A remark about the Orlicz-Pettis theorem and the statistical convergence (Q5962267): Difference between revisions

The authors prove a version of the Orlicz-Pettis Theorem using statistical convergence. If \(A\subset \mathbb{N} \), let \(\left| A\right| \) be the cardinality of \(A\) and if \(n\in \mathbb{N} \), let \(A(n)=\{i\in A:i\leq n\}\). The density of \(A\), \(d(A)\), is \(d(A)=\lim\frac{1}{n}\left| A(n)\right| \) if this limit exists. Let \(X\) be a real Banach space and \(\{x_{n}\}\subset X\). The sequence \(\{x_{n}\}\) is statistically convergent to \(x\in X\) , written \(St\lim x_{n}=x\), if for every \(\epsilon>0\), \(d(\{i\in \mathbb{N}:\left\| x_{i}-x\right\| <\epsilon\})=1.\) The sequence \(\{x_{n}\}\) is statistically weakly convergent to \(x\) if \(St\lim f(x_{n})=f(x)\) for every \(f\in X^{\prime}\). A family of subsets \(\mathcal{F}\) of \(\mathbb{N} \) is natural if \(\mathcal{F}\) contains the finite subsets of \(\mathbb{N} \) and \(\mathcal{F}\) has property (M) if for every disjoint sequence of finite sets \(\{F_{n}\}\) there exist \(B\in\mathcal{F}\) and infinite \(M\subset \mathbb{N}\) such that \(\bigcup_{n\in M}F_{n}\subset B\subset\bigcup_{n\in \mathbb{N}}F_{n}\). Using these concepts the authors establish the following version of the Orlicz-Pettis Theorem. If \(\mathcal{F}\) is a natural family with property (M) and if the series \(\sum x_{i}\) is such that \(\sum_{i\in B}x_{i}\) is statistically weakly convergent for every \(B\in\mathcal{F}\), then \(\sum _{i}x_{i}\) is unconditionally Cauchy. The authors also give characterizations of unconditionally convergent series in terms of statistical (weak) convergence and establish a matrix result related to statistical convergence analogous to the Antosik-Mikusinski Matrix Theorem.