [405], [425], [459], [463], [465], [466], [467],
[468], [469], [471], [484], and in a number of other works as well.
The motivation for studying integro-di erential problems
comes from the many physical models in such elds as elec-
tromagnetic wave propagation, heat transfer, nuclear reactor
dynamics, and thermoelasticity. Besides the integro-di erential
equations arise in many spheres of human activity as well. For
example, the second order fully nonlinear integro-di erential
equations are derived from the pricing problem of nancial deriva-
tives and optimal portfolio selection problem in a market [59].
In [84] nonlinear integro-di erential equations that arise from
stochastic control problems with purely jump Levy processes
are considered.
Many problems of modern science and engineering can be
described by partial integro-di erential equations. Since quite
a lot of these problems are time-dependent, most of them are
evolution equations and especially nonlinear evolution parabolic
equations, see [93], [177], [179], [180], [191], [216], [310], [311],
[312], [322], [332], [345], [346], [358], [451], and references therein).
Many scienti c works are devoted to investigation and nu-
merical solution of parabolic integro-di erential models, see [6],
[18], [45], [53], [62], [63], [69], [74], [83], [86], [87], [93], [98], [103],
16 CHAPTER 1. INTRODUCTION

[108], [115], [118], [128], [137], [139], [145], [155], [156], [163],
[175], [177], [181], [182], [191], [198], [247], [253], [263], [265],
[266], [267], [276], [279], [281], [306], [308], [310], [311], [312],
[313], [319], [320], [322], [323], [327], [353], [358], [370], [376],
[378], [389], [393], [427], [430], [432], [433], [447], [448], [449],
[450], [451], [456], [479], [483], [488], [492], [497], and references
therein.
Study of the models of type (1.6), (1.7), and (1.8) has begun
in the works [138], [187] and [188]. In these works, in particu-
lar, the theorems of existence of solution of the initial-boundary
value problem (with rst (Dirichlet) boundary conditions) for
scalar equation with one-dimensional space variable are proved.
Investigations of higher space dimensions for model (1.10) car-
ried out initially in [137] and [140].
In [306] some generalization of the system of type (1.6), (1.7)
was proposed. In particular, assuming that the temperature
of the considered body depends on the time variable, but in-
dependent of the space coordinates, then the same process of
penetration of a magnetic eld into the material is simulated
by the averaged (as the author of [306] has named it) integro-
di erential models (1.9) and (1.11).
Study of the models of type (1.9) and (1.11) has started in
the works [217] and [219].
One must note that some works were devoted to the study
of modeling of physical process of electromagnetic eld pene-
tration in the case of cylindrical conductors. In this case, the
above-mentioned integro-di erential model (1.6), (1.7), written
in cylindrical coordinates, was given in [148]. The work [333]
is devoted to the investigation of periodic problem for one-
dimensional (1.8) type model in cylindrical coordinates.
Interest in the above-mentioned integro-di erential (1.6) -
(1.11) models is increasing. Some generalizations of (1.10) and
(1.11) models, which have the forms (1.12) and (1.13) corre-
1.1. COMMNETS AND BIBLIOGRAPHICAL NOTES 17

spondingly, are given and studied in [219], [305], and in a num-

ber of other works as well. Equation (1.12) is investigated, for
example, in [145], [219], [305], and [306]. Equation (1.13) was
investigated in [145] and [219]. In the scienti c literature some
more general models have also appeared, see [137], [139], [145],
[219], [305], [306], and [322].
Many di erent kinds of initial-boundary value problems with
a variety of boundary and initial conditions are considered for
the above-mentioned integro-di erential equations. In the works
[246], [280], [281] investigation and numerical approximation of
problems with mixed boundary conditions, for (1.10) and (1.11)
type one-dimensional scalar models, are studied.
Let us also note that rst kind initial-boundary value prob-
lems with nonhomogeneous boundary condition on one side of
lateral boundary are also considered and studied in many works.
This type of the problem statement is dictated by mathematical
simulation of the physical processes, see, for example, [148] and
problem (2.64), (2.65) given in "Comments and bibliographical
notes" section in Chapter 2.
The theorems proved for investigating the asymptotic behav-
ior as time tends to in nity in some cases show the di erence be-
tween stabilization character of solutions with homogeneous and
nonhomogeneous boundary conditions of the rst kind initial-
boundary value problems. More precisely, in homogeneous case
stabilization has an exponential character, whereas in nonho-
mogeneous case it has power-like form.
The works [49], [50] are also worth mentioning, where in-
vestigation of inverse problems for multidimensional models of
(1.10) type is carried out.
Another integro-di erential model studied in this monograph
is (1.14). This model describes heat ow in material with mem-
ory [196], [345]. It also arises in the theory of viscoelasticity
[109], [344], and [346].
18 CHAPTER 1. INTRODUCTION

