Random ergodic theorems and regularizing random weights (Q1600118)

The authors consider a sequence of variables \(X_k\) (\(k\geq 1)\) of random independent variables defined on a complete probability space. The variables do not have necessarily the same distribution. The main result provides asymptotic estimates for the trigonometric polynomials \[ \sum_{k=1}^N[\exp(2i\pi\alpha X_k(w))-{\mathbf E} \exp(2i\pi\alpha X_k)]. \] (\({\mathbf E}\) denotes expectation.) The estimates are uniform in \(\alpha\in[0,1]\). The authors give applications of these estimates to ergodic theory, namely, to the study of pointwise ergodic averages for random operators. Besides the use of random trigonometric polynomials, the methods involve Gaussian tools and spectral theory. The authors use in particular work of M. Weber. The paper contains only statements with sketches for some of the proofs. The details of the proofs are contained in a forthcoming paper [``Random ergodic theorems and regularizing random weights'', to appear in Ergodic Theory Dyn. Syst.].