Get Inverse Problems of Mathematical Physics Mikhail M. (Jn.) Lavrent'Ev PDF ebook with Full Chapters Now

Visit https://ebookgate.
com to download the full version and

explore more ebooks
Inverse Problems of Mathematical Physics Mikhail M.

(Jn.) Lavrent'Ev
_____ Click the link below to download _____

https://ebookgate.com/product/inverse-problems-of-
mathematical-physics-mikhail-m-jn-lavrentev/
Explore and download more ebooks at ebookgate.com

Here are some recommended products that might interest you.
You can download now and explore!
Contemporary Problems in Mathematical Physics 1st Edition

J. Govaerts
https://ebookgate.com/product/contemporary-problems-in-mathematical-
physics-1st-edition-j-govaerts/
ebookgate.com
Inverse Problems and Inverse Scattering of Plane Waves

Dilip N. Ghosh Roy
https://ebookgate.com/product/inverse-problems-and-inverse-scattering-
of-plane-waves-dilip-n-ghosh-roy/
ebookgate.com
Computer Modelling in Tomography and Ill Posed Problems

Mikhail M. Lavrent'Ev
https://ebookgate.com/product/computer-modelling-in-tomography-and-
ill-posed-problems-mikhail-m-lavrentev/
ebookgate.com
Dynamical Inverse Problems of Distributed Systems

Vyacheslav I. Maksimov
https://ebookgate.com/product/dynamical-inverse-problems-of-
distributed-systems-vyacheslav-i-maksimov/
ebookgate.com
N Geometry of the generalized geodesic flow and inverse
spectral problems 2nd Edition Vesselin M. Petkov
https://ebookgate.com/product/n-geometry-of-the-generalized-geodesic-
flow-and-inverse-spectral-problems-2nd-edition-vesselin-m-petkov/
ebookgate.com
Inverse and Ill Posed Sources Problems Yu. E. Anikonov
https://ebookgate.com/product/inverse-and-ill-posed-sources-problems-
yu-e-anikonov/
ebookgate.com
Differential equations inverse and direct problems 1st

Edition Angelo Favini
https://ebookgate.com/product/differential-equations-inverse-and-
direct-problems-1st-edition-angelo-favini/
ebookgate.com
Inverse problems in atmospheric constituent transport 1st

Edition I. G. Enting
https://ebookgate.com/product/inverse-problems-in-atmospheric-
constituent-transport-1st-edition-i-g-enting/
ebookgate.com
Inverse Problems for Kinetic and Other Evolution Equations

Yu. E. Anikonov
https://ebookgate.com/product/inverse-problems-for-kinetic-and-other-
evolution-equations-yu-e-anikonov/
ebookgate.com
INVERSE A N D ILL-POSED PROBLEMS SERIES
Inverse Problems of Mathematical Physics

Also available in the Inverse and Ill-Posed Problems Series:
Ill-Posed Boundary-Value Problems Radiation Transfer
S.E.Temirbolat VS. Antyufeev
Linear SobolevType Equations and Degenerate Introduction to the Theory of Inverse Problems
Semigroups of Operators AL Bukhgeim
GA Sviridyuk andV.E Fedorov Identification Problems ofWave Phenomena - Theory
Ill-Posed and Non-Classical Problems of Mathematical and Numerics
Physics and Analysis S.I. Kabanikhin and A Lorenzi
Ed/tors: M M . Lavrent'ev and S.I. Kabanikhin Inverse Problems of Electromagnetic Geophysical Fields
Forward and Inverse Problems for Hyperbolic, Elliptic P.S. Martyshko
and Mixed Type Equations Composite Type Equations and Inverse Problems
AG. Megrabov Al. Kozhanov
Nondassical Linear Volterra Equations of the First Kind Inverse Problems ofVibrational Spectroscopy
AS. Apartsyn AG.Yagola, I.V. Kochikov, GM. Kuramshina andYu.A. Pentin
Poorly Visible Media in X-ray Tomography Elements of the Theory of Inverse Problems
D.S. Anikonov, V.G. Nazarov, and I.V. Prokhorov A M . Denisov
Dynamical Inverse Problems of Distributed Systems Volterra Equations and Inverse Problems
V.l. Maksimov A L Bughgeim
Theory of Linear Ill-Posed Problems and its Small Parameter Method in Multidimensional Inverse
Applications Problems
V.K Ivanov.V.V.Vasin andV.P.Tanana AS. Barashkov
Ill-Posed Internal Boundary Value Problems for the Regularization, Uniqueness and Existence ofVolterra
Biharmonic Equation Equations of the First Kind
MAAtakhodzhaev A. Asanov
Investigation Methods for Inverse Problems Methods for Solution of Nonlinear Operator
V.G. Romanov Equations
Operator Theory. Nondassical Problems V.P.Tanana
S.G. Pyatkov Inverse and Ill-Posed Sources Problems
Inverse Problems for Partial Differential Equations Yu.E Anikonov, BA Bubnov and G.N. Erokhin
Yu.Ya. Belov Methods for Solving Operator Equations
Method of Spectral Mappings in the Inverse Problem V.P.Tanana
Theory Nondassical and Inverse Problems for
V.Yurko Pseudoparabolic Equations
Theory of Linear Optimization A Asanov and ER. Atamanov
I.I. Eremin Formulas in Inverse and Ill-Posed Problems
Integral Geometry and Inverse Problems for Kinetic Yu.E Anikonov
Equations Inverse Logarithmic Potential Problem
AKh.Amirov V.G. Cherednichenko
Computer Modelling in Tomography and Multidimensional Inverse and Ill-Posed Problems for
Ill-Posed Problems Differential Equations
M M . Lavrent'ev, S.M. Zerkal and O.ETrofimov Yu.E Anikonov
A n Introduction to Identification Problems via Ill-Posed Problems with A Priori Information
Functional Analysis V.V.Vasin andALAgeev
A. Lorenzi Integral Geometry ofTensor Fields
Coefficient Inverse Problems for Parabolic Type VA. Sharafutdinov
Equations and Their Application Inverse Problems for Maxwell's Equations
PC Danilaev V.G. Romanov and S.I. Kabanikhin
Inverse Problems for Kinetic and Other Evolution
Equations
Yu.E Anikonov
Inverse Problems ofWave Processes
AS. Blagoveshchenskii
Uniqueness Problems for Degenerating Equations and
Nondassical Problems
S P. Shishatskii, A Asanov and £R Atamanov
Uniqueness Questions in Reconstruction of
Multidimensional Tomography-Type
Projection Data
V.P. Golubyatnikov
Monte Carlo Method for Solving Inverse Problems of
INVERSE AND ILL-POSED PROBLEMS SERIES
Inverse Problems
of
Mathematical Physics
M.M. Lavrentiev, A.V. Avdeev,

M.M. Lavrentiev, Jr. and V.l. Pnimenko
IIIVSP///
UTRECHT · BOSTON
2003
VSP Tel: +31 30 692 5790
an imprint of Brill Academic Publishers Fax: +31 30 693 2081
P.O. Box 346 [email protected]
3700 AH Zeist www.brill.nl
The Netherlands www.vsppub.com
© Koninklijke Brill NV, 2003
First published in 2003
ISBN 90-6764-396-3
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system,
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or
otherwise, without the prior permission of the copyright owner.
Printed in The Netherlands by Ridderprint bv, Ridderkerk.

Preface
The present book describes a part of the theory of the so-called Inverse
Problems of Mathematical Physics and some applications of such problems.
Mostly the theoretical aspects of Inverse Problems are discussed. Besides,
we also consider some applications and numerical methods of solving the
problems under study. Descriptions of particular numerical experiments are
also included.
The theory of Inverse Problems of Mathematical Physics is a vast and
intensively developing field of modern mathematics. Plenty of publications
appear, and even a number of specialized journals are published. Because
of extended area of applications, many various statements of problems are
considered, and diverse methods are used for their solution.
We stress attention at providing a concept of versatility and complexity
of inverse problems arising in applications. We did not pursue the aim of
giving the complete review of literature, instead we pointed out the most
popular textbooks and characteristic statements of the problems. Also, we
often pointed out the connections of such problems with various applications
of methods of mathematical simulation. The references cited are mainly of
illustrative character. At the same time, we gave references to the most
frequently cited monographs which contain further references and a more
complete account of the history of this field. Meaning to provide an in-
troduction intended for specialists in other fields, we tried to emphasize
the basic general principles and approaches to solution of various problems,
supplying them with concrete examples of results obtained.
The monograph is arranged as follows. In Introduction we explain our
understanding of the concept of Mathematical Modeling, outline the general
differences between direct and inverse problems, and give strict mathemat-
ical definitions of correct and ill-posed problems.
In Chapter 1, we show up a list of applied areas, where inverse problems
have been successfully used for years. Of course, it is impossible to list
ii M. M. Lavrentiev et al. Inverse problems .
all such applications, so we took a liberty to mention somehow the basic

fields, some of which historically developed along with the theory of inverse
problems.
In practice, almost all problems are solved approximately, in mathemat-
ical sense. So, in Chapter 2, we give some of the basic definitions related to
various regularizations of inverse problems. We hope that the reader who
does not need strict definitions can get all the general ideas just browsing
the text part of the chapter.
Chapter 3 is dedicated to the problems of Integral Geometry, which are
both classical (accounting Radon transform, e.g.) and non standard. Some
recent results, previously available only through specialized journals, are
included.
Chapter 4 is arranged in non traditional way providing the reader with
a sort of overview of one-dimensional inverse problems. In spite of usual
separation of model equations of hyperbolic and parabolic type, the chapter
is compiled as follows. First, Lamé system is described from physical state-
ment to model simplifications and uniqueness and stability results. Second,
the so-called "quasi-stationary approximation" of Maxwell system is con-
cerned. The point is that the complete proof through analytic relations
between solutions to equations of hyperbolic and parabolic types is given.
Then, a concept of relations among inverse problems for different type gov-
erning equations are discussed. Such concept is not that new, but perhaps is
not well known. The next sections is dedicated to brief descriptions of such
fundamental methods of investigation as the separation of singularities and
the reduction of one-dimensional inverse problem with a focused source of
disturbances to a linear integral equation. Finally, the determination of the
piece-wise constant coefficient for wave equation is considered. The example
of inverse problem (arising in applications) is given, in which case recurrent
algorithm for exact determination of equation coefficients is available.
In Chapter 5 we consider some inverse problems for the coupled Maxwell
and Lamé systems. First of all, the solution of the one-dimensional inverse
problems for the equations of electromagnetoelasticity in the case of seismo-
magnetic interaction is studied. Then we give some results of the solution
of inverse problems for the system of electromagnetoelasticity in the case of
piezoelectric effect. In the next section, the linear process of interaction of
electromagnetic and elastic waves in a weakly conducting elastic medium is
considered. Finally, we give some results of solution of direct and inverse
problems for the system of electromagnetoelasticity in the case of nonlinear
interactions between electromagnetic and elastic fields.
Preface iii
Chapter 6 contains some examples of numerical solution of the different

type inverse problems arising in applications. The first section is a small
survey in numerical methods for inverse problems. The next section repre-
sents the numerical solution of a 3D inverse kinematic problem of seismics.
Then we describes how the proposed algorithm of numerical solution could
be used for the determination of the structure of the Earth's upper mantle.
The numerical solutions of two inverse problems of electromagnetoelasticity
are discussed in the next section. We would like to draw attention on the
last section, which represents the results of numerical modeling of coastal
profile evolution. This application of inverse problems is new and not well
developed at the moment.
Acknowledgement
The authors axe grateful to Academician of RAS, Professor A. S. Alekseev,