As a rule we cannot nd exact solutions of the considered

nonlinear di erential and integro-di erential models. Therefore,
particular attention should be paid to construction of numerical
solutions and to their importance for integro-di erential models.
The rst steps in this direction, for the models studied in this
monograph, are made in the works [139], [220], [274], [376], and
[378]. Now the research in this direction has intensi ed.
Let us note that the models being considered in this mono-
graph have arisen from practical tasks. But they can be con-
sidered as models, generalizing known nonlinear parabolic equa-
tions, which are studied in many known scienti c papers, books,
and monographs, see [76], [133], [134], [169], [199], [298], [327],
[332], [381], [385], [461], and references therein.
Many scienti c researches are devoted to numerical solution
of partial di erential and nonlinear parabolic equations as well,
see [23], [43], [104], [167], [171], [183], [193], [210], [213], [225],
[226], [301], [327], [400], [404], [407], [411], [415], [435], [439],
[440], [447], and the references listed in these papers and books.
As we have already mentioned the main part of integro-
di erential structures considered here has arisen from Maxwell’s
systems of partial di erential equations. There are many scien-
ti c papers and books on the numerical solution of Maxwell’s
systems and Maxwell’s-type systems as well, see [4], [5], [15],
[26], [27], [54], [68], [139], [142], [145], [153], [212], [224], [247],
[253], [260], [276], [314], [315], [316], [317], [336], [341], [369],
[434], [442], [472], [480], [493], [498], and references therein.
The results of these researches very often are used in construc-
tion and investigation of the numerical algorithms for the cor-
responding integro-di erential models.
The detailed description of investigation and numerical so-
lution of the above-mentioned models is given in Chapters 2-4.
More complete references and comments are given in a section
entitled "Comments and bibliographical notes" in each chapter.
Chapter 2

Mathematical Modeling

Abstract
The chapter consists of six sections. The chapter concerns
mathematical modeling of the investigated in the monograph
equations. Some mathematical features of these models are
studies as well. In the rst section general statement of di usion
process is given. Sections two and three are dedicated to a re-
duction of Maxwell’s equations to the integro-di erential mod-
els. Consequently, two types of integro-di erential equations
are obtained which are called Model I and Model II accordingly.
Both models in di erent physical assumptions describe process
of penetration of an electromagnetic eld into a substance. In
the third section Model III is considered, which represents a
special model for one-dimensional heat ow in materials with
memory. This model arises in the theory of one-dimensional
viscoelasticity as well. Next two sections are devoted to some
mathematical features of all three models above. The existence
and uniqueness properties of the solutions as well as asymp-
totic behavior of solutions of the appropriate initial-boundary
value problems are presented. This chapter closes with some
concluding remarks and bibliographical overview of the three

Numerical Solutions of Three Classes of Nonlinear Parabolic Integro-Differential Equation.

http://dx.doi.org/10.1016/B978-0-12-804628-9.50002-8
© 2016 Elsevier Inc. All rights reserved. 19
20 CHAPTER 2. MATHEMATICAL MODELING

investigated models.

Key words: Mathematical modeling, electromagnetic eld

penetration, Maxwell’s equations, heat ow in materials with
memory, one-dimensional viscoelasticity, integro-di erential mod-
els, existence and uniqueness, asymptotic behavior.

2.1 Electromagnetic di usion process

2.1.1 General statement of di usion process
Let us consider the phenomena which occur in a substance
placed in an external variable electromagnetic eld.
Suppose the electromagnetic eld and the currents satisfy
quasi-stationary conditions, see, e.g., [300] and [419]. Maxwell’s
system of equations is given by:

1 @( H)
= r E; (2.1)
c @t
r ( H) = 0; (2.2)
4
E = r H; (2.3)
c
where E = (E1 ; E2 ; E3 ) and H = (H1 ; H2 ; H3 ) are vectors of
electrical and magnetic elds, respectively, is the magnetic
permeability, is the conductivity of substance, and c is the
speed of light in vacuum.
In (2.3), following an assumption of quasi-stationarity, the
term proportional to @E=@t is omitted and Ohm’s law is used,
relating the vector E to a vector of density of a current J by
the following relation

