Tauberian conditions for almost convergence (Q1041505)

Let \((X,\|.\|)\) denote a Banach space and let \(\Im =t\{ f_{i},i\geq 0\} \) denote a sequence in \(X\). The de la Vallée-Poussin mean and the Cesàro mean are defined as \(V_{n,N}(\Im )=\frac{1 }{N}\sum_{k=n}^{n+N-1}f_{k}\) and \(\sigma _{n}(\Im )=V_{0,n+1}(\Im )\). The authors consider the following limit statements. In each case \(f\in X\). (A) \(f_n@>\|.\|>>f\) if \(\|f_n-f\|\rightarrow 0\) as \(n\rightarrow \infty \); (B) \(f_n@>ac>>f\) if \(\sup_{n\geq 0}\|V_{n,N}(\Im )-f\|\rightarrow 0\) as \(N\rightarrow \infty \); (C) \(f_n@>w-ac>>f\) if \(\sup_{n\geq N}\|V_{n,N}(\Im )-f\|\rightarrow 0\) as \(N\rightarrow \infty \); (D) \(f_n@>(C,1)>>f\) if \(\|\sigma _{n}(\Im )-f\|\rightarrow 0\) as \(n\rightarrow \infty \). It is clear that (A) implies (B), (C) and (D) and the author proves that (B) is equivalent to (C). In the paper the author proves several Tauberian theorems. Among others, he proves: (1) (A) \(\Longleftrightarrow\) (B) and \(\|f_{n+1}-f_{n}\|\rightarrow 0\); (2) (A) \(\Longleftrightarrow\) (D) and \(n\|\sigma _{n+1}(\Im )-\sigma _{n}(\Im )\|\rightarrow 0\). As an application the author considers the spaces \(L^{p}(T)\) and \(C(T)\) and the sequence \(f_{n}=s_{n}(f)\), where \(s_{n}(f)\) denotes the \(n\)th partial sum of the Fourier series of \(f\).