Limiting distributions and large deviations for random walks in random environments

TL;DR

This thesis investigates the limiting behaviors and large deviations of random walks in random environments, revealing conditions for quenched and annealed distributions and their properties in different dimensions.

Contribution

It provides new results on quenched CLTs with environment-dependent centering and characterizes annealed large deviation rate functions for non-nestling environments.

Findings

Quenched CLT with environment-dependent centering in 1D RWRE

No quenched limit distribution when annealed limit is non-Gaussian

Analyticity of annealed large deviation rate function near its zero

Abstract

This thesis concerns the study of random walks in random environments (RWRE). Since there are two levels of randomness for random walks in random environments, there are two different distributions for the random walk that can be studied. The quenched distribution is the law of the random walk conditioned on a given environment. The annealed distribution is the quenched law averaged over all environments. The main results of the thesis fall into two categories: quenched limiting distributions for one-dimensional, transient RWRE and annealed large deviations for multidimensional RWRE. The analysis of the quenched distributions for transient, one-dimensional RWRE falls into two separate cases. First, when an annealed central limit theorem holds, we prove that a quenched central limit theorem also holds but with a random (depending on the environment) centering. In contrast, when the annealed limit distribution is not Gaussian, we prove that there is no quenched limiting distribution for the RWRE. Moreover, we show that for almost every environment, there exist two random (depending on the environment) sequences of times, along which random walk has different quenched limiting distributions. While an annealed large deviation principle for multidimensional RWRE was known previously, very little qualitative information was available about the annealed large deviation rate function. We prove that if the law on environments is non-nestling, then the annealed large deviation rate function is analytic in a neighborhood of its unique zero (which is the limiting velocity of the RWRE).