J = E: (2.4)
2.1. ELECTROMAGNETIC DIFFUSION PROCESS 21

If at some moment of time the magnetic eld is solenoidal,

i.e., r ( H) = 0, then (2.2) is always satis ed. From (2.1) we
have
@
r ( H) = 0:
@t
We shall assume that the substance is isotropic with = 1
and the coe cient of electroconductivity depends on tempera-
ture, i.e., = ( ). In applications, the form of is power-
1
like, for example, for metals ; for homogeneous plasma
3=2
, and so on.
For de nition of temperature it is necessary to use the equa-
tion of balance of heat. Let us make the assumption that the
characteristic time of resistive di usion is much less than that of
heat transfer. Then, neglecting the e ect of heat conductivity,
the change in temperature is de ned only by Joule’s heating and
taking into account (2.4), we have

@
cv = EJ = E 2 ; (2.5)
@t
where cv is the speci c heat capacity of the environment. Thus
the factor of heat capacity can also depend on temperature (as
a rule, in a power-like form).
Equations (2.1), (2.3), (2.5) form the closed system for def-
inition of an electromagnetic eld and temperature under an
appropriate initial and boundary conditions.
The questions of existence and uniqueness of the solutions of
linear di erential problems ( = (x)), in general enough state-
ment, are considered in [136], [297], and in a number of other
works. Thus, in [297] the movement of environment is supposed
within the framework of magnetic-hydrodynamical approach. In
these works the transition from classical statement to the gener-
alized one is made. The requirement that the functions satisfy
equation (2.2) and boundary conditions is replaced with the re-
22 CHAPTER 2. MATHEMATICAL MODELING

quirement of their belonging to special functional spaces. The

problem for equations (2.1), (2.3) is formulated in the variational
form, in particular, in terms of variational inequalities.
The study of system (2.1)-(2.3) can be made also on the
basis of the equations which have been written down only for
vectors E or H. As an example of similar research, see [504], in
which the nonlinearity caused by dependence of the magnetic
permeability on H is entered.

2.1.2 A reduction of system of nonlinear dif-

ferential equations to the integro-di er-
ential model (Model I)
We assume that a massive body is placed in a variable magnetic
eld. It is necessary to describe the distribution of the eld
inside the body.
Inserting the resistance = 1= in (2.4) we shall express
E = J and substitute in the penultimate equation (2.3), we
have
c
E= r H:
4
Substituting this expression in (2.1), we receive
@H c2
+ r ( r H) = 0: (2.6)
@t 4
A variable magnetic eld penetrating into a substance in-
duces a variable electrical eld, which causes occurrence of cur-
rents. The currents result in heating of the substance and in-
crease of its temperature, which in uences its resistance. From
reasonings given in [399], it follows that one order change of
temperature results in change of resistance on several orders,
so at large uctuations of temperature the dependence = ( )
must be taken into account. Last essential restriction, which
2.1. ELECTROMAGNETIC DIFFUSION PROCESS 23

will be made, is connected to the assumption that the change in

body temperature under action of a current J submits to Joule’s
law

@
0 cv = J 2: (2.7)
@t
Here is density of the substance, 0 is the density at time t = 0,
and cv (as was already mentioned in 2.1.1) is its speci c heat
capacity. Generally they also depend on temperature. Equation
(2.7) does not take into account transfer of heat due to heat
conductivity and radiation. Number of other physical e ects
are not considered. However, from the mathematical point of
view system (2.6), (2.7) is complex enough.
Now we shall begin the reduction of system (2.6), (2.7) to a
system of nonlinear integro-di erential equations.
Let us rewrite equation (2.7) in the following form

cv @
0( ) = J 2:
( ) @t

Introducing the function

Z
cv
S( ) = 0( ) d ;
0
( )

we have
@S
= J 2:
@t
Let us assume that the process begins at the moment of time
t = 0 in which the temperature 0 of the substance is constant.
Integrating this equation on a segment [0; t]; we shall nd
Z t
S( (x; t)) S( 0 ) = J 2d :
0
24 CHAPTER 2. MATHEMATICAL MODELING

The functions 0 ; , and cv are positive in physical sense;

therefore, the function S( ) is monotonically increasing. From
here follows that the inverse function = (S) is uniquely de-
termined, connected with S( ) by parity (S( )) = : Thus, it
is possible to write down
Z t
(x; t) = J 2d :
0
From (2.3), (2.4) we have
c
J= r H;
4
so, Z t
c 2
(x; t) = r H d :
0 4
Substituting this expression in (2.6) as argument of function
= ( ), we receive

t
c2
Z
@H c 2
+ r r H d r H = 0;
@t 4 0 4

r H = 0:
Let us introduce the notations

c2 c
a(S) = ( (S)); W = H
4 4
and rewrite the system in the following form:

Z t
@W
+r a jr W j2 d r W = 0;
@t 0 (2.8)
r W = 0:
2.1. ELECTROMAGNETIC DIFFUSION PROCESS 25

If we consider a at eld of the form W = (0; 0; U ), where

the function U = U (x1 ; x2 ; t) depends on two spatial variables,
then

@U @U
r W = ; ;0
@x2 @x1
and (2.8) takes the form
Z t
@U
=r a jrU j2 d rU : (2.9)
@t 0

Let us note again that equations such as (2.8), (2.9) for the
rst time have arisen in [187] and [188]. In these works, as well
as in [137], [138], [139], [140], [141], [145], [186], [219], [223],
[238], [247], [253], and [261] together with other questions, the
uniqueness of the solutions of the initial-boundary value prob-
lems for equations (2.8) and (2.9) is given, under a rather general
assumptions on the function a = a(S).
Assume that 0 and cv are constants. Let us give examples
of function ( ), inducing a(S), see, e.g., [145], [304], and [306].
If ( ) = , > 1, then

a(S) = C1 (C0 S) 1;
where C0 and C1 are some positive constants. Thus, the power
growth of the resistance ( ) gives a coe cient determined only
on a nite interval. Let us notice that the physical substances
do not have such property.
If ( ) = , then

a(S) = C1 eC0 S ;
with positive constants C0 and C1 , i.e., the linear growth of the
function ( ) gives exponential function a(S). The linear growth
26 CHAPTER 2. MATHEMATICAL MODELING

of resistance with temperature is characteristic of metals.

If ( ) = , 0 < < 1, then

a(S) = C1 (C0 + S) 1 ;
with positive constants C0 and C1 . Thus, under-linear growth
of the function ( ) gives sedate growth of the coe cient a(S).
If ( ) = , < 0, then

a(S) = C1 (C0 + S) 1 + > 0;

with positive constants C0 and C1 , i.e., a decreasing function
( ) gives decreasing functions a(S). This is a general rule in
case 0 and cv are constants. Suppose that function ( ) is
di erentiable for 0 . Then according to the de nition we
have

da c2 d d c2 d =d c2
= = = ( ) 0 ( ):
dS 4 d dS 4 dS=d 4 cv

From here it is obvious that the function a(S) decreases or

grows simultaneously with the function ( ). Let us note that
the decrease of resistance ( ) with growth of temperature is
characteristic of semiconductors in solid, gaseous, and plasma
phases, for which, as it was already remarked in 2.1.1, the the-
oretical formula ( ) = K 3=2 is valid.

2.2 On the averaged Model II

We shall begin now a derivation of the averaged equation [306]
describing again the process of penetration of an electromagnetic
eld into a substance.
2.2. ON THE AVERAGED MODEL II 27

Following [419] and as in section 2.1, we shall consider the

system of the Maxwell’s equations, describing the interaction of
an electromagnetic eld into a substance:
1 @H
= r E; (2.10)
c @t
r H = 0; (2.11)
4
J = r H; (2.12)
c
J = E: (2.13)
Joule’s law (2.7) is a localization of the law of allocation of
heat [419]
Z
dQ = JEdxdt: (2.14)

Here dQ is in ow of a thermal energy absorbed by a body in an

electromagnetic eld in time dt on a substance, occupying area
R3 . Joule’s law does not take into account the process of
transfer of heat inside a body, which is fair if one to consider
the temperature time-dependent but independent of the spatial
coordinates. Thus, in this case it is possible to write that =
(t) and dQ = mcv d , where m is the weight of the substance.
Therefore, (2.14) becomes
Z
d
mcv ( ) = JEdx:
dt
Using Ohm’s law E = ( )J and repeating the process of
exception, we shall receive the following analogue of system (2.8)
in the same designations:

Z tZ
@W 1
+r a jr W j2 dxd r W = 0;
@t j j 0
28 CHAPTER 2. MATHEMATICAL MODELING

r W = 0:
Here j j is the volume of the substance. Due to averaging the
coe cient of this equation depends only on the variable t, and
consequently the equation can be rewritten in the form
Z tZ
@W
=a jr W j2 dxd W: (2.15)
@t 0

Here we used the known vector identity

r r W = W + r(r W ):
For a at eld W = (0; 0; U ), where U = U (x1 ; x2 ; t) is a
function of two spatial variables, system (2.15) becomes
Z tZ
@U
=a jrU j2 dxd U: (2.16)
@t 0

Laptev [306] remarked that the research of the so-called aver-

aged integro-di erential models (2.15) and (2.16) requires inde-
pendent theory than he applied for investigation of (2.8), (2.9),
and such type equations.