Corresponding Members of RAS, Professors V. G. Romanov, V. V. Vasin,
and B. G. Mikhailenko, and to Professors Yu. E. Anikonov, S. I. Kabanikhin,
V. G. Yakhno, and V. A. Cheverda for useful discussions and kind permission
to use some their results.
Also, the authors are grateful to A. Nazarov and D. Nechaev for their
kind and helpful association in preparation of the camera-ready.
Finally, we would like to mention that this book was written under finan-
cial support, in part, of the Russian Foundation for Basic Research (projects
Nos. 01-05-64704, 02-05-64939) and the U.S. Civil Research Development
Foundantion (project No. RG1-2415-NO-2).
Mikhail M. Lavrentiev,
Alexander V. Avdeev,
Mikhail M. Lavrentiev, Jr.,
Viatcheslav I. Priimenko
iv M. M. Lavrentiev et al. Inverse problems ...
Contents
Introduction 1
0.1. The concept of mathematical simulation 1
0.2. Direct and inverse problems 3
0.3. On correctness of direct and inverse problems of mathematical
physics 5
Chapter 1. Some physical motivations of inverse problems 13

1.1. Inverse problems of geophysics 13
1.2. Inverse tomography problems 22
Chapter 2. Approximate methods of solution of ill-posed

problems 29
2.1. On some aspects of statement and solution of ill-posed problems . 29
2.2. Solutions on compact sets. The concept of a quasi-solution . . . . 30
2.3. The method of quasi-inversion 41
2.4. Regularization methods 47
Chapter 3. Integral geometry problems 53

3.1. Statement of integral geometry problems 53
3.2. The Radon problem 53
3.3. The problem of general form on the plane 57
3.4. Problems of general form in the space 65
3.5. Problems of Volterra type with manifolds invariant under
the motion group 77
3.6. Integral geometry problems with perturbation on the plane . . . . 81
3.7. Mathematical problems of tomography and hyperbolic mappings . 88
vi M. Ai. Lavrentiev et al. Inverse problems
Chapter 4. One dimensional inverse problems 103

4.1. Some inverse problems for Lamé system 106
4.2. Inverse problems for quasi-stationary Maxwell's equations . . . . 112
4.3. Connections among inverse problems of hyperbolic, elliptic,
and parabolic type 118
4.4. Problems with a focused source of disturbances 125
4.5. Reducing the problem with a focused source of disturbances to a
linear integral equation: necessary and sufficient conditions
for solvability of the inverse problems 135
4.6. Determination of the piece-wise constant coefficient for wave
equation 140
Chapter 5. Inverse problems for the coupled Maxwell

and Lamé systems 147
5.1. One-dimensional inverse problem of electromagnetoelasticity
in the case of the seismomagnetic effect 148
5.2. Inverse problems of electromagnetoelasticity in the case of
piezoelectric interaction 156
5.3. Inverse problems of electromagnetoelasticity for weakly
conducting media 1Ö5
5.4. An inverse problem of electromagnetoelasticity in the case of
nonlinear interaction 179
Chapter 6. Numerical solution of inverse problems:

some examples 191
6.1. Short review of numerical approaches to solving inverse problems 191
6.2. Numerical solution of a 3D inverse kinematic problem of seismics 196
6.3. Determination of the structure of the Earth's upper mantle . . . . 209
6.4. Numerical solution of inverse problems of
electromagnetoelasticity 229
6.5. Simulation of the long-term coastal profile evolution 245
Bibliography 259
Introduction
0.1. THE CONCEPT OF MATHEMATICAL SIMULATION
The statement of direct and inverse problems of mathematical physics im-

plies preliminary schematic representation of the real process in a certain
mathematical form. At present, modern technology is inconceivable without
mathematical modeling, which is understood as replacement of the original
object under investigation by its model. The aim of such a replacement is
to study the properties of the object with the help of its model.
There are experimental (physical) modeling and theoretical (mathemat-
ical) simulation. In the former case the process is studied on the basis
of real experiments on mock-ups, physical models, laboratory installations,
etc. Mathematical simulation consists in construction and investigation of
quantitative values for physical parameters through a simplified model of
the process under study, which is formulated in mathematical terms. As a
rule, a mathematical model should be adequate to the physical process.
Mathematical models contain unknown characteristics. The choice and
evaluation of these characteristics is a difficult problem which is solved on
the basis of accumulated experience, available experimental data, physical
laws, etc. To develop a mathematical model it is necessary to pass through
two main stages: identification, i.e., the choice of the type (structure) of the
model, and determination of the numerical values of the model parameters
and characteristics.
The structure of a model is conceived as a qualitative character of the
mathematical description of the processes under investigation. Thus, phys-
ical laws are most frequently represented in the form of differential equa-
tions. One can distinguish models with concentrated parameters, in which
case physical parameters are independent of space and only their evolution
in time is the subject of study. Such models are described by systems of
2 M. M. Lavrentiev et al. Inverse problems
Ordinary Differential Equations. On the contrary, models with distributed

parameters involve spatial distribution of physical fields in addition to tem-
poral evolution. Such models are based on Partial Differential Equations
(equations in partial derivatives).
Such classification is very rough. Indeed, one can distinguish stationary
(steady-state) and non-stationary (dynamic) models, linear and nonlinear
models, one-dimensional and multidimensional (in space variables) models,
etc.
Henceforth we shall deal with models represented in the form of differ-
ential equations in partial derivatives (PDE's). According to the general
theory of such equations, a number of additional conditions should be at-
tached to the equation itself to single out a unique element from the whole
of the set of solutions to this equation.
We shall mostly speak about linear governing equations of second order.
It means that the changes of physical variables (displacement, temperature,
concentration, field intensity, etc.) axe described by linear differential equa-
tions, whose coefficients represent the physical properties of the medium
where the process being modeled takes place.
Example 0.1.1. The equation

d2u o d2u tn „ „.
describes small oscillations of a string. In this case the function u(z,t)

is the displacement of the string, at a point ζ and at a moment i, from
the equilibrium state, which is supposed to coincide with the z-axis. The
coefficient c characterizes the speed of disturbance propagation along the
string. If the string is bounded, ζ G [0,/], and fixed at the endpoints ζ = 0
and ζ — I, then the boundary conditions for (0.1.1) have the form
u(0, t) = u(l, t) = 0, t e K+, (0.1.2)
where R + = [0, oo). To set the boundary conditions is not sufficient to single
out a unique solution to (0.1.1). We arrive at a unique solution only if some
additional initial conditions are used. In this case they axe the initial string
displacement and the initial velocities of its points:
du
η(ζ,0) = φ(ζ), —(ζ,0)=φ(ζ), ze[0,l}. (0.1.3)
One can prove that conditions (0.1.2) and (0.1.3) define a unique solution
to (0.1.1).
Introduction 3
0.2. D I R E C T A N D INVERSE PROBLEMS
In some statements, the coefficients of the governing equations aie considered

as given, thus, the problem is to study the properties of solutions to the
model equations. In the case where the initial and boundary conditions are
set "properly" (there exists a unique solution to the problem under study
and this solution depends continuously on the parameters of the problem),
we shall refer to such statements as Direct Problems.
The statement of each direct problem implies prescribing some set of
functions, i.e., the coefficients of the equations, sources (right-hand sides of
the equations), external actions (nonhomogeneities and the coefficients of
the boundary conditions), etc. In the result of solving a direct problem, a
new set of functions — the solutions to the direct problem — is placed into
correspondence to that original set. Thus, the operator of a given direct
problem is defined; i.e., the operator which maps the data of the problem
into its solution.
Assume now that some functions among the data of a "properly stated"
direct problem are unknown; and, instead, some additional information on
the solution of the problem is given. Such problems will be referred to as
Inverse Problems.
In particular, inverse problems of mathematical physics are often under-
stood as problems of determining the internal characteristics of a medium
(as a rule, they cannot be measured directly) from a certain information
on the values of various physical fields (parameters) at the boundary of a
certain domain.
Example 0.2.1. The wave propagation in a vertically-inhomogeneous

medium can be described by the following problem:
d u
c2! \ A
= fojAu, zeR3, tei, (0.2.1)
u
lt<0 Ξ leR3, (0.2.2)
ÄT =/(*)S(*1.32), (ii,i2,i)Gl2 xl+, (0.2.3)

ox* 13=0
where c(x3) is a function characterizing the velocity of wave propagation in
the medium, and f(t) and g(x\,x2) describe the duration of action and the
space distribution of sources on the free surface, respectively.
Inverse Problem 0.2.1 (IP 0.2.1). Let the wave propagation be

described by the system (0.2.1)-(0.2.3). It is necessary to reconstruct the
4 M. M. Lavrentiev et al. Inverse problems .
velocity distribution (to find the function c(x$)) in the medium and/or the
characteristics of the sources, f(t) and g(x 1,0:2), using the additional infor-
mation about the vibration regime of the observation surface £3 = 0
«1x3=0 = «ο(χΐ,Χ2,ί),
3
t G [0, Τ], (n,i2)e5c{iel I a;3 = 0}. (0.2.4)
To clarify better the relations and distinctions between Direct and In-
verse Problems of Mathematical Physics, we give their "cause-and-effect"
interpretation. Consider the given physical parameters of a medium (e.g.,
density, conductivity, etc.) along with the boundary and initial conditions,
the geometry of the domain, etc. as causai characteristics. As effects we ob-
tain the states of physical fields (temperature or concentration distributions,
velocity fields, etc.), which are determined by solving the corresponding di-
rect problem. So, to solve a Direct Problem means to describe the effect
of given causal factors. On the contrary, solution of an Inverse Problem is
interpreted as reconstruction of causal characteristics from their effect.
Therefore, in contrast to Direct Problems, the statements of some Inverse
Problems do not correspond to any physically realizable events. Indeed, one
cannot invert the direction of time-flow (in order to reconstruct the initial
distribution of a physical field from its state at a given moment); it is also
impossible to reverse the process of reagent diffusion or heat propagation.
In this sense, one can say that a number of inverse problems are "physically
incorrect". In mathematical statements, naturally, this difficulty displays
itself as mathematical incorrectness, which results in such complications as
instability of a solution, multiple solutions, even absence of solutions, etc.
These natural causes give rise to difficulties in development of reliable meth-
ods and algorithms to solve inverse problems.
That is why, in spite of existence of many general methods for solution of
inverse problems, each concrete statement requires special theoretical treat-
ment. Note that without such preliminary "analytic" investigation, it is
practically impossible to create cost-effective and efficient numerical algo-
rithms.
A natural approach to solving complex problems consists in constructing
a series of models with increasing complexity that describe the initial state-
ment more and more comprehensively. Consecutive study of these models
allows us to determine, at initial stages, the most general qualitative proper-
ties of solutions. Later these general properties are determined more exactly
in the course of study of more complex models.
Introduction 5
0.3. ON CORRECTNESS OF DIRECT A N D INVERSE

