No label defined (Q4040244)

[For parts I and II (1984) see Zbl 0573.65001 and Zbl 0573.65002.] This book is based on the notes for a series of lectures given at universities and at a number of courses for professionals to provide an advanced level with basic knowledge of numerical analysis. It presents a great deal of approximation and iteration methods for numerical solutions of partial differential equations and inequalities, both of stationary and evolution problems. In particular, elliptic, quasi-elliptic and parabolic problems are considered. Special attention is paid to continuous problems as they arise in practice. The book consists of the following sections: table of contents, short preface, 7 Chapters, bibliography (each chapter concludes a list of references, at the end of the book there is a general (complementary) list of references) and subject index. Chapter 1 represents some background information and relevant notions of functional analysis scattered in the literature. First some results in Hilbert space setting are introduced: Riesz theorem, Lax-Milgram theorem; the notion of weak compactness and the notion of the projection on a closed convex set. Afterwards several results including the projection theorem of Helly and Milman, the closed graph theorem and the Brouwer theorem are demonstrated for Banach spaces. Chapter 1 is completed with the convolution and distribution theory. This chapter provides a sufficient background. To give the reader a better understanding of the foundations needed for representation of the material of this book all the theorems are given with proof. Chapter 2 treats the theory of nonlinear monotone-type equations and inequalities and shows their close ties with the problem of optimization, both in differentiable and nondifferentiable cases. Chapter 3 deals with the problem of numerical approximation of linear and nonlinear equations and variational inequalities of monotone type. In order to enlarge the region of application several strategies are suggested, for instance the elliptic regularization and the conversion of constraints into penalty functions. Chapter 4 is devoted to the problem of discretization (widespread finite element methods are left out of the discussion) and direct methods for solving stationary problems. Chapter 5 concerns primarily the Galerkin method for solving linear parabolic evolution problems (within the Galerkin frame-work one can generate finite-element, finite differences and spectral methods. An approximation method using finite difference for space and time variables is investigated for solving a nonlinear monotone-type evolution problem. Chapter 6 discusses an iterative method for solving discretized problems and the error propagation. More precisely, the author deduces an overall estimate of the error in the solution and gives its asymptotic formula assuring the convergence of the total error to zero. In Chapter 7 the results of the previous ones are applied to solution of a linear parabolic problem and another nonlinear problem. Furthermore the estimate of total error is specified in case the effect of errors is expressed merely as a function of discretization steps of time and space variables. Much of the material of this book reports on work by the author. It is meant to be useful to several types of readers at a high level. It contains many figures, examples and exercises and presents methods that are easily used on computers. Turbo Pascal implemenation of these algorithms are available. In the reviewer's opinion this book is an excellent monograph, comprehensive and complete, and it is very useful for all the people who are working in this very important field of mathematics.