2.3 Mathematical Model III

The classical linear theory of heat conduction for homogeneous
and isotropic media is given by

@U @ 2U
=k 2;
@t @x
where U (x; t) is the absolute temperature and k > 0 is a con-
stant called thermal di usivity. This parabolic equation allows
a thermal disturbance at any point in the medium to be in-
stantly felt at other points, see [196]. Gurtin and Pipkin [196]
2.3. MATHEMATICAL MODEL III 29

have suggested a heat ow model based on a memory e ect in

the medium. The linear one-dimensional version of their the-
ory assumes that the internal energy, (x; t); and the heat ux,
q(x; t), satisfy the following relations:
Z 1
(x; t) = bU (x; t) + B( )U (x; t )d ; (2.17)
0
Z 1
@U (x; t )
q(x; t) = c( ) d ; (2.18)
0 @x
where b is the instantaneous speci c heat, B is an energy re-
laxation function, and c is the thermal conductivity (see [345]).
Thus is a functional of the history of the temperature and q
is a functional of the history of the temperature gradient. Mac-
Camy [345] treated a partially nonlinear version of the above
model by replacing (2.18) by
Z 1
@U (x; t )
q(x; t) = K( ) d : (2.19)
0 @x
Now assume that the material is at zero temperature and in-
ternal energy up to time t = 0. This is not restrictive since
we can incorporate nonzero initial values with the forcing term.
Incorporating (2.17), (2.19) with the balance of heat equation,
we have
Z t
@U (x; t) @U (x; )
b + B(t ) d =
@t 0 @
Z t (2.20)
@ @U (x; )
K(t ) d + r(x; t);
0 @x @x
where r(x; t) is the heat source. Let R(x; t) be the resolvant for
B, i.e., the function U given by
Z t
1
U (x; t) = (x; t) + R(t ) (x; )d
b 0
30 CHAPTER 2. MATHEMATICAL MODELING

solves the equation

Z t
(x; t) = bU (x; t) + B(t )U (x; )d :
0

Now de ne a(t) and f (x; t) by

Z t
1
a(t) = K(t) + R(t )K( )d ;
b 0
Z t
1
f (x; t) = r(x; t) + R(t )r(x; )d ;
b 0

then (2.20) becomes

Z t
@U (x; t) @ @U (x; )
= a(t ) d + f (x; t); (2.21)
@t 0 @x @x

which is (1.14).

2.4 Some features of Models I and II

2.4.1 Existence and uniqueness of the
solutions
In the present section we give some features of those types of
equations considered in sections 2.1 and 2.2.
As models we shall consider the following two nonlinear par-
tial integro-di erential equations:

" Z ! #
t 2
@U @ @U @U
a d = f (x; t) (2.22)
@t @x 0 @x @x

and
2.4. SOME FEATURES OF MODELS I AND II 31

!
Z tZ 1 2
@U @U @ 2U
a dxd = f (x; t): (2.23)
@t 0 0 @x @x2

These equations are one-dimensional one-component case of

the equations arising at mathematical modeling of process of
penetration of an electromagnetic eld into a substance with
temperature-dependent coe cient of electroconductivity. As we
noted in section 2.1, equation (2.22) given here was o ered in
[187] and [188] at rst and then was generalized in numerous
works [49], [50], [137], [139], [141], [145], [148], [219], [303], [304],
[305], [306], [333], and [334]. Equation (2.23) describes the same
process as (2.22) and is rst given in [306].
Under certain conditions on the initial data for these equa-
tions the existence and uniqueness of the initial-boundary value
problems with rst kind boundary conditions are given.
For investigations of initial-boundary value problems dis-
cussed in the book, we use usual Sobolev spaces Wpk ( ), W kp ( ),
Lp ( ), C k ( ), Lq 0; T ; Wpk ( ) , Lq 0; T ; W kp ( ) and their
properties, see, for example, [7], [136], [175], [327], and refer-
ences therein. Some description of these spaces will be given in
section 3.6 as well.
In the domain QT = (0; T ) of the variables (x; t); where
= (0; 1) and T is a xed positive constant for equations (2.22)
and (2.23) we shall consider the following initial-boundary value
problem:

U (0; t) = U (1; t) = 0; t 2 [0; T ]; (2.24)

U (x; 0) = 0; x 2 : (2.25)
We shall search the solution of problem (2.22), (2.24), (2.25)
which satis es the identity
32 CHAPTER 2. MATHEMATICAL MODELING