PROBLEMS OF MATHEMATICAL PHYSICS
The notion of correctness is usually considered in the theory of Direct Prob-
lems of Mathematical Physics. When dealing with Inverse Problems of
Mathematical Physics it is convenient to alter this notion a little. Below
we give a short overview.
0.3.1. General notes about correct problems
The theory of differential equations states that a differential equation defines

a whole set of its solutions which depend on a certain number of arbitrary
constants or arbitrary functions. For the problem to have definite physical
sense, we need to single out a unique solution. Usually it is achieved by
setting initial and boundary conditions. This is illustrated by examples
below.
Example 0.3.1.
Consider the so-called heat equation, which describes, e.g., the tempera-
ture evolution in a cooling body,
du .
(0.3.1)
in a cylindrical domain Q = Ω χ R+, where Ω C R3 is a domain bounded by

a closed surface S. To single out a unique solution in this case it is sufficient
to set the heat regime on the surface 5, for instance,
u(a;,i)=0, χ e S, f G M+, (0.3.2)
and the initial distribution of temperature inside Ω,
it(z,0) =g(x), χβΩ. (0.3.3)
Setting of the initial and boundary conditions is aimed at singling out a

unique solution from the whole class of solutions to a differential equation.
But the number of these conditions should be minimal, for otherwise they
may contradict one another, in which case a solution to such problem does
not exist.
As it is known (see, e.g., Ladyzenskaya, Solonnikov, and Ural'tseva,
1968), there exists a unique solution to the problem (0.3.1)-(0.3.3). More-
over, small enough perturbations of the initial profile g(x) cause arbitrarily
6 M. M. Lavrentiev et a1. Inverse problems
small deviations of this solution over any finite time interval t G [0,T]. One
should remember that the main goal of solving mathematical problems is
to describe certain physical processes in mathematical terms. In this case
the initial data are obtained experimentally; and since measurements cannot
be absolutely precise, the data contain measurement errors. For a mathe-
matical model to describe a real physical process, the problem should be
supplemented with some additional requirements reflecting, in a physical
sense, the fact that the solution should have only small variations under
slight changes in initial data or, to put it conventionally, the stability of the
solution under small perturbations in the initial data.
Generally speaking, in such case the problem is said to be correct, while
in alternative cases, ill-posed or incorrect.
0.3.2. Mathematical definitions
Now we put the above general idea on the strict theoretical (mathemati-
cal) basis. Given a differential equation with concrete initial and boundary
conditions, we can pose the problem of finding its solutions belonging to
various functional spaces. The choice of a concrete function class depends
on the physical interpretation of the problem. For example, we can consider
the problem of finding a solution to (0.1.1)—(0.1.3) in the class of func-
tions C2(D), where D = {{z,t) \ ζ G [0,/], t G R+}, or in other classes. In
other words, one can choose a functional space of solutions to a differential
equation in quite an arbitrary way.
The functions involved in the boundary and initial conditions of the
problems for differential equations cannot be chosen arbitrarily; they should
ensure that the solution belongs to the chosen functional space. For this,
they should belong to the certain special functional space corresponding to
the space of solutions. This becomes clearer if one considers problems for
differential equations from the viewpoint of functional analysis. Choose a
space U for the solutions of a differential equation. The differential equation
together with some additional conditions defines the operator A that relates
any solution u G U to the set of functions involved in the additional (initial
and/or boundary) conditions. For (0.1.1) these are the functions φ and ψ, for
(0.3.1) it is the function g. Considering this set of functions as an element /
of a functional space F , one comes to a conclusion that solving a problem for
a differential equation is equivalent to solving the formal operator equation
Au = f (0.3.4)
under the condition that u G Í7.
Introduction 7
The solution to this equation exists if and only if the element / is the
image of a certain element u € U, i.e., it should belong to the set of values of
the operator A. Thus, the set of data of the problem is defined by specifying
the space of solutions U.
If there exists a unique solution to (0.3.4), then the inverse operator A'1
exists, too,
u = A-1/, (0.3.5)
that solves the problem by fitting the solution u G U to the initial data of
the problem, i.e., to the element / .
Let u be a solution to (0.3.4) wherein the operator A acts from the
normed space U into the normed space F. A solution to (0.3.4) is said to be
stable under small variations in the right-hand side / 6 F if for every ε > 0
there exists <5 > 0 such that for every element f G F satisfying
11/ - ¡\\f < δ,
the inequality
||u — {¿Hυ < ε
holds true. Here Au = / and Aü = / .
The stability of a nonlinear operator A depends, generally speaking,
on the element / ; the operator may be stable for one set of elements and
unstable for another set. In the case of a linear operator A, either stability
or instability takes place for all elements / G F at once.
A mathematical problem of solving (0.3.4) which obeys the require-
ments of existence, uniqueness, and stability of the solution under small
variations of the initial data is called a correct problem. The concept of
well-posedness (correctness) was developed by a prominent French mathe-
matician Hadamard. Below we give a detailed discussion. Let us abstract
ourselves from the concrete nature of the operator A. Let A be an operator
acting from a normed space U to a normed space F.
Definition 0.3.1. The problem of solving (0.3.4) is said to be formu-

lated correctly in the spaces F and U if the solution to the problem
1) exists for every f £ F and belongs to the space U;
2) is unique in the space U;
3) is stable at every element / E F.
Problems of solving (0.3.4) that do not satisfy one or more of the above
requirements will be called ill-posed.
Condition 1 means that the problem should not have redundant data
which make it overdetermined. Condition 2 means that the data should
be sufficient for singling out the unique solution. Condition 3 is associated
with the following circumstance. If the problem is related to a physical
phenomenon, its data cannot be considered as known exactly. We can only
assume that the exact values are approximated arbitrarily close. Conse-
quently, if the solution is not stable under small variations of data, it is not
actually determined at all.
Note that a problem may be correct in one pair of spaces and ill-posed in
another pair. It is clear, for instance, that when the space F is extended, the
requirement that a solution should exist for every f E F may be violated.
In the case of a linear operator A, the problem (0.3.4) is correct in a pair
of Banach spaces U and F if and only if for the operator A there exists a
bounded inverse operator A~l acting from F into U, the domain of definition
of the inverse operator coinciding with the space F.
Below we give an example of a correct problem.
Example 0.3.2. Consider equation (0.1.1) (with c = 1) in the half-plane

(z,t) e R x I f . In this case, it is sufficient to set the Cauchy data
du
η(ζ,0)=φ(ζ), —{ζ,0)=φ(ζ), ζ E R. (0.3.6)
The solution to the problem (0.1.1), (0.3.6) is unique and is given by the
d'Alembert formula
1 1 rz+t
u{z, t) = - [φ(ζ -t) + φ(ζ + t)] + - J φ(ξ) άξ. (0.3.7)
For arbitrary Τ G R+ consider the characteristic triangle

DT = {(z,t) I 0<t<T-\z\}.
Formula (0.3.7) yields that for Φ E C [ - T , T ] and Ψ E C [ - T , T ] the solu-
2 L
tion u(z,t) belongs to the function class C (DT)·

2
The uniqueness of the solution to the problem (0.1.1), (0.3.6) in the class
C (DT)
2 is proved easily. Moreover, it can be shown that the solution given
by (0.3.7) is stable:
MC*(DT) < Κ (|Mlc2[-r,T] + IMIC1[-T,T])· (0.3.8)

In other words, for arbitrary ε > 0 we have |M|C2(Dt) < ε in case the initial
functions φ and φ are small enough (small variations of the functions ψ
and ψ result in small changes of the solution in the norm || · ||c2(dt))·
Introduction 9
Thus, the problem is correct for φ G C2[-T,T], ψ G Cl[-T,Τ], and

u G C2{Dt) for every Τ G R + .
Note that the same problem becomes ill-posed if one assumes that both
functions φ and φ are only continuous, still considering the same space
C2(DT) for solutions. Indeed, the first and third conditions of correctness
are not satisfied in this case.
Now consider an example of an ill-posed problem which remains ill-posed
for arbitrarily smooth initial information. Such a problem was first formu-
lated by Hadamard to emphasize the importance of the third condition of
correctness.
E x a m p l e 0 . 3 . 3 . Let u(x 1,0:2) be a solution to the Laplace equation
Au = 0 (0.3.9)
in the domain G = {{xi,%2) \ xi £ (—π,π); X2 G R + } satisfying the condi-

tions
u(-7T, X2) = u(n, £2) = 0, u(x 1,0) = 0,
du , /s . (0.3.10)
-—(ii,0) = e v sinnxi-
OX 2
It is easy to check that the solution to this problem is
u ( x \ , x 2 ) = — e _ v ^ s i n n x i sinhna;2. (0.3.11)
η
One can prove that the solution to this problem is unique. As τι —> oo
the function e _ v / " sin nari, which represents the data of the problem (0.3.9),
(0.3.10), tends uniformly to zero together with all its derivatives. As is seen
from (0.3.11), for every fixed X2 G M+ the solution to the problem has the
shape of a harmonic curve with arbitrarily big amplitude if η is sufficiently
large. Hence, arbitrarily small changes in the problem data in Ck[—π,π]
at every finite k result in substantial variations in the solution. Thus, the
problem (0.3.9), (0.3.10) is ill-posed due to its instability.
The above examples show that, generally, there exist two types of ill-
posed problems. There are problems that are ill-posed in one set of spaces
but can be made correct in another set. There also exist ill-posed problems
(such as, for instance, (0.3.9), (0.3.10)) which are ill-posed in any normed
spaces whose norms involve a finite number of derivatives.
For problems which are not correct in the classical sense,
Tikhonov (1943) suggested a new notion of correctness which was physi-
cally justified for many applied problems. Within his approach, Tikhonov
considered a very restricted set M C U that is essentially narrower than the

whole of the normed space U. Usually M is a compact set. We shall denote
by R the image AM of the set M in the normed space F under the mapping
realized by the operator A; R = AM.
Definition 0.3.2. The problem of solving (0.3.4) is called conditionally

correct (or well-posed in the Tikhonov sense) if
1) it is known a priori that the solution u exists and belongs to a certain

given set M of the normed space U, i. e., u G M C U\
2) the solution is unique in the set M;
3) for every ε > 0 there exists δ > 0 such that for any f , f £ R = AM
satisfying the condition ||/ — f\\p < δ the inequality ||u — ü\\¡j < ε
holds.
The set M is called a set of correctness (set of well-posedness) of the prob-

lem.
Below we explain the difference between problems well-posed in the
Tikhonov sense and problems well-posed in the classical sense. In the clas-
sical definition of correctness (Definition 0.3.1) the problem data / are as-
sumed to belong to a certain normed space F. The first condition demands
to prove that for every / G F the solution to (0.3.4) exists and belongs to U.
The corresponding condition of correctness in the Tikhonov sense postulates
that the solution to (0.3.4) belongs to the given set M without discussing
the necessary properties of the set R = AM. This is associated with the fact
that for many ill-posed problems the conditions of the belonging / G AM
cannot be verified practically.
There is no difference between studying the questions of uniqueness for
problems well-posed in the Tikhonov sense on the set M and for problems
well-posed in the classical sense on the space U.
Now consider the third condition of correctness. Within the classical
approach the solution to the problem exists for every / G F and belongs to
the space U. It is important that any small change in the element / in the
norm of the space F leads to a small change in the solution in the norm of
the space U. Tikhonov stated that correctness does not require numerous
initial data at all and postulated, instead, that the solution belongs to a
certain set M. In this case any small change in the initial data in the norm
of F can result either in the nonexistence of a solution to the problem or,
Introduction 11
even if the solution still exists, its non-belonging to M. Therefore it would

be reasonable to modify the requirement of continuous dependence of the
solution on the problem data by demanding the continuous dependence for
such variations in / that retain the solution within the set M; this is reflected
in the third condition of correctness.
In connection with the definition of conditional correctness note also that
two of the three conditions of correctness are modified in it. Of course, if in
the case of an unstable problem one can constructively describe the set of
data, i.e., the set R = AM, then it is not reasonable to abandon the first
requirement of correctness in its classical variant.
Many problems that are ill-posed in the classical sense are conditionally
correct. For instance, the problem (0.3.9), (0.3.10) is conditionally correct
in the set of functions u(x 1,2:2), xi G [—π, π], X2 G [0,1], belonging at
every fixed £2 £ [0,1] to the space Ζ,2[—π, π] and satisfying the additional
condition
where C is a given constant.

