Unconditional convergence of functional series in problems of probability theory (Q2255688)

The paper is devoted to the exposition of all necessary preliminaries of the author's results about unconditional convergence of series of vectors and of series of random vectors in Banach spaces. From the introduction: ``This paper consists of the introduction, the list of main notions and notation, and two chapters. Each chapter is preceded by information and facts that facilitate the perception of its content. In Chap. 1, we consider the unconditional convergence of (nonrandom) series in Banach spaces. [\dots] We recall well-known definitions of various types of convergence of series, absolutely summing operators, various types of bases, and the type and cotype of Banach spaces. For completeness, we also recall proofs of some properties of unconditionally convergent series (note that we could not find these proofs in sources available to us). We examine the structure of unconditionally convergent series in Banach spaces by using \(p\)-absolutely summing operators. [\dots] By using Hadamard matrices in Banach spaces, we construct series of special type, solve some new problems, and simplify solutions of known problems. We introduce a numerical characteristic of Hadamard matrices, which is essentially used in the study of the convergence of series and is of independent interest. The main theme of Chap. 2 is the study of almost sure unconditional convergence of random series (mainly Gaussian) with values in Banach spaces. [\dots] We prove the condition of the almost sure unconditional convergence of Gaussian series in Banach spaces in terms of covariance operators of their summands. [\dots] We study almost sure unconditional convergence of Gaussian series of the form \(\sum_{k=1}^\infty a_kg_k\), where \((a_k)\) is a sequence of elements in a Banach space and \((g_k)\) is a Gaussian sequence of standard Gaussian (number-valued) random variables and examine its relationship to the geometry of Banach spaces. Further, we clarify some properties of nuclear operators and operators related to them. [\dots.] The results proved here are used in the study of the problem of the \(r\)-absolute decomposability of Gaussian measures in Banach spaces. [\dots] We also consider canonical Gaussian measures in Banach spaces\dots We prove the existence of a Banach space \(X\), which does not contain \(\ell_\infty^n\) uniformly, and a Gaussian random element with values in \(X\), which does not possess an almost surely unconditionally converging decomposition.''