Z Z
@U
+ A(U ) V dxdt = f V dxdt; (2.26)
QT @t QT

where V is any function from the appropriate space (this space

for case a(S) = 1 + S will be speci ed in section 3.6), and
" Z ! #
t 2
@ @U @U
A(U ) a d :
@x 0 @x @x

The basic characteristics of equations such as (2.22) is that

higher derivatives of the nonlinear term depend on time inte-
gral. The coe cients a = a(S) in equation (2.22) contains inte-
gral determining nonlocal operator. In the theory of di erential
equations
Z one often uses the fact that the Volterra’s operator of
t
a kind U d is increasing [175]. However, in a combination
0
with derivative on spatial variable this property might be lost.
For example, the equation
Z t
@U @ @U
= a d
@t @x 0 @x
Z t
is parabolic but if we replace U d = V then the nonlinear
0
equation becomes hyperbolic

@ 2V @ @V
2
= a ;
@t @x @x
which generally does not possess global solutions.
The following existence and uniqueness statement takes place
for problem (2.22), (2.24), (2.25) (see [138]).
2.4. SOME FEATURES OF MODELS I AND II 33

Theorem 2.1 If

a(S) = (1 + S)p ; 0 < p 1;

and f 2 W21 (QT ); then there exists a unique solution U of prob-

lem (2.22), (2.24), (2.25) satisfying the following properties:

U 2 L2p+2 (0; T ; W 12p+2 ( )); @U=@t 2 L2 (QT );

@ 2U
2 L2 (QT );
@x2
p @ 2U
T t 2 L2 (QT ):
@t@x

Theorem 2.1 and some of its generalizations for (2.22), (2.23)

type equations and for their multidimensional variants are proved
in [137], [138], [139], [141], [145], [187], [188], and [219], with ap-
plication of Galerkin’s and compactness methods [327], [328],
and [461].
It should be noted that Theorem 2.1 in case p = 1 will be
proved completely in section 3.6.
Below we discuss the asymptotic behavior of the solutions of
the initial-boundary value problems. The existence result can
also be proved based on the a-priori estimates given in the inves-
tigation of the asymptotic behavior of the problem for Models I
and II.
In [305] and [306] it was shown that equation (2.22) gives rise
to coercive monotone operators in the space W 1p (0; 1; L2 (0; T ))
but not in the traditional L2 (0; T ; W 1p (0; 1)). In mentioned works
by the same scheme the analogous results are obtained for (1.12)
type models.
34 CHAPTER 2. MATHEMATICAL MODELING

2.4.2 Asymptotic behavior of the solutions as

time tends to in nity
This section is devoted to longtime behavior of the solutions
of initial-boundary value problems for the nonlinear integro-
di erential models (2.22) and (2.23). These equations arise at
mathematical modeling of process of penetration of an electro-
magnetic eld into a substance, with temperature dependent
coe cient of electroconductivity. These equations were derived
in sections 2.1 and 2.2.
The initial-boundary value problems with rst-type bound-
ary conditions are stated. Investigations, which are made in
[217], [218], [228], [231] are given. Let’s formulate the state-
ment for asymptotic behavior of the solutions of the problem
with homogeneous boundary conditions.
At rst for one nonlinear integro-di erential problem (2.22)
we shall give a-priori estimates of the solutions independent of
t. From these estimates the stabilization of the solution follows
as t ! 1.
In the cylinder Q = (0; 1); we consider problem (2.22),
(2.24), (2.25).
Let us assume that

a(S) a0 = const > 0:

If
2a0
;
C
where C here and below is a positive constant independent of t,
the following estimate takes place

t=2
jjU ( ; t)jj e jjU0 jj :
Another random document with
no related content on Scribd:
 Not wanting to put a high poultry netting on my fence to keep my
poultry at home, I devised the method shown, and was thus enabled
to use 24-in. wire instead of 48-in., thereby effecting a saving, as
well as not marring the appearance of the fence. Chickens will not
attempt to fly over this wire but once or twice, as they endeavor to
alight on the pickets of the fence instead of the wire, and
consequently are thrown back by striking the wire. The wire should
be supported at each post, as shown in the sketch.—Roy H. Poston,
Flat River, Mo.
A Fort Built of Snow Blocks Made in Box Form

The Making of the Snow-Block Fort Is Nearly as Much Fun as Daring the
Enemy to Come and Capture It or Capturing the Enemy’s Barricade
 What boy does not enjoy a regular snowball fight—a battle, in fact
—especially when he has a fort of his own building to defend, and
his friends have one which he wants to capture? A way to build a
snow fort that will not crumble easily when the enemy charges it is to
pack the snow into a box form and build the structure of the snow
blocks, as shown in the photograph. Portholes can be provided and
when the snow melts slightly, and freezes again, a very solid wall
results. The form for the block, as detailed in the sketch, is made of a
box, arranged so that it is hinged at one corner and has a catch at
the corner diagonally opposite. The form is set on a wooden board
and when the snow is packed in solidly, the catch is opened and the
form lifted away from the block. It is then carried into position and
another block made on the same board. Several boards are handy in
making the block quickly.—H. S., Chicago, Ill.