When the operator A is continuous and M is a compact set, general
theorems of functional analysis yield that if the first and second conditions of
correctness are satisfied, then the third condition is satisfied, too. This result
serves as the basis for defining the conditional correctness of the problem.
A problem unstable throughout the whole space F may be stable on the
set R = AM; the correct type of stability estimation depends crucially on
the set M. Within the new approach to the definition of correctness the
focus point is the uniqueness of the problem solution, i.e., establishing the
uniqueness theorem.
Let us explain why the concept of the conditional well-posedness is nat-
ural in studying many applied and theoretical problems and especially in
studying the inverse problems for differential equations. Consider, e. g., the
inverse problem connected with oscillations of a string. When the density
of the string changes from one point to another, the process of small os-
cillations of the string is described by equation (0.1.1) with nonconstant
parameter c = c(z). Consider the inverse problem of determining the func-
tion c(z) with the help of some additional information on the solution to
the direct problem (0.1.1)—(0.1.3). It follows from the physical statement
of the problem that the solution to the problem should be sought in the
class of positive functions that are bounded from above and below by some
constants. One can assign the values of these constants if one knows the
12 M . M . Lavrentiev et al. Inverse problems .
possible set of materials of which the string can be made. As a result, we

can determine the set M to which the function c(z) belongs.
If the data of the problem are taken from a concrete physical experiment
with a string such that c(z) G M, then, obviously, there exists the actual
function c(z) corresponding to this experiment. Now the question is how to
find this function and whether this information is sufficient for its unique
determination.
It can be easily understood that the theory of ill-posed problems con-
sists not only of inverse problems (though they constitute the widest field of
application of the theory of ill-posed problems). To understand this it suf-
fices to recall such classical problems as differentiation of a function which is
given approximately, or the problem of numerical summation of the Fourier
series. Consider, for example, the problem of differentiating a function f(z)
which is given approximately. Let f[(z) = ui(z), 2 6 [0, π]. The func-
tion f2(z) = fi{z) + sin(nz) differs from the function f\(z) in the norm of
the space C[0, π] by the value ||/i — /2||c[o,7r] = 1 for every η G Ν. But
the derivative fîz) = U2(z) differs from u\(z) in the norm of C[0, π] by
the value ||tii — «2||c[o,jr] = τι which can be arbitrarily large for sufficiently
large n. Therefore, the problem under consideration is ill-posed in this pair
of spaces, since the third condition of Definition 0.3.2 is violated.
However, if the distance between the functions f i and /2 is calculated in
the norm of the space C 1 ^ , π]:
ll/i - /allcip»,*] = m « . {1/iM - / 2 (*)| + \f[(z) - f2(z)|},

z£[ 0,7Tj
and the distance between the functions u\ and 112 is estimated in the norm
of the space C[0, π], then the problem under study is well-posed in this pair
of spaces, since the smallness of the quantity ||/i — f2\\c1[o,w] implies the
smallness of \\ui - w2||c[o^]·
In the sequel, we shall sometimes transcend the frames of the study of
inverse problems for differential equations and shall consider such ill-posed
problems that are not inverse problems for differential equations.
Ill-posed problems are called weakly ill-posed if there exits a pair of
functional spaces whose norm is defined by a finite number of derivatives
and the solution depends by a finite number of derivatives and the solution
depends continuously on data in these spaces. An example of such problem
is the noted problem of differentiation. Ill-posed problems for which such a
pair cannot be found are called strongly ill-posed (Examples 0.1.1, 0.3.3).
Chapter 1.
Some physical motivations of

inverse problems
1.1. INVERSE PROBLEMS OF GEOPHYSICS
One of the widest and most developed application area of the theory of
inverse problems is connected with geophysics. It includes the study of the
Earth's structure, search for natural resources, study of the ocean, etc.
The geophysical methods are based on observations of physical fields
whose distributions in space (and in time, if the fields are nonstationary)
depend on the structure and composition of the Earth and, in particular,
the Earth's crust.
Various methods can be classified according to the physical nature of
the fields involved. In that view we distinguish seismics, which involves
the study of elastic waves; geoelectrics, which involves the study of electric
and electromagnetic fields; gravimetry, which deals with gravitational fields;
magnetometry, which deals with stationary magnetic fields; and radiometry,
which involves the study of the natural or induced radioactivity of rock and
soil.
Geophysical prospecting involves two major types of problems:
- structural problems which deal with the structure of sedimentary de-

posits and the Earth's crust, in particular, the structures that axe
probably associated with oil and gas resources;
- location problems connected with determining local inhomogeneities

in sedimentary deposits and crust that are due to mineral deposits
such as unstructured oil and gas deposits, ore bodies, etc.
Without going into details, here we just mention that we refer to inverse
problems of geophysics as the problems to recover the inner structure of the
Earth (Earth's crust) through the surface measurements of physical fields.
Depending on the prospecting goals, different types of "model structure"
hypotheses are used to solve inverse problems. For structural prospecting,
the basic model is that of a stratified medium in which the medium's pa-
rameter α — ot(x3) is supposed to be a piecewise continuous function of
depth. Usually this parameter is taken to be piecewise constant to describe
a layer-homogeneous medium. A more complicated model is that of a quasi-
stratified medium in which seams with constant αk have variable thickness
hk(xi,x2)·
For location problems, two classes of models are usually considered:
- a homogeneous medium containing a system of local bodies with con-

stant parameters
- a homogeneous medium with a zone with variable parameter a(x),
X = (XI,X2,X3) 6 R 3 .
1.1.1. Inverse problems of seismics
The problems of seismics are understood as the study of elastic vibrations

in the Earth's crust using measurements made at its surface. Elastic fields
carry information about, for instance:
- properties of the field sources;

- the structure of areas of the Earth's interior, where seismic waves travel
through.
So, as was mentioned above, such problems are referred to as inverse

problems.
These problems can be mathematically formulated in several different
ways depending on the nature of the seismic waves (volume, surface, etc.),
on relative dimensions of the irregularities sounded (beam and wave ap-
proximations), and so on. In any case, a description of the Earth's interior
structure results from certain additional information about seismic fields.
The corresponding mathematical problems are ill-posed (see Introduction).
Chapter 1. Some physical motivations 15
The basic problem of geophysics consists in studying the internal Earth's

structure on the basis of observations on its surface or in a near-surface layer.
Consider some examples of the inverse problems of seismics.
At the end of the 19th century the following question was posed in geo-
physics. Is it possible to determine the velocity of propagation of seismic
waves inside the Earth if we know the data on the seismic wave fronts on
the Earth's surface from different earthquakes? It should be noted that
this velocity depends on the elastic properties of the matter. For example,
the velocities of longitudinal waves vp and transverse (shear) waves vs are
connected with the elastic properties of the medium λ, κ, and ρ by the
formulas
where λ and κ stand for the Lamé parameters and ρ is the density of the
material.
Prom the point of view of mechanics, the Lamé parameters λ and κ
and the density ρ characterize completely an elastic substance. Therefore,
one of the most important problems of seismology consists in finding the
propagation velocities vp and vs of longitudinal and transverse seismic waves,
respectively, as functions of spatial location.
The physical statement of the above problem is as follows. In the domain
D C R3 bounded by a surface S we consider the process of propagation of
seismic waves generated by sources of perturbations concentrated at points
x° 6 S; the waves from the sources of perturbations propagate in the domain
D with finite velocity v(x); at the points of the boundary we measure the
"travelling time" τ(χ, x°) that it takes the wave to travel from the source
x° E S to the receiver χ € S.
In the case of an isotropic medium the function r{x, x°) inside the do-
main D satisfies the differential equation of first order (the well-known
eikonal equation)
(1.1.1)
under the condition that

τ{χ,χ°) = 0 ( | χ - Χ ° | ) , ,o
X^X°,
where grad x is the gradient with respect to x. The equation r(x, x°) = t
defines the wave front from the source x° at the time t; and τ(χ°,χ°) = 0.
The mathematical formulation of our problem is as follows.
Inverse Problem 1.1.1 (IP 1.1.1). Given the times τ(χ,χ°), where χ
is an arbitrary point of the surface S and x° runs over some set M C DUS,
it is required to fìnd the wave propagation velocity v(x) inside the domain D
satisfying equation (1.1.1).
This problem is known as the inverse kinematic problem of seismics. In
1905-1907 Herglotz and Wiechert (see (Herglotz and Wiechert, 1905)) con-
sidered the first inverse kinematic problem assuming a spherically symmetric
model of the Earth and, moreover, that the velocity v{x) grows monoton-
ically with depth. They showed the unique solvability of the problem and
obtained explicit formulas for finding the function r/v(r), where r = |:r|. In
alternative cases the inverse kinematic problem, as was shown, may have
multiple solutions. On this basis the first conclusions on the internal struc-
ture of the Earth were made.
The above "pure kinematic approach" has certain limitations due to a
possible ambiguity and instability of solutions. Besides, to extract kinematic
data from real seismograms observed, it is necessary, as a first step, to solve
a very complicated problem of wave correlation (to determine the nature
of waves whose travel times are used in solving the inverse problem). In
complex subsurface geometry this problem of wavefield decomposition to
kinematically isolated waves leads to the loss of its physical meaning.
These limitations of the kinematic approach compel us to use the wave
field on the Earth's surface. It is possible to register, on the Earth's surface,
the displacement vector U(x,t,x°) of surface points as a function of t, too.
The function U(x, t, x°), as a function of the variables χ and t, satisfies the
dynamical system of elasticity
where
and F(x,t,x°) represents external forces. For the stress tensor Τ we have
the following defining relation in the case of an isotropic medium:
T = XtrS-I + 2xS,
where S is the strain tensor defined by the formula
i,j = 1,2,3
and I is the unit matrix of order 3 x 3 .

