Theoretical numerical analysis. A functional analysis framework

From MaRDI portal





This book arises from courses given by the authors at the University of Iowa. It ``covers basic results of functional analysis and applies them mainly to the computations of the solutions to partial differential equations (PDEs) and integral equations. Distribution theory and measure theoretic aspects are bypassed. Chapter one is a standard introduction to linear normed spaces, Hilbert spaces, \(L^p\) spaces. Chapter two is on linear operators. It is seen how the geometric series theorem allows to characterize the solution to integral equations. Then one continues by introducing duality, one proves the Hahn-Banach theorem, the Riesz representation theorem, one defines the adjoint operator and weak convergence. The chapter ends with compact operators (and puts emphasis on integral operators), their spectral properties and characterizes the resolvent operator. Chapter three is on approximation theory. First one defines various types of interpolation (polynomial, piecewise polynomial and trigonometric) and then one deals with best approximation topics (existence, uniqueness, the case of Hilbert spaces which leads to a mean square criterion). Finally, one considers projection operators and error bounds. Chapter four handles an important topic of numerical nonlinear functional analysis namely the solution to operator equations via the Banach fixed-point theorem. After that, generalized derivatives are introduced and Newton's method in Banach spaces is presented. A last section treats the conjugate gradient iteration. Chapter five is a first short course on finite difference methods. Issues of stability and convergence are included. Chapter six is a first short course on Sobolev spaces and one section is devoted to periodic Sobolev spaces. From now on PDEs or integral equations and the functional analytic motivation for their numerical study are considered. Chapter seven deals with elliptic boundary value problems (mainly linear) and weak formulations allow to express the problem in a functional analytic way. Some general existence and uniqueness results are presented. The Lax-Milgram lemma is proved and used for well-posedness of the above-mentioned problem. A boundary value problem of linear elasticity is also studied with this tool. Mixed and dual formulations are presented. The last section deals with a nonlinear version. Chapter eight is a first short course on the Galerkin method. Chapter nine concentrates on the finite element methods. Error estimates are provided. Chapter ten describes elliptic variational inequalities and discusses some approximation techniques (regularization, Lagrange multipliers, numerical integration). Chapter eleven deals with the numerical solution to integral equations of the second kind. Galerkin and collocation methods are considered. The Nyström method is described and an error analysis provided. Product integration and nonlinear versions end the chapter. The last chapter is devoted to boundary integral equations and their numerical solution. A list of references and an index end the book. Exercises are proposed and the chapters end with some bibliographical information. The book embeds numerical analysis in functional analysis and the title is well chosen. It illustrates how functional analysis may appear in applications. This volume can be recommended to advanced undergraduate students in numerical analysis.




Cited in
(only showing first 100 items - show all)








This page was built for publication: Theoretical numerical analysis. A functional analysis framework

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2711447)