¶In applying varnish over an old wood surface, or on new wood that
has become soiled, care should be taken to remove all traces of oil
or grease with benzine.
 Model Steam Tractor Made by Boy

This Model Steam Tractor Is 28 Inches Long and Drives Itself, Wood and Coal
being Used as Fuel

Boys, and others who like to make models, may be interested in a

small steam tractor I designed and built of scrap materials around
the farm shop. I used only the ordinary tools on hand in such a
place. The steam engine and tractor run smoothly, and the engine
propels the tractor just like a large machine of this kind. The tractor is
28 in. long and 14 in. high, and the boiler holds 1¹⁄₄ gal. The bore of
the cylinder is ⁷⁄₈ in. and its stroke 1³⁄₄ in. It was made from an old
iron window weight. I use wood and coal to fire the boiler, and steam
can be got up in a few minutes.—Albert H. Bennington, Park River,
N. D.
 Toboggan Made of Old Buggy Shafts

This Toboggan with Stout Runners Made from Buggy Shafts can be Used
like One of the Ordinary Type, but Works Best on Frozen Snow

A method of making a strong toboggan sled by nailing a suitable

bottom over sections cut from old buggy shafts, as shown in the
sketch, can be adapted easily for use with poles of about the same
size. They can be steamed and bent at the ends and the
construction otherwise made as shown. The shafts were cut to the
length indicated, the curved front end being sawed off. The bottom
was carried up around the curved rear end, and a strong bracing
strip was bolted across them. One-half-inch rope with leather loops
provided handholds, and was also used as a brace at each side of
the front of the sled. Twisted wire may also be used for this purpose.
Cushions of the usual type can easily be made for the toboggan.
 Bicycles Driven as Three-Wheel Tandem after
Breakdown

When the Front Wheel of One of the Bicycles Broke Down, They were
Rigged into a Tandem, and a Distance of Five Miles to Town was Covered

On a bicycle trip, when about five miles from town, the front axle of
my partner’s bicycle broke, and as neither of us wanted to walk, we
had to find some way of getting us both home. We removed the front
wheel from one of the bicycles, and set the front fork on the rear axle
of the undamaged machine, as shown in the sketch. The axle was
just long enough so that the sides of the fork could be bolted
securely. By careful riding we made the distance to town with only
the slight inconvenience of my partner having to carry the extra
wheel on his back—J. F. G., Mountain Grove, Mo.

¶Do not touch the inner surface of a tent or shelter in the rain, as the
water will drip from the spot touched.
 Test Exposures for Bromide Enlargements
To avoid expensive waste of paper in making large bromide
enlargements the following device is useful: Cut an oblong sheet of
stiff cardboard, 5¹⁄₄ by 4 in., with a central opening, measuring 3³⁄₄ by
1¹⁄₂ in. Along the lower edge of the central opening mark off five
spaces of ³⁄₄ in. each, and number them from 1 to 5. Glue a strip of
cardboard, 1 in. wide, at the top and the bottom of the oblong, on the
back. Now glue a second strip of cardboard over each of those just
attached, making the second strips slightly wider to form grooves
similar to those in a sliding-top box.
Cut a piece of cardboard, wide enough to slide easily in these
grooves, and about 7 in. long. In the center of it, cut an opening, 1¹⁄₂
in. long and ³⁄₄ in. wide, with the longer dimension extending across
the strip. By inserting this slide into the grooves, the opening in it
may be shifted to correspond to any of the five ³⁄₄-in. indications at
the lower edge of the opening in the oblong frame.
Focus the enlargement on the easel and put the cap on the lens.
Then pin a 1¹⁄₂-in. strip of bromide paper where a good range of
detail will fall upon it. Pin the oblong card over the bromide strip and
make five successive exposures, of increasing duration, by sliding
the strip so that its opening is opposite the number indications on the
oblong. Note the particulars of exposure at the time and later
develop the strip of bromide paper. Suppose the exposures were of
8, 16, 24, 32, and 40 seconds’ duration, and upon development it
was found that the first and second exposures were much
underexposed and that the fifth was overexposed; the third and
fourth being only slightly under and overexposed, respectively. Then
the mean of these two, that is 24 plus 32 divided by 2, or 28
seconds, will be approximately the correct exposure. The use of this
device will effect a considerable saving.—H. J. Gray, Lewes, Sussex,
England.
Scale on Vaulting Pole Indicates Points of Grip
 Indications on the Vaulting Pole Aid the Athlete in Finding the
Proper Grip