It is necessary to add to this equation some boundary condition which,

for example, in the case of a free surface S has the form
where η is the outward unit normal vector to S. In some cases it is reason-

able to simulate the earthquake by a function F(x,t,x°) of the type
F{x,t,x°) = δ(χ - x°)f(t,x°),
where δ(χ — x°) is Dirac's mass and f(t,x°) = 0 for t < 0. The func-
tion f{t,x°) may be regarded as an unknown function together with the
Lamé parameters λ, κ, and the density of the medium ρ which describe the
medium. In this case the inverse problem can be formulated as follows:
Inverse Problem 1.1.2 (IP 1.1.2). Find the density of the medium p,
the elastic Lamé parameters λ and κ, and the vector-function of sources
f(t, a;0) in the domain D if the vector u(x, t, x°) is known for χ G S, t € M+
and x° G M C D for a certain M belonging to the boundary S or "near-
boundary" layers of D.
This problem is called the inverse dynamic problem of seismics.
1.1.2. Inverse problems of geoelectrics

The electromagnetic methods in geophysical prospecting use the fact that
alternating electromagnetic fields penetrate into the Earth according to their
electromagnetic properties and so it is possible to determine the rock struc-
ture. For example, the electrical conductivity of granite and basalt is com-
paratively low whereas that of sediments is relatively high. Interaction of
the electromagnetic field with a medium is described by Maxwell's system
of equations
β
rot Η = ε— E + σΕ + j, diveE = pe,
ot
β
rot Ε = —μ—Η, div μΗ = 0,
ot
where E and Η are the intensities of the electric and magnetic fields, respec-
tively; ε and μ are the dielectric and magnetic permeabilities, respectively;
σ stands for the electric conductivity; pe is the density of free charges; and j
is the source of electric oscillations.
Thus, the task of electromagnetic prospecting is to determine the dis-

tribution of electromagnetic parameters ε, μ, and σ within the Earth by
measuring the electromagnetic field at the Earth's surface.
Before the 50's, all electrical prospecting problems were based on steady
electric currents. Dynamic statements involving nonstationary electromag-
netic (EM) fields became available after the well-known paper of Tikhonov
(Tikhonov, 1950).
Electromagnetic methods of geophysical prospecting use both artificial
(generated) EM fields and natural ones (magnetotelluric fields), the sources
of the latter being located in the Earth's magnetosphere and ionosphere.
Natural fields vary slowly in time (with periods from one second to 24 hours)
and, therefore, penetrate deeply into the Earth. Also these fields change
slowly over the Earth's surface since the sources of the magnetotelluric fields
are located high above the Earth. This make it possible to simplify the math-
ematical model of the magnetotelluric field substantially and to consider the
field as depending only on the depth X3 — ζ in a stratified medium. Thus, a
one-dimensional problem is reasonable in view of various applications. Such
model looks like:
e"{ζ) + ίωμ0σ(ζ)β{ζ) =0, ζ G [0, Η],
(e' - iy/ίωμοσΗ e)\ ζ = Η = 0,
where e(z) = E\ {z)/E\ (0) is the relative electric field in the Earth, ω is the
frequency of the field variations over time, μο is the magnetic permeability of
vacuum, and συ is the electric conductivity of the homogeneous half-space
(ζ > Η).
In the above approximation, the inverse problem of magnetotelluric
sounding (MTS) is stated as follows.
Inverse Problem 1.1.3 (IP 1.1.3). Determine the electric conduc-

tivity σ(ζ) from the known frequency dependence of the admittance of the
electromagnetic held
dEi i de
ωμοΕι(0) dz ζ=ο ωμο dz 2=0
measured at the Earth's surface ζ = 0.
The MTS method was proposed by Tikhonov (1946) along with a proper
theoretical background. A uniqueness theorem for the MTS inverse problem
in the class of piecewise analytic functions σ(ζ) was proved. We would also
like to refer to the papers by Cagniard (1953), Kato and Kikuchi (1950),
Rikitake (1950) which were very influential for the further development of
the MTS method.
1.1.3. Inverse problems of gravimetry and magnetometry
The gravitational field at the Earth's surface has a complicated structure,

which is determined by the interior mass distribution. This field is conven-
tionally subdivided into a standard component (the field of a "geoid" with
an averaged density), a regional component which varies slowly across geo-
logical regions (in a large spatial scale, in other words), and an anomalous
component due to local inhomogeneities (sources).
The purpose of gravitational prospecting is to study local sources of
anomaly. For this, the standard field is excluded by conventional methods,
while the regional component is considered as a background over which the
anomalous field can be discerned using an initial treatment.
It is natural that the inverse problems of determining the characteristics
of the field sources from an observed surface field are classified according to
the components of the total field. Hence the inverse problems of Gravimetry
and Magnetometry can be divided into:
- problems of determination to single out ore bodies (or groups of bod-
ies), which result in anomalies;
- problems involving multilayered media or their fragments.
In addition, there is the third important class of problems, namely, to
determine how a potential field could be extended towards its source. This
is also an inverse problem since the procedure of stable extension is used
to obtain approximate information about the depth of occurrence of the
sources in accordance with the location of some characteristic points, such
as the center of mass, or some surface points, etc., of the gravitating bodies.
The problems of magnetometry in their scalar statements are formally
equivalent to those of gravimetry. Many problems of gravitational and mag-
netic prospecting are connected with the so-called inverse potential problem.
Its physical formulation is as follows: the potential created by a body lying
inside some domain bounded by a surface S is known outside this domain;
it is necessary to find the shape and density of the body.
Consider the simplest mathematical problem connected with the inverse
problem of Newton's potential. Suppose that we have a body with the
density p(x) which is located inside a finite domain in the space K3 bounded
by a surface S. Denote the support of p(x) (i. e., the set of points χ such that
p(x) φ 0) as X>, and the characteristic function of D as Xv(x)· Newton's
potential u(x) created by this body satisfies the Poisson equation
Au = - 4 πρ(χ)χτ>(χ). (1.1.2)
Inverse Problem 1.1.4 (IP 1.1.4). Determine the functions p(x)

and Xv(x) if we know the solution to equation (1.1.2) outside S.
Since the solution to (1.1.2) has the form
" M - Jv
f ì \rx -^ ς|
· «-«>.&·«•
the above IP 1.1.4 is reduced to investigation of integral equations of first
kind with respect to the functions p(x) and χζ>(χ). Poincaré (1899) has
shown that the geometrical characteristics of a body and its density can-
not be determined simultaneously and uniquely from the gravitational field.
Therefore, it is necessary to impose some simplifying assumptions about
the density, for example, that the density is known. Under such assump-
tion the first uniqueness theorem was proved. Later the inverse potential
problem in different statements was studied by Lavrentiev (1962), Lavren-
tiev, Romanov, and Shishatskii (1986), Strakhov (1969), Tikhonov and Ar-
senin (1977) and others.
Nowadays the theory of inverse potential problems still attracts much
attention, both in views of the theory and numerical methods of their solu-
tion.
1.1.4. Inverse problems in combined statements
Investigation of inverse problems in combined statements began in the mid-

dle of the fifties of the last century. The interest of geophysicists to this
theme was caused by the fact that there appeared a possibility to analyze
simultaneously different geophysical fields in the study of the Earth's struc-
ture both at the qualitative and quantitative levels by using computers.
The work by Lines, Schultz, and Treitel (1988) should be mentioned
here, in which the numerical study of a possibility of reconstructing the
medium structure was carried out using the data of the wave and gravity
fields given at the surface. In this case it was assumed that there exists a
functional relation between the velocity and density. Two approaches were
compared: minimization of a complex functional taking into account the

relation between the parameters of the medium and sequential minimization
of data misfit functionals for each of the problems with recalculation of the
parameters.
Numerical study of a combined inverse problem of seismics and geo-
electrics was considered in Avdeev, Goryunov, and Priimenko (1996a),
Avdeev, Goryunov, and Skazka (1995) and others.
The problem of multidisciplinary earthquake prediction study deserves
special mention. The main concept of prediction is based on the maximally
reliable prediction of large earthquakes. This leads to the use of the largest
possible number of precursors, which provides a comprehensive control of
the process of preparation of large earthquakes.
However, the use of multifactor criteria makes more complicated the
conditions of reliable application of statistical methods and techniques of
pattern recognition because in this case it is necessary to have a larger vol-
ume of α priori information. Moreover, "logical deadlocks" appear in cases
where the multifactor and multidisciplinary criteria in different combinations
give contradictory estimates of the fact itself of earthquake preparation and
its tempo.
The direct problems of geophysics give only a description of transfor-
mation of the variations of the medium properties into the changes of the
fields-precursors. The reverse procedure is necessary for the earthquake pre-
diction: passing from the fields-precursors to the mechanical parameters of
the medium.
However, in the geophysical monitoring of a seismic zone, each method
of observation can give only insufficient reliability and few details in the
result of solution of the corresponding inverse problem when the time and
coordinates of the expected earthquakes are not definite. Therefore, the
monitoring based on the solution of combined inverse problems using all the
observed geophysical anomalies as initial data is necessary.
Omitting here the geophysical aspect of the problem, we would like to
point out the paper by Alekseev (1992) which is, in our opinion, of substan-
tial mathematical and methodological importance. It turned out that, in
essence, the inverse problems in combined statements provide a possibility
of a more successful solution than the study of each of the inverse problems
separately, taking into account the data obtained in order to get the general
idea of the medium in study.
The property of individual inverse problems to "bring" more information
and "of better quality" into the combined inverse problem than it might ap-
pear at first glance was called by Alekseev the property of complementability

(Alekseev, 1992). He gave a mathematical definition of the combined inverse
problem and showed that it is not equivalent to a simple set of individual
problems. Here it should be pointed out that the widely used term "joint
inversion" is not the same as the solving of the combined inverse problem.
The first term denotes a technical method whereas the second one denotes
a new mathematical problem. There exist many theoretical questions which
are connected with the uniqueness, existence, and stability of solutions of
combined inverse problems. These questions are more difficult as compared
with the same questions concerning individual inverse problems. Here we
have no possibility to discuss in detail their distinctions and would like to
refer the readers to Priimenko and Vishnevskii (2003a).
1.2. INVERSE TOMOGRAPHY PROBLEMS
Tomography is a scientific-technical direction that develops intensively in

recent years. The word originates from the Greek words "tomos" (layer) and
"graphy" (image). This word appeared in the scientific-technical literature
in the twenties of the last century in connection with creation of medical X-
ray tomographs. Intensive development of tomography began in the 70s of
the last century. There are two basic aspects in a modern tomograph. First,
this is a technical device that allows one to obtain the data of X-raying of
an object under examination. Second, this is a transformation of the data
obtained into an image of a layer of the object. The second aspect is based
on mathematical theory.
Though until recently the term tomography was used mainly in con-
nection with the problems of medicine, virtually the same principles have
already been used in geophysics for a long time in reconstruction of sections
of regions of the Earth's crust from the data of geophysical measurements
along a profile.
1.2.1. Mathematical problems of tomography and hyperbolic

mappings
Below, in Chapter 3, we consider a new class of mathematical problems

related to interpretation of tomography data. The main assumption is that
the sought distribution of absorption is an identically one function in the
domain to be determined. These problems are connected with three known
directions of mathematical physics: the Dirichlet problems for hyperbolic
equations, the problems of small oscillations of a rotating fluid, and the

problems of supersonic flows of an ideal gas.
By inverse tomography problems we mean the mathematical problems
that are related to interpretation of tomography data. The first high-quality
tomogram of a human brain was obtained in 1972. Since then tomography
has found rather broad applications in medicine. Progress was made in
industrial tomography, tomography in gas dynamics and plasma physics,
and seismic tomography (see Natterer, 1986; Nikolaev, 1988; Pikalov and
Preobrazhenskii, 1·987).
It is X-ray tomography that is used most widely in medicine and in-
dustry. Interpretation of tomography data in medical X-ray tomography is
connected with the Radon transform. We state the corresponding mathe-
matical problem.
Suppose that u(x,y) is a continuous function compactly supported in
a bounded domain D (we can assume that D is the disk D — {{x,y) |
x2 + y2 < 1}). The Radon transform of u{x,y) is the function
We can consider (1.2.1) as a linear operator equation in u(x, y). It is required

to determine u(x,y) if the function f(x,y,a) is given.
In the mathematical model of X-ray tomography u(x, y) relates to ab-
sorption of X-rays. In medical tomography it is reasonable to consider the
sought function to be an arbitrary continuous or piecewise continuous func-
tion, since different patients possess rather different distributions of absorp-
tion. Moreover, the distribution of absorption of the same patient may vary
depending on whether he or she stands or lies.
In industrial tomography another situation also arouses interest. The
distribution of absorption has usually a standard shape. The goal of the
tomography exploration is to find deviations from this standard shape, flaws
of a product. If an industrial product is obtained by casting, then there
may be some cavities and cracks inside. Thus, we arrive at the following
mathematical problem: the function u(x,y) in (1.2.1) equals unity in some
unknown domain Do C D and vanishes outside Do:
u{x,y) = 1, (x,y)eD0CD·, u{x,y) = 0, (x,y) <£ Do-
Moreover, the boundary of Do is a piecewise smooth curve. Since the func-

tion u(x, y) in this statement is determined by two functions of one variable,
it is natural to suppose that a solution to (1.2.1) can be obtained from an
Figure 1.2.1
essentially less complete information on the right-hand side f(x,y,a) than

in the general case. In this section, we discuss the questions of solving (1.2.1)
in the case where the values of f(x,y,a) are known for two values of the
parameter or; i. e., we suppose that the functions
fifav) = f(x,V,<* i)> h (ζ,ν) = f(z,V,a 2)

axe given (Figure 1.2.1).
The properties of the formulated inverse problem are connected with
the properties of those mappings of the plane under which the families of
straight lines are invariant.
1.2.2. Problems of seismic tomography

