On the static and dynamic points of view for certain random walks in random environment (Q2565439)

The models of random walk in random environment on \(Z^d\) are considered.The random environment is described by a collection of i.i.d. \((2d)\)-dimensional vectors \((p(e_1),p(e_2),\dots,p(e_{2d}))\), such that \(p(e_i) \in [\kappa,1]\), \(| e_i| =1\), \(1 \leq i \leq 2d\), \(\sum_{i=1}^{2d} p(e_i)=1\), with common law \(\mu\), which specify transition probability of the walk at each site of \(Z^d\). In these models ``the environment viewed from the particle'' naturally defines a Markov chain. The equivalence between static and dynamic points of view means that there exists an invariant measure of this chain which is absolutely continuous with respect to its initial distribution. The main results of the article concern the case of ``non-nestling'' walks for which \(\kappa>0\) and \(\mu\)-a.s. \((\sum_{i=1}^{2d}p(e_i)\cdot e_i) \cdot \ell \geq \eta\), for some \(\ell \in S^{d-1}\) and \(\eta>0\) (the inequality characterizes the local drift of the walk at each site). In this case the equivalence between static and dynamic points of view is proved if \(d \geq 4\) and the ``disorder is low'', i.e. \(\alpha=\sup_{| e| =1} \text{ess} \sup \log \frac{p(e)}{\widetilde{p}(e)}\), is small enough, where \(p(\cdot)\), \(\widetilde{p}(\cdot)\) are independent \(\mu\)-distributed. In the same setting a functional central limit theorem is derived for almost every realization of the environment, i.e. (1) \(X_n/n \to v\), where \(X_n\) is a random walk, \(v\) is deterministic and \(v \cdot \ell>0\); (2) \(n^{-1/2}(X_{\cdot n} -[\cdot n] v)\) converges weakly to the Wiener measure with non-degenerate matrix. Non-elliptic walks are also considered in case \(\kappa=0\) and \(\mu\)-a.s. \(\sum_{i=1}^d p(e_i)=1\), with \(e_i\), \(1\leq i \leq d\), the canonical basis of \(Z^d\). An example where the equivalence between static and dynamic points of view breaks down is given.