Unnecessary delay in running off the pole-vaulting events at

athletic meets is caused by the repeated testing of vaulters for the
proper height at which to grip the vaulting pole. This was readily
overcome by marking intervals of 1 ft. on the upper end of the
vaulting poles. Vaulters judge the proper height once and thereafter
may easily take the same grip with the aid of the indications on the
pole.—A. B. Wegener, Madison, N. J.
 Current of Canal Raises Irrigation Water

The Current of the Canal Causes the Wheel to Revolve, Carrying the Water
to the Trough

The current of a canal was utilized to raise water for irrigation

purposes, and other uses, by means of the homemade device
shown in the illustration. A large-size buggy wheel was set on an old
axle, and extensions were fixed to eight of the spokes. Paddles, 9 by
18 in. in size, were secured to the extensions, and uniform-size tin
cans were fixed to the paddles, as shown. The current revolved the
wheel, carrying the loaded buckets upward, and the water was thus
poured from them into a wide trough braced on the platform. A flume
led the water to the irrigation ditches, and pipes were used to
conduct some of it to the ranch house for general purposes. When
the water is not desired for useful purposes the trough is withdrawn
and the water drops into the canal.—Joseph Gray, Heber, Calif.
 Protecting an Air Pump against Denting

Denting of Tire Pumps is Prevented by the Use of a Case Like That Shown

The brass cylinders of many tire pumps are made of thin metal,
and a slight blow may dent them. Such injuries interfere with the
action of the plunger, making the pump almost useless. A case which
will afford protection to the pump is shown in the sketch. It was made
of two strips of wood cored out on the inner sides to fit the pump and
provided with small hinges and hooks, to lock it into place.
 A Feather as an X-Ray Lens
The outline of the bones in the hand may be seen by holding it
before a strong light and looking through a lens made of a piece cut
from the tip of a feather. The sketch shows such a lens, which is
mounted in a small piece of cardboard that may be carried in the
pocket, and which affords diversion wherever exhibited.
 The device is made as follows: Procure a soft white feather from a
fowl, and cut off the tip as indicated in the upper sketch. Cut a piece
of cardboard, about 2 in. wide and 6 in. long. Fold it to form a folder,
3 in. long, and cut a hole, ¹⁄₄ in. in diameter, through both parts of it.
Glue the tip of the feather between the folder at the hole, taking care
that the fibers lie flat and that no rib is exposed. Glue the edges of
the folder, and the device is ready for use. Hold the hand, or other
object, to be examined about 15 in. from the lens, and place the
latter close to the eye. Looking through the hole toward an electric,
or other strong, light, it will be observed that the bones of the hand,
for example, may be seen clearly. They will be outlined with a band
of varicolored light. A frosted globe is better than one which exposes
a filament.—Leander L. Droesch, Chickasaw, Ohio.

¶A safety-razor blade, fitted into a wooden handle, provides a

convenient tool for ripping seams and cutting cardboard.
 Guard Saves Wear on Vest
Bookkeepers and others who work close to a desk nearly all day
find that unusual wear results on the front of the vest, forcing them to
discard their suits sooner than would otherwise be necessary. By
providing a strip of cloth of a color to match the goods of which the
suit is made and buttoning it across the front of the vest from the
lower to the upper button this wear is prevented.—J. A. McGrath,
Hancock, Michigan.
 Bilge Water Siphoned Through Water-Jacket
Outlet

The Small Pipe Fixed to the Water-Jacket Outlet Siphons the Bilge Water
from the Bottom of the Motorboat

The sketch shows a simple method of removing the bilge water

from a motorboat automatically while the engine is in operation. A
small pipe is fitted into the outlet from the water jacket, as shown at
A, and extended to the bottom of the boat where the bilge water
accumulates. A sack is fixed over the end of the pipe to prevent
particles of dirt from entering the pipe. The water passing out of the
water-jacket outlet causes a siphoning action in the small pipe, and
the bilge water is drawn up and passes out of the larger outlet. At the
left, in dotted outline, is shown a method of removing the bilge water