A coupling approach to randomly forced nonlinear PDE's. II (Q1850158): Difference between revisions

A Markov chain in a Hilbert space \(H\), defined by \(u_{n} = S(u_{n-1}) + \eta_{n}\), \(n\geq 1\), is considered, where \(S: H\to H\) is a mapping Lipschitz continuous on bounded sets and \(\{\eta_{n}\}\) is a sequence of independent identically distributed random variables in \(H\). Let \(P\) be the transition probability of the chain \((u_{n})\). In Part I of the paper by \textit{S. Kuksin} and \textit{A. Shirikyan} [ibid. 221, No. 2, 351-366 (2001; Zbl 0991.60056)] sufficient conditions on \(S\) and \(\eta_{n}\) were found for \(P\) to have a unique invariant probability measure \(\mu\). Moreover, it was shown that \(P^{k}\) converges to \(\mu\) with the rate \(\exp(-ck^{1/2})\). In Part II, this result is strengthened by showing that the convergence is exponentially fast.