Notes on functional analysis (Q5901522)

The book is based on a one semester course taught (for the first time in 1987) by the author to second year Master of Statistics students at the Indian Statistical Institute in New Delhi. The readers are assumed to have a solid background in real analysis, linear algebra, measure theory and probability. The book is an almost verbatim record of the lectures; it is divided into 26 lectures, each one corresponding to a 90 minutes classroom session. In its turn, each lecture is broken into small units that are numbered. The book covers the standard material in functional analysis, with emphasis on the spectral theory of Hilbert space operators. The basic results on Banach spaces are treated in the first ten lectures: Banach spaces, new spaces from old ones (direct sums, quotients), the Hahn-Banach theorem, the Uniform Boundedness Principle (UBP), the Open Mapping Theorem, dual spaces and reflexivity, weak topologies and the Alaoglu-Bourbaki theorem (a discussion of the use of nets in topology is included). Besides this standard material, some more specialized topics are included as, for instance, a proof of the Lomonosov invariant subspace theorem. The spectral theory of linear operators is treated in detail, including the fine subdivision of the spectrum, the Riesz theorem on the spectrum of compact operators, and the spectral theorem for Hilbert space operators. To treat this last topic, a short introduction to projection operator valued measures is provided. There are numerous examples from classical analysis (both real and complex) illustrating and motivating the abstract results: the existence of divergent Fourier series as an application of the UBP, the Riesz-Herglotz integral representation theorem, the Müntz theorem, the classical orthogonal polynomials of Hermite, Laguerre and Legendre as examples of complete orthonormal systems, etc. The author gives simple and elegant proofs of many results -- personally, I enjoyed the proofs of the open mapping theorem and of Lomonosov's theorem. There are many exercises, completing the main text with further results or with details to be filled in by the students. There is no reference list at the end of the book, but references to other books and papers are provided within the text. The result is a well written textbook which can be used for teaching functional analysis (the division into lectures could help in planning the course), as well as by students for self-study.