Using Bernoulli maps to accelerate mixing of a random walk on the torus

Abstract

We study the mixing time of a random walk on the torus, alternated with a Lebesgue measure preserving Bernoulli map. Without the Bernoulli map, the mixing time of the random walk alone is $O(1/\epsilon^2)$, where $\epsilon$ is the step size. Our main results show that for a class of Bernoulli maps, when the random walk is alternated with the Bernoulli map $\varphi$ the mixing time becomes $O(|\ln \epsilon|)$. We also study the \emph{dissipation time} of this process, and obtain $O(|\ln \epsilon|)$ upper and lower bounds with explicit constants.