Diffusive estimates for random walks on stationary random graphs of polynomial growth (Q2364140)

It is well known (to the specialist!) that random walks on \(\mathbb Z^d\) exhibit diffusive behaviour (the mathematical expectation of the square of the deviation from the initial position behaves like \(n\)), and more explicitly \textit{N. T. Varopoulos} [Bull. Sci. Math., II. Sér. 109, 225--252 (1985; Zbl 0583.60063)] and \textit{T. K. Carne} [Bull. Sci. Math., II. Sér. 109, 399--405 (1985; Zbl 0584.60078)] showed that a random walk on a graph of polynomial growth exhibits a speed which is slightly super-diffusive. The main purpose of the paper is to improve the Varopoulos-Carne approach using the theory of harmonic functions, which has been proved to produce new results as, for instance, in the case of the central limit theorem. To this end the authors reduce the diffusive estimate to a statement on random walks restricted to finite subgraphs and use the mass transport principle.