Théorème ergodique pour les opérateurs positifs à moyennes bornées sur les espaces Lp(1 < p < ∞)

From MaRDI portal
Publication:4024950





This paper presents a highly nontrivial extension of the classical pointwise ergodic theorem for sub-Markovian operators to the case of Cesàro bounded operators in \(L^ p\), \(1<p<\infty\). In particular this result can be applied to power bounded operators. The exact formulation is the following: Let \(T\) be a positive linear operator on the space \(L^ p(m)\), \(1<p<\infty\), where \(m\) is a positive \(\sigma\) finite measure, and let \(M_ n\) be the Cesàro sum: \(M_ n={1\over n}\sum^ n_{k=1}T^ k\). Then the following properties are equivalent: (i) The sequence \(M_ n\) is bounded in \(L^ p(m)\); (ii) The sequence \(M_ nf\) converges \(m\) a.e. in \(L^ p(m)\) for any \(f\in L^ p(m)\). The essential step is achieved by proving that if we denote the \(A\) the operator formally defined by \(A={1\over T}(I-\sqrt{I-T})\) then when \(T\) is Cesàro bounded and \(m\) is a probability measure one has \(A\mathbf{1}\leq \text\textbf{1}\) or \(\sup A^{*n}\mathbf{1}\in L^ 1(m)\), where \(A^*\) is the dual operator of \(A\) on \(L^ q\).











This page was built for publication: Théorème ergodique pour les opérateurs positifs à moyennes bornées sur les espaces Lp(1 < p < ∞)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4024950)