Introduction to functional analysis
The core text encompasses 287 pages. In addition there are four short appendices, a bibliography counting 67 items, and an index -- all in all 340~pages.\N\NThe core text is divided into 10 chapters, each divided into sub-chapters. Let me now very briefly indicate the way the text is built up: The first three chapters cover the basic corner stones of Banach spaces, including the Uniform boundedness, Open mapping, and Hahn-Banach theorems. Next comes a chapter with focus on dual spaces and reflexivity, before the text turns to a chapter on Hilbert space theory. The authors then present the theory of weak- and weak star topologies followed by a chapter on spectral theory. Again the focus shifts, the last three chapters are on basics of topological vector spaces, an introduction into non-linear analysis, and finally, five selected special topics from Banach space theory.\N\NThe authors indicate in the preface that the chapters up to and including the one on spectral theory may form the content of a first course while the remaining three chapters are intended for a second course. They also emphasize that functional analysis is deeply connected to applied mathematics, and their presentation can perhaps be described as ``an abstract one with the flavour of an applied one. Functional analysis has a long and rich history; each chapter has an ending sub-chapter called Comments and historical notes.\N\NNow it is time for some deeper glimpses into this rich text. Chapter~1 on normed spaces can be said to follow the line: All finite-dimensional normed spaces are complete -- \(L_p\)-spaces -- compactness and dimension (Riesz's lemma) -- examples of separable spaces. Chapters~2.1 and~2.2 present the characterizations of continuity for linear operators together with some illustrating examples. 2.3 proves the Baire category theorem and the Banach-Steinhaus theorem, with an application to bilinear operators. Concise and clear proofs of the Open mapping- and Closed graph theorems form the main content of 2.4 and 2.5. Example 2.4.4 applies the Open mapping theorem, Ascoli's theorem and Riesz' lemma to prove that every closed infinite-dimensional subspace of \(C[0,1]\) contains an \(f\) with \(f'\) not being continuous.\N\NIn 3.1, the ``algebraic Hahn-Banach extension theorem is first proved in the real case (Theorem~3.1) and then extended to the complex case (Theorem~3.2). The normed versions are now immediate as corollaries. Chapter~3.2 discusses complementability and norm-preserving extension of continuous linear operators. In 3.3 the Hahn-Banach theorem is used to show that a normed space is separable when its dual is, and that \(\ell_\infty\) contains all separable normed spaces as subspaces. The separation theorems are thoroughly proved in 3.5 in the real case; the complex versions are left as exercise~3.26.\N\NChapter 4.1 is quickly described; the Riesz representation theorem for \(L_p\), \(1\leq p<\infty\), and its proof. In the much easier 4.2 we learn how and why \(c_0^\ast\) is \(\ell_1\). In 4.3 the bidual is introduced, and thus reflexivity. A detailed proof that \(L_p\), \(1<p<\infty\), is reflexive is given, an important but often neglected ``detail in light of James's famous example of a non-reflexive space isometrically isomorphic to its bidual (this is also commented on in the Comments and historical notes section~4.4).\N\NChapter 5 on Hilbert spaces starts with the basic properties of inner products in 5.1. In 5.2 orthogonality is introduced, the best approximation theorem is proved and applied to orthogonal decomposition of elements in a given subspace. It is emphasized in Corollary~5.2.7 that subspaces of Hilbert spaces are 1-complemented. 5.3 discusses the standard properties of orthonormality, the ``Bessel-Parseval business. In 5.4 we learn about the Gram-Schmidt process and how this produces a fundamental orthonormal set in separable Hilbert spaces \(H\). From this the route is quick to the fact that such an \(H\) ``is just \(\ell_2\), here called the Riesz-Fischer theorem. That the dual of a Hilbert space is given via the inner product (the Fréchet-Riesz theorem), and the consequences of this is the content of 5.5. Lax-Milgram theorems are extremely important in applied mathematics, and Chapter 5.6 presents two versions, with thorough proofs.\N\NChapter 6.1 is just a reminder on how to topologize a set from a family of functions on that set, with values in a topological space. 6.2 then first applies this to give a normed space its weak topology and establishes some fundamental properties, including Mazur's theorem. It is also shown that the norm and the weak topology coincide if and only if the space is of finite dimension, and this is followed up by proving Schur's theorem. 6.2 introduces the weak star topology, compares it to the weak topology, and gives a detailed proof of Alaoglu's theorem (here called the Banach-Alaoglu-Bourbaki theorem). It is obvious that the weak and weak-star topologies coincide when the space is reflexive; in 6.4 the opposite is proved (called Kakutani's theorem), via Helly's lemma and Goldstine's theorem. In 6.5 we first learn when the weak and weak-star topologies are metrizable, and then that \(C[0,1]\) is universal for separable spaces (like the non-separable \(\ell_\infty\) was shown to be, in Chapter~3.3). Chapter~6.6 contains the theory of uniformly convex spaces; it is first proved that uniformly convex spaces are reflexive, and then it is shown that \(L_p\), \(1<p<\infty\), is uniformly convex. 6.6 ends with a proof of the reflexivity of \(L_p\), \(1<p<\infty\), in the real case, without using the Riesz representation theorem. Further and deeper properties of the weak and weak-star topology are discussed in the Comments and historical remarks section.\N\NChapter 7 ends the part of the book that the authors consider the introductory part. 7.1 presents the spectrum of an operator, 7.2 gives the basic properties of compact operators, including Schauder's theorem, and 7.3 is on spectral theory for compact operators (the Fredholm alternative). In 7.4 we learn about self-adjoint operators on a Hilbert space, and in 7.5 about their spectral theory, ending with the Spectral decomposition theorem.\N\NFor the last three chapters I will only indicate the content. Chapter~8 starts with topological vector spaces, then focuses on locally convex vector spaces. Hahn-Banach theorems end the chapter. Chapter~9 on non-linear analysis begins with discussions on continuity of certain functions, first in the norm topology and then in the weak topology. The next themes are Minimization problems, Fixed point theorems (including both Brouwer's and Schauder's fixed point theorems), and Fréchet differentiability before the chapter ends with a short trip into the Bochner integral. The themes of the final Chapter~10 are the Dvoretzky-Rogers theorem, Grothendieck's inequality, Schauder bases and basic sequences, and finally, as a continuation of the previous theme, the Bessaga-Pełczyński selection principle.\N\NThe themes of the four short appendices are Zorn's lemma, general topology, measure and integration, and last but definitely not least, answers and hints for selected exercises.\N\NIt has been a pleasure to read the book. For an introduction to functional analysis there are not too many different ways to do it as some fundamentals just have to be included, and necessarily in a certain logical order. A clear strength of this book, in addition to being well written, is that it takes the time needed to explain in detail some very important points, as I have given examples of above. Also, I am very sure that the answers and hints appendix will be highly appreciated.
This page was built for publication: Introduction to functional analysis
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7021915)