Here we would like to single out seismic tomography problems in view of very
wide modern applications. Usually, the term "seismic tomography" stands
for the inverse problems previously referred to as inverse kinematic problems
and even some inverse dynamic problems. Apparently, in the first place this
is due to the fact that the statements of these problems are similar to those
of medical tomography and, secondly, due to the development of systems
and facilities of gathering seismic information since modern systems used
in medicine as far as their geometric scanning characteristics are concerned
(Herman, 1980).
As to mathematical background, all tomography problems are the prob-
lems of integral geometry: to reconstruct a function from given integrals of
it along a given set of curves determined by the locations of sources and

receivers.
However, in practice, the methods and algorithms developed for medical
tomography have been observed to be unsuitable for the seismic tomography
problems. In contrast to the problems of medical tomography, the problems
of seismic tomography are characterized, on the whole, by the nonlinearity
of governing equations (the ray traces depend on the reconstructed velocity),
which results in "curved rays" (refraction) which, in addition, form a non-
regular set. It should also be noted that unlimited scanning of the medium
is not always possible. Besides, practical observational systems do not al-
ways provide total covering of the reconstruction domain with rays. Finally,
seismic tomography problems belong to the class of ill-posed problems and
this is their principal peculiarity.
All these complications require development of new efficient numerical
algorithms and methods based on the principles of geometrical optics.
The inverse kinematic problem appears first in geophysics (see IP 1.1.1).
It arose in connection with the attempt to study the Earth's internal struc-
ture from observations, on the Earth's surface, of the propagation of seismic
wave fronts generated by earthquakes.
Prom the mathematical viewpoint the inverse kinematic problem can be
formulated as the problem of finding the right-hand side of equation (1.1.1)
such that the solution to this equation under the condition τ(χ°,χ°) = 0
takes on the given values at χ e S, x° G M. Here the point x° is the
parameter of the direct problem for equation (1.1.1).
It is important to extend consideration from one-dimensional (ID) mod-
els, where the function ν is dependent on one variable, to multidimensional
models, where ν is essentially dependent on all three coordinates.
Along these lines we would like to mention the linear variant of the
inverse kinematic problem of seismics. The function v(x) is assumed to be
represented in the form
v(z) =v0(x) + v1(x), (1.2.2)
where VQ(X) is supposed to be known and υι(α;) is small as compared

with vo (x). In this case, the problem of determining the small addition v\ (x)
to the function VQ(X) is reduced to the problem of solving an integral equa-
tion of first kind, i. e., the problem of integral geometry.
The main theoretical results concerning the questions of uniqueness and
stability of solutions to such problems were obtained by Bukhgeim (1975,
1983a, 1999, 2000), Lavrentiev and Romanov (1966), Lavrentiev, Ro-

manov, and Shishatskii (1986), Lavrentiev, Romanov, and Vasiliev (1970),
Mukhometov (1977, 1978, 1981), Mukhometov and Romanov (1978), Ro-
manov (1967a, 1969, 1978a, 1978b).
In the nonlinear problem, the first results related to the uniqueness of
solution to the multidimensional inverse kinematic problem were obtained
by Anikonov (1978), Romanov (1973a, 1973b, 1974a, 1974b, 1974c, 1975).
It is also necessary to point out the results of Bernstein and Gerver (1978,
1980), Goldin (1986).
In view of theoretical and applied reasons it is important to specify the
necessary and sufficient conditions upon which the solution of the inverse
kinematic problem exists and belongs to the given functional space. For the
general case this problem is still unsolved.
1.2.3. Numerical methods and algorithms of seismic tomography
Numerical techniques for solving multidimensional seismic tomography

problems can be conventionally divided into three groups:
- computer trial-and-error techniques;
- methods of projecting into a finite-dimensional subspace;
- general numerical methods.
Trial-and-error techniques are based on the possibility of effective solution
of the direct problem for a certain selected class of models of the medium.
The first results see, e.g., Alekseev (1967), Matveeva (1972), Matveeva and
Alekseev (1964).
If the available information is not sufficient, trial-and-error techniques
can give a general idea of the medium structure. State-of-the-art in the
methods of solution of direct problems increases the possibilities of trial-
and-error techniques. Note, however, that this requires high computational
costs and even the best agreement between the calculated and observed data
does not guarantee that the profile obtained is the right one.
The inverse kinematic problem is nonlinear since the geometry of seismic
rays is unknown. Such geometry must be determined in order to compute
the travel times. In seismic investigations an approximate mathematical
model of the inhomogeneous medium under study, i.e., a certain velocity
model corresponding to the refraction coefficient, can be pre-selected on the
basis of "a priori" data or can be obtained by approximate methods based
on selected data from the wave field observed.
Then a nonlinear seismic tomography problem is reduced by lineariza-

tion to a linear one, i.e., to the solution of a linear integral equation of
first kind. This is the usual way, which is used in most works on seismic
tomography. The advantage of this approach is that for the linear inverse
problems, computational techniques are developed on the basis of efficient
algorithms.
One of such methods is the projecting of the infinite-dimensional model
space into a suitably selected finite-dimensional subspace. Combinations of
different forms of vo, which is given, and v\, which is to be reconstructed,
see (1.2.2), constitute that variety of methods and algorithms which are used
in seismic tomography.
If vq is constant, the rays are straight lines and v\ is determined from
the Radon inverse transformation. To implement this transformation, algo-
rithms of back projection a n d algebraic reconstruction techniques are n o r m a l l y
used which are most commonly employed in medical tomography. However,
under conditions of limited scanning geometry, which are specific for seismic
problems, their application is rather difficult.
These methods are used in seismics when studying the medium be-
tween the boreholes (Dines and Lytle, 1979; Ivansson, 1985; McMechan,
1983), in investigation of an inhomogeneous layer by reflected or refracted
waves (Bishop et al., 1985; Neumann-Denzau and Behrens, 1984; Nolet,
1985, 1987), and in determining global horizontal inhomogeneities along the
Earth's surface from the observations of surface waves.
Modifications of the algebraic reconstruction technique and back projec-
tion algorithms are also used in the case where VQ is a one-dimensional func-
tion of depth and in the inverse problem in terms of the generalized Radon
transform (Fawcett and Clayton, 1984; Humphreys, Clayton, and Hager,
1984). In the papers Anderson (1984), Anderson and Dziewonski (1984),
Dziewonski (1984), Nakanishi and Anderson (1982), Nataf, Nakanishi, and
Anderson (1984), Woodhouse and Dziewonski (1984), a spherically sym-
metric model corresponds to the initial profile VQ(X) and the velocity varia-
tion vi (χ) is represented as an expansion in a series with respect to spherical
harmonics.
Under the same assumptions, the first results on the Earth's seismic
tomography were obtained as early as in 1969-1971; see Chapter 6, where
the problem is considered to determine the P-wave propagation velocity
in the upper layers of the Earth's mantle from the earthquake travel-time
curves. In that problem, it is assumed that the Earth without the crust is
a sphere and that we know the travel-times along seismic rays between two
arbitrary points of a certain connected domain on the surface of the sphere.
28 M. M. Lavrentiev et ai. Inverse problems .
The reduction of the seismic tomography problem to an inverse problem

for a differential equation of the Hamilton-Jacobi type makes it possible
to develop the general numerical method aimed at solving the problems of
integral geometry, both linear (Romanov, M. E., 1975, 1983) and nonlinear
(Romanov, M. E., 1972, 1988). This method employs the same common al-
gorithm in which the reconstruction of a medium is accomplished by combi-
nation of downward continuation of data along the rays and local projection
techniques.
Chapter 2.
Approximate methods of solution of
ill-posed problems
2.1. ON SOME ASPECTS OF STATEMENT AND SOLUTION

OF ILL-POSED PROBLEMS
The study of any ill-posed problem connected with physical phenomenon is
carried out on the basis of a certain mathematical model. It means that
the operator A and the normed space U (see Introduction) containing the
sought characteristic ü of the process are chosen. The normed space F for
the right-hand side / of the equation
Aü = f (2.1.1)
is defined by the type of the experimental information which is used to

solve the ill-posed problem (2.1.1). Here we assume that in the normed
space F there exists an element / = Aü which is given as an experimental
measurement. But these measurements are made with some error, therefore,
the element / is unknown. Instead, an approximate element fg such that
| | / j - / | | f < $ is given along with the value of the error δ, which characterizes
the distance between fg and / in the norm of the space F.
Thus, of special importance in the problem concerned axe the questions
of uniqueness and stability of the solution. Indeed, the study of the unique-
ness of the solution to an ill-posed problem gives an answer to the question
whether the experimental information is sufficient for unique determination
of the sought characteristic of the process under study. The problem of

stability is connected with construction of such methods that would allow
us to determine the approximate solutions us, which solve the equations
Aus = Is- Such solutions must be close to the sought function ü. It is
natural to look for these approximations on the basis of the information fs,
which is given approximately, and the number δ.
The basic idea of how to construct the approximate solutions us to ill-
posed problems was proposed in the 50's (see Tikhonov and Arsenin, 1977).
The point is to use the α priori information on the exact solution to the
ill-posed problem (2.1.1) in construction of the approximate solution us. As
a rule, the α priori information is connected with the fact that the unknown
characteristic stands for some physical quantity. This quantity has certain
properties (positiveness, boundedness, the monotonicity of the derivative,
etc.). The α priori information of such kind makes it possible to solve the
desired inverse problem on a certain subspace (or subset) M of the original
space, M C U, on which the solution to the problem is stable. The idea of
reducing the class of possible solutions of an ill-posed problem to some set on
which its solution is stable lies in the foundation of the concept of correctness
by Tikhonov (see Introduction) which was introduced by Lavrentiev (1962),
Lavrentiev, Romanov, and Shishatskii (1986).
As a rule, the set M on which the solution is stable is compact.
However, one often encounters the situation where the a priori informa-
tion on the belonging of the solution to the set M on which the solution
to the problem is stable is absent. To construct approximate solutions to
the ill-posed problem in this case, the fundamental concept of a regularizing
operator is used (see, e.g., Tikhonov, 1963; Tikhonov and Arsenin, 1977).
On the basis of this concept approximate methods of solution of ill-posed
problems are developed, these methods being stable to small perturbations
of initial data. Concrete examples realizing this approach are given in this
chapter.
2.2. SOLUTIONS ON COMPACT SETS.

