Two Consequences of Brunel's Theorem
Let \(\{T_ k\}_{k=1}^ n\) be a set of linear operators on \(L_ p(\mu)\), \(1<p<\infty\), where \(\mu\) is a probability measure. For each \(f\in L_ p\) let \(M(T_ 1,...,T_ n;f)\) be the function defined as \(M(T_ 1,...,T_ n;f) + sup_{m_ 1...m_ n}| A(m_ 1,...,m_ n;T_ 1,...,T_ n;f)|\), where \[ A(m_ 1,...,m_ n;T_ 1,...,T_ n;f)=(m_ 1...m_ n)^{-1}\sum_{k_ 1=0}^{m_ 1- 1}...\sum_{k_ n =0}^{m_ n-1}T_ 1^{k_ 1}...T_ n^{k_ n}f. \] If there exists a constant \(C\) such that \(\parallel M(T_ 1,...,T_ n;f)\parallel_ p \leq C \parallel f \parallel_ p\) for all \(f\in L_ p\), \(\{T_ k\}_{k=1}^ n\) is said to admit a dominated estimate. If \(T_ 1,...,T_ n\) are majorized by positive power-bounded operators \(S_ 1,...,S_ n\) for which the pointwise ergodic theorem holds, then \(\{T_ k\}_{k=1}^ n\) admits a dominated estimate so that \(lim_{m_ 1,...,m_ n\to \infty}A(m_ 1,...,m_ n;T_ 1,...,T_ n;f)\) exists a.s. for all \(f\in L_ p\). Further, the pointwise convergence for each \(T_ k\) can be taken along uniform sequences. These are shown as consequences of a theorem of Brunel.
This page was built for publication: Two Consequences of Brunel's Theorem
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3978882)
