Vector-valued transference and maximal ergodic theory in UMD-valued function spaces (Q1915574)

We prove two transference theorems for maximal convolution operators on vector-valued \(L^p\)-spaces. We then present applications of these results towards ergodic theory. In particular, let \(R\) be a distributionally controlled representation of \(G\), a locally compact Abelian group, acting on \(L^1(\Omega, X)\cap L^\infty(\Omega, X)\), where \(X\) is a Banach space while \((\Omega, {\mathcal F}, \mu)\) is an abstract measure space. We show that, for \(p\in [1, \infty)\), if the associated representation \(R^{(p)}\) acting on \(L^p(\Omega, X)\) is strongly continuous, then \(R^{(p)}\) transfers strong-type and weak-type bounds for maximal convolution operators from \(L^p(G, X)\) to \(L^p(\Omega, X)\). The transference theorems hold for any Banach space \(X\); however when seeking ergodic theorems related to singular integral kernels we need to require that \(X\) satisfy the UMD condition introduced by D. Burkholder.