THE CONCEPT OF A QUASI-SOLUTION
The following theorem (Tikhonov and Arsenin, 1977) is the substantiation

of the approach to solution of ill-posed problems based on singling out a
compact set M C U.
Chapter 2. Approximate methods 31
Theorem 2.2.1. Let a set M of a normed space U be mapped by an

operator A onto a set R (R = AM) of a normed space F in a one-to-one
manner. Let the operator A be continuous on the set M C U, and let M be
a compactum. Then the inverse operator A~l is continuous on the set R.
Proof. Since the set R = AM is the one-to-one image of the set M,

therefore, there exists the inverse operator A"1 on the set R. Let us show
that this operator is continuous. Assume that it is not so. Then there exist
an element /o G R, a constant ε > 0, and a sequence {fn}neN> In G R, such
that lim ||/ n - / 0 | | f = 0, while
n—>oo
||un-«o||tf > e > 0 , (2.2.1)
where un = A~lfn and UQ = Since the sequence {un}neN belongs to
the compact set M, therefore, one can select a subsequence {unm}meN from
the sequence such that linim—>oo ll u n m - u\\u = o, where u Ε M.
Then it follows from inequality (2.2.1) that
\\u — Uollt/ > £·
But A is continuous on the compact set M, and lim ||u nm — u\\u — 0.
Then we have
lim II fnm - /OIIF = l i m \\Aunm - f0\\F - 0.

m—>oo m—>oo
Hence,
\\Au-fo\\F = 0,
which implies Au — Auo = fo-
Since the mapping A is a one-to-one mapping on the set M, therefore,
u = UQ. But this equality contradicts the inequality \\U — t¿o||c/ > ε. Conse-
quently, our assumption at the beginning of the proof was wrong, and the
operator A~l is continuous on R. •
In case the problem (2.1.1) is conditionally well-posed (see Introduction)
on a pair of normed spaces U and F and the set M C U is a compactum,
one can construct an estimate of the conditional stability.
Theorem 2.2.2. Let the operator A be a continuous one-to-one map-
ping of the compact set M of the normed space U on the set R = AM of the
normed space F. Then there exists a function ω (τ) : ω(0) = 0 continuous
at zero and such that for any elements ui,it2 G M the following estimate
takes place:
I k - u2\\u < - AU2\\F)· (2.2.2)
Proof. Assume that such a function ω (τ) does not exist. Then for every
number η € Ν there exist elements u^, u^ G M, a number r n , lim τη = 0,
and a number ε > 0 such that
WAul-AulWpKTn, (2.2.3)
ll"n ~ u n\\u > ε. (2.2.4)
Since the sequences k = 1,2, belong to the compact set M, there-
fore, one can select subsequences {u^m}meN, k = 1,2, such that
lim ||ti* m -u f c ||y = 0, A; = 1,2,

m—>oo
where uk e M, k = 1,2. Then it follows from the estimate (2.2.3) that

H-Aii1 — Aîi2||(/ = 0. Since the operator Λ is a one-to-one map on the compact
set M, we have ul — u2. But inequality (2.2.4) implies that Hu1 - u2\\u >
ε > 0. We have arrived at a contradiction, hence, the initial assumption
was wrong, and the function ω(τ) possessing the necessary properties does
exist. •
Stability estimates of the type (2.2.2) obtained under the assumption
that the solution belongs to some set M are called the conditional stability
estimates. Examples of estimates of this type will be given below.
Consider the case where A is a continuous operator acting from a normed
space U into a normed space F such that the problem of solving equation
(2.1.1) is ill-posed.
Assume that for the exact right-hand side / of equation (2.1.1) there
exists the unique solution ü from some compact set M. But the element /
is unknown, and instead of / an approximate right-hand side fs is given,
along with the value of the error δ, such that | | / — /¿||F < δ. With f¿ and δ
known, it is required to propose a method of construction of approximate
solutions us which would tend to the exact solution ü as δ 0 in the norm
of the space U. Consider the set
Us = {ueU I ||AU-/,||f<J}.
The problem to solve equation (2.1.1) is ill-posed, therefore, we cannot con-

sider an arbitrary element of this set as an approximate solution. But we
have the a priori information that the exact solution belongs to the compact
set M. Therefore, let us narrow down the set Us, taking its intersection with
the compactum M, namely, consider the set U^ = U¡ Γ\ M. It is clear that
the set U ^ is not empty since it contains the exact solution u. Let us show
that the elements of the set U¡* can be treated as approximate solutions to
equation (2.1.1).
Theorem 2.2.3. Under the above notations, we have
lim sup j|€Z — «Hü- - 0.
Proof. Assume that the theorem is not true. Then there exist a se-
quence of numbers {¿n}neN> limôo δη — 0, and a sequence of elements
(u5n}neN) ugn € Ug^, such that ||u — u¿n\\u > ε > 0. But the sequence
{^<5„}neN belongs to the compact set M, therefore, it contains a converg-
ing subsequence {usnk }fceN (for simplicity, we shall denote it as {u^ }fceN)i
lim Htij. — «olii/ = 0, where UQ G M. Then we have
fc->oo
0 < ε < lim ||ü — ||t; = ||ü — «ο||ί7· (2.2.5)
A;->oo
Since ugk 6 therefore, \\Ausk — / ¿ J | f < Consequently, we have
HM) - I\\F = limoo \\Ausk - fs \\F < lim Sk = 0.

k—yoo
This is possible if ||A«o — I\\F = 0. Thus, AUQ = f — AU, and from the
uniqueness of the solution to equation (2.1.1) with the right-hand side / we
have uo = ü. But this equality contradicts inequality (2.2.5). Thus, the
initial assumption was wrong, and the theorem is proved. •
Remark 2.2.1. In fact, the proof of Theorem 2.2.3 coincides with

the proof of Theorem 2.2.1 of the continuity of the inverse operator on
the compact set M. But in contrast to Theorem 2.2.1, it was assumed in
Theorem 2.2.3 that the solution to equation (2.1.1) is unique not on the
whole of the compact set M but only for the exact right-hand side / .
Let us present an example showing the absence of the convergence of all
elements of the set U¿ to the exact solution ü.
Example 2.2.1. Consider the operator A : M R which acts as follows:

Au = u/(u2 + 1). Then the equation Au = f with the exact right-hand side
/ = 0 has the unique solution ü — 0. Let the approximate right-hand side
be fs = δ/2 and the value of the error δ be given instead of / . The set Us
in this case is the set of real numbers satisfying the inequality
Obviously, for every δ > 0 the number u¡ = 2/5 will be a solution to this
inequality. This number does not tend to the exact solution ü = 0 as δ —>• 0.
On the other hand, if as the compact set M we consider the segment [—1,1],
then for a sufficiently small δ the set U^1 is the set of real numbers satisfying
the inequalities
-l + yfr7^ 1 - VI - 9δ2
δ ^ 35 '
and for it Theorem 2.2.3 holds true. Indeed, in this case
lim sup ||it — ü\\u = lim max |u| = 0
U€U6M SÔUEUF
since
—1 + ¿2
lim ; = lim — . »= 0
<5-^0 δ <5->oá[l + V l - ¿ 2 ]
and, similarly,
rlim— - 0.
n
3Í
Prior to proceeding to the concept of quasi-solution, we would like to
stress attention on some difficulties connected with solution of ill-posed prob-
lems on an example of the simplest variant of a classical inverse problem:
the problem with inverse time for the heat equation.
As the direct problem, consider the problem of determining the function
u(z, t) which satisfies the heat equation
^ = ζ e (o,/), te( ο, Γ), (2.2.6)

the boundary conditions
u(0,t) = u(l,t) = 0, t e [0,Τ], (2.2.7)
and the initial condition
u(z,0) = ψ(ζ), ζ e [0,1]. (2.2.8)
Inverse P r o b l e m 2.2.1 (IP 2.2.1). The inverse problem will be un-

derstood as the problem of finding the function φ(ζ) if we know the values
of the solution to the direct problem (2.2.6)-(2.2.8) at the moment t = T
u(z,T)=f(z)i ζ e [0,1]. (2.2.9)
Such problem is ill-posed.

Let us assume that the additional information (2.2.9) is known from an

experiment not exactly but approximately with some error δ
/δ (ζ) = f (ζ) + δ sin (ηπ ζ), π G Ν,

where δ is a small positive number. The exact solution to this problem has
the form
ug(z, t) = u(z, t) + ¿ β ( η π ) 2 ( τ _ ί ) sin (rarz), η G Ν.

The point is that for any given δ and t G [0, Τ), the multipliers can
be arbitrarily large if η is chosen properly. Let us show how this problem can
be made correct by proper choice of norms. The idea is almost evident: one
needs to define the norm || · ||¿? in the space of data of the inverse problem in
such a way that the quantity ¿||sin (ηπζ)\\ρ grows rapidly with the number n,
to balance the above-mentioned exponential multipliers. We shall define the
norm of a function f(x) G F in terms of its Fourier coefficients fn. Let
00
z sin
f( ) = Σ/ra (ηπζ)·
71=1
We define
-[ΐΣίy
1 0 0
n ^2T]1/2, (2.2.10)
π=1
= [j\2{z)àz]1'2 = ML2[qA.
Now the inverse problem is correct. Indeed, if two functions f(z) and fs(z)
differ from each other by the value
oo
/(*) - fs{z) = Σ(Ιη - /in) sin (ηπζ),
η-1
then the corresponding solutions <p(z) and ψξ{ζ) to the inverse problem differ
by the value
00
φ ) - φδ{ζ) = - fsn)en2n2T sin {ηπζ).
η-1
It is clear that
1 00 ^ „2η 2 π 2 τΙ 1/2 < δ
\\ψ - Ψδ\\υ ^= [g
> Σ
1 I (« / η I-Jf<nfin? e 2
71=1
when 11/ - fs\\F < δ.

Random documents with unrelated
content Scribd suggests to you:
Merkki on meidän lampaissa, 6930 pykälä kylän piioissa.
Miekka maassa miehettömän.
Mies rukka soutelee, mieli rukka melailee.
Mies se, joka mielensä hallitsee.
Mies se, joka sanansa pitää.
Mies silmiin, asia mieleen.
Miestä myöten ripeys, taloa myötea kaakku.
Mikä jääpi terältä, se jääpi hamaralta.
Mikä liikkuu, se läikkyy.
Mikä suden suussa, se suden hännässä.
Mikä tänäpäiwänä maksetaan, 6940

sitä huomena ei kysytä.
Milloin miniä kiitetään,

pojan lasta ei polwenaan.
Missä hepo piehtaroipi,

siihen karwa kariseepi.
Mitä enemmin kissaa silitetään,

sen ylemmäksi se häntänsä nostaa.
Mitä selkiänä mielessä,

se juopuneena kielessä.
Mitä siitä lehmästä,
joka maitonsa kaataa.
Mitä sodassa saadaan,

sitä sodassa syödään.
Mitä wanha syöpi, menee kuin hukkaan,

mitä wanha tekee, antaa kuin Jumala.
Monena mies eläissään, koirana kuoltuansa.
Monta maalla wiisasta, kun wahinko merellä tapahtuu.
Monta Mattia maantiellä. 6950
Märkään kaikki tarttuu.
Nauraa hywälle, nauraa pahalle

Suomalaisen sanalasku.
Nauraa lapsi syötyänsä,

wanha yöt lewättyänsä.
Nauroksi köyhä kuolee.
Nauta yksin nautitsepi, seurass' ihminen iloinen.
Neidon mieli miehelle korkiastaki kodosta, mataloihinki

majoihin.
Niin kauan lapsi, kun lattialla, niin kauan poika, kun polwen
päällä.
Niinkuin ruoka syntyy, niin se syödään.

