An averaging principle for stochastic flows and convergence of non-symmetric Dirichlet forms (Q2118843)

Let \(X^\kappa\) be a diffusion process on a Riemannian manifold \(M\) with generator \(A^\kappa=A+\kappa V\), where \(A\) is a second order differential operator, \(V\) is a vector field on \(M\) and \(\kappa\) is a large positive parameter. After an appropriate time change, \(X^\kappa\) is a random perturbation of the dynamical system \(\frac{dx_t}{dt}=V(x_t)\). Suppose that one is able to construct a certain metric space \(\tilde{M}\) and a continuous map \(\pi: M\rightarrow \tilde{M}\). Using non-symmetric Dirichlet forms and their convergence in a sense close to the Mosco convergence, the paper under review mainly obtains that under suitable conditions, \(\tilde{X}^\kappa:=\pi(X^\kappa)\) converges in law towards a diffusion process \(\tilde{X}\) on \(\tilde{M}\) as \(\kappa\rightarrow \infty\). This result permits higher dimensions than \(d=2\), and the case of dimension \(2\) was paid particular attention to in the book [\textit{M. I. Freidlin} and \textit{A. D. Wentzell}, Random perturbations of dynamical systems. Translated from the Russian by J Szücs. 3rd ed. Berlin: Springer (2012; Zbl 1267.60004)]. Furthermore, this result is also applied to a system of \(n\) particles in a turbulent fluid with a shear flow generated by \(V\).