Nuori terwe syötyänsä, wanha yön lewättyänsä.
Nälkä pereihen tora. 6960
Nöyrällä hewolla on hywä ajaa.
Ohut karja ostajalla, wajaa karja waihtajalla.
Oma ruoska, wieras hewoinen.
Oma waiwa opettapi

toisen waiwaa tuntemahan.
On maata raatajalle,
wäljää wettä soutajalle.
Onni osa rehellisen,

riemu woitto rauhallisen.
Opettajan omat jäljet

owat opille kehoitus.
Oppinut tuntepi opit,

wiisas ymmärtää enemmin.
Ota wettä kaatunutta.
Paha luulo petetyllä, 6970

pelko kielensä polttajalla.
Paha paljon lupailia,

hywä puheinsa pitäjä.
Paha puhe pahan suussa,
nuokassa wariksen nimi.
Paha purepi sanoilla,

koira hampailla kowilla.
Paha walo warkahalle, päiwän koitto pettäjälle, paha

walitus pahoille.
Paljon koira haukkuu,

enemmin tuuli kantaa.
Paljon nuorena kaunihita,

paljon wanhana pahoja.
Paljon talossa tekemistä,

paljon koissa koskemista.
Paljon työtä orjuudessa,

enemmin emäntyessä.
Paljon äiti kaunihimpi,

waan ei kuiten armaampi.
Palttina kudettaan kiittää, 6980

leipä hierejauhojansa.
Paniaisiks' köyhän panos menee.
Parain welka maksettuaan.
Parempi hywä sowinto, kun paha ystäwyys.
Parempi kauppa rikki, kun mieli rikki.

Parempi, kun kallis kulta, ompi puhdas omatunto.
Parempi kuolla nuorena, kun nälkään.
Parempi mies kojassa, kun korwessa.
Parempi on katsoa naurawan suuhun, kun itkiän.
Parempi on pio totuutta, kun kyynärä wääryyttä.
Parempi paikka, kun paljas. 6990
Parempi ruumis, kun waate.
Parempi waraa pitää, kun jälkein katua.
Pellon mukaan siemen paras,

luonnon mukaan miehen wirka.
Pian näkyy kullan pilkka,

pian wika wiisahalla.
Pian paha tapahtupi,

kauan haawa kirwelepi.
Pitää kuningaski kulunsa.
Pojasta polwi muuttuu,

latwasta laki kääntyy.
Puhtaus on paras lahja,

siinä on wara waiwaisenki.
Puolihullu puheissansa
wasta on wiisas maatessansa.
Puulla pirtti lämpiäpi, 7000

tawallansa talo eläpi.
Raha rakas rikkahalle,

watsaparka waiwaiselle.
Rehellisen rintaluissa
rauha pysyy puuttumata.
Rikkaan riista kuluu,

köyhän pää menee.
Rinta suoja sotamiehen,

jaloissa jäniksen onni.
Rohkialla waaran woitto,

hätäillä hämmentypi.
Saapi leipä syöjiä, työ tekiöitä.
Sana auttaa, sana painaa.
Se parhain nauraa, ken wiimeiseksi nauraa.
Seipäällä wierasta asiaa ajetaan.
Siinä miestä merkitään, 7010

missä korwat leikataan.
Sitte on huulet hukassa,

kun kieli maata wetääpi.
Sitte on myöhä myllyyn mennä,
kun owat akanat syöty.
Siiwo tarwitaan sikoinki ruoassa.
Sopu tilaa antaa.
Sotamiehen akka, kalamiehen koira,

jääwät rannalle raukumaan.
Sulhaisena ilo suurin,

toiwottaissa onni täysin.
Sulholla suden ikenet,

kosiolla koiran leuat.
Suopi Jumala suuttumaan, waan ei wihastumaan.
Suuhunpäin on sudella hampaat.
Suuren weräjän salpaat, waan et suurta suuta. 7020
Suussa on sääskellä henki.
Suusta neittä, päästä neittä, jaloista jaloa miestä.
Syrjällensä syötyänsä: ken ei palaista paina, sitä pala

painakohon.
Syödään häissä wehnästäki.
Syöminen on saamisessa, elämä kylän nojassa.
Taaksensa panee, edestänsä löytää.

Taidossa kunnia kunki,
waan ei parran pituudessa.
Taidota on mieli tyhmä,

käsi tyhjänä kewiä.
Taito miehellä tawara,

toiwo waimon paras turwa.
Taito näkyy taitawalla 7030

taitawaisista tawoista.
Terwe ruumis, sielun rauha,

tawara on täydellisin.
Terä miekalla teräwä,

kieli häjyllä kamala.
Tie taloon, toinen siwutse.
Tiellä on, tietä kysypi.
Toimesta tunnetaan mieli,

käytöksistä jalo järki.
Toiwo ompi waiwaan woide,

onnen liianki lisäjä.
Tulee itraa sokianki lusikkaan.
Tulee loukku leipään,

jos lapsiki leikatkoon.
Tulee mies meren takaa,
waan ei tule turpeen alta.
Tulee wälistä linnaki. 7040
Tuntuuha kuningas kypärästä.
Turhat tyhmän kerskaukset, työ on taidolle todistus.
Turkki on miehen turwa.
Tyhjää jauhain kiwi tylsä,

häpiäksi halpa puhe.
Tyhmän täytypi hekkua,

wiisas kelpaa hekkumata.
Tyhmät taikoja pitäwät

luonnon laatua tutkimata.
Tyttäret owat talon warkaat.
Tywestä puuhun noustaan.
Työ kiittää tekiätänsä, leipä hierejauhojansa.
Työ on palkkansa wäärti.
Työ warajaa uurasta, ei wäkewää. 7050
Työläs köyhänä koreilla, raskas ylpeillä rahata.
Täi miehessä, kirppu koirassa.
Uusi on aina uhkiampi, ehkä on entinen parempi.

Wahinko wiisaaksi tekee.
Wajaalla wiisas ylpeyttä, ylpeä tyhmyyttä täynnä.
Waiwainen on monta wajaa, ei yhtä sianlihaa, tahi kahta

kaalinpäätä.
Waiwainen waiwaisen powesta muruja poimii.
Waiwainen warresta wääntää,

kun ei saa sormesta nuolta.
Waiwaisella on waski kirkas,

autuaalla hopia tumma.
Walistus on sielun silmä, 7060 raamattu sokean saua,

tunnon teillä kuljettaissa.
Wanha suola janottapi.
Watsa waiwa nälkäisellä, selkä tyhmän omatunto.
Weri wettä sakiampi.
Wieraaseen ihastuu,
takawaraansa peljästyy.
Wähän päiwässä,
paljon wiikossa.
Wälein kissa pojat saapi,

waan sokiat ne syntywätki.
Ylenkatse köyhän arwo, 7067
raha rikkahan Jumala.
*** END OF THE PROJECT GUTENBERG EBOOK SUOMEN KANSAN
SANALASKUJA ***
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 ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
Let us accompany you on the journey of exploring knowledge and

personal growth!
ebookgate.com

Get Inverse Problems of Mathematical Physics Mikhail M. (Jn.) Lavrent'Ev PDF ebook with Full Chapters Now

Uploaded by

Copyright:

Available Formats

Get Inverse Problems of Mathematical Physics Mikhail M. (Jn.) Lavrent'Ev PDF ebook with Full Chapters Now

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Get Inverse Problems of Mathematical Physics Mikhail M. (Jn.) Lavrent'Ev PDF ebook with Full Chapters Now

Uploaded by

Copyright:

Available Formats

Visit https://ebookgate.

com to download the full version and

Inverse Problems of Mathematical Physics Mikhail M.

_____ Click the link below to download _____

Explore and download more ebooks at ebookgate.com

Contemporary Problems in Mathematical Physics 1st Edition

Inverse Problems and Inverse Scattering of Plane Waves

Computer Modelling in Tomography and Ill Posed Problems

Dynamical Inverse Problems of Distributed Systems

Inverse and Ill Posed Sources Problems Yu. E. Anikonov

Differential equations inverse and direct problems 1st

Inverse problems in atmospheric constituent transport 1st

Inverse Problems for Kinetic and Other Evolution Equations

Inverse Problems of Mathematical Physics

M.M. Lavrentiev, A.V. Avdeev,

© Koninklijke Brill NV, 2003

First published in 2003

Printed in The Netherlands by Ridderprint bv, Ridderkerk.

all such applications, so we took a liberty to mention somehow the basic

Chapter 6 contains some examples of numerical solution of the different

The authors axe grateful to Academician of RAS, Professor A. S. Alekseev,

Chapter 1. Some physical motivations of inverse problems 13

Chapter 2. Approximate methods of solution of ill-posed

Chapter 3. Integral geometry problems 53

Chapter 4. One dimensional inverse problems 103

Chapter 5. Inverse problems for the coupled Maxwell

Chapter 6. Numerical solution of inverse problems:

0.1. THE CONCEPT OF MATHEMATICAL SIMULATION

The statement of direct and inverse problems of mathematical physics im-

Ordinary Differential Equations. On the contrary, models with distributed

Example 0.1.1. The equation

describes small oscillations of a string. In this case the function u(z,t)

u(0, t) = u(l, t) = 0, t e K+, (0.1.2)

0.2. D I R E C T A N D INVERSE PROBLEMS

In some statements, the coefficients of the governing equations aie considered

Example 0.2.1. The wave propagation in a vertically-inhomogeneous

ÄT =/(*)S(*1.32), (ii,i2,i)Gl2 xl+, (0.2.3)

Inverse Problem 0.2.1 (IP 0.2.1). Let the wave propagation be

0.3. ON CORRECTNESS OF DIRECT A N D INVERSE

0.3.1. General notes about correct problems

The theory of differential equations states that a differential equation defines

in a cylindrical domain Q = Ω χ R+, where Ω C R3 is a domain bounded by

u(a;,i)=0, χ e S, f G M+, (0.3.2)

and the initial distribution of temperature inside Ω,

it(z,0) =g(x), χβΩ. (0.3.3)

Setting of the initial and boundary conditions is aimed at singling out a

0.3.2. Mathematical definitions

11/ - ¡\\f < δ,

Definition 0.3.1. The problem of solving (0.3.4) is said to be formu-

Example 0.3.2. Consider equation (0.1.1) (with c = 1) in the half-plane

For arbitrary Τ G R+ consider the characteristic triangle

tion u(z,t) belongs to the function class C (DT)·

MC*(DT) < Κ (|Mlc2[-r,T] + IMIC1[-T,T])· (0.3.8)

Thus, the problem is correct for φ G C2[-T,T], ψ G Cl[-T,Τ], and

E x a m p l e 0 . 3 . 3 . Let u(x 1,0:2) be a solution to the Laplace equation

in the domain G = {{xi,%2) \ xi £ (—π,π); X2 G R + } satisfying the condi-

considered a very restricted set M C U that is essentially narrower than the

Definition 0.3.2. The problem of solving (0.3.4) is called conditionally

1) it is known a priori that the solution u exists and belongs to a certain

2) the solution is unique in the set M;

_ Click the link below to download _

ÄT =/()S(1.32), (ii,i2,i)Gl2 xl+, (0.2.3)

ll/i - /allcip»,] = m « . {1/iM - / 2 ()| + \f[(z) - f2(z)|},