Recurrence rates for shifts of finite type

TL;DR

This paper investigates recurrence rates in mixing shifts of finite type with Gibbs measures, establishing conditions for measure zero or one of certain recurrence sets and identifying a new critical threshold, with applications to self-similar dynamics.

Contribution

It introduces new criteria for recurrence set measures and uncovers a previously unknown phase transition threshold.

Findings

Identified conditions for measure zero and one of recurrence sets.

Discovered a new critical threshold for recurrence measure transition.

Applied results to dynamics on self-similar sets.

Abstract

Let $\Sigma_{A}$ be a topologically mixing shift of finite type, let $\sigma:\Sigma_{A}\to\Sigma_{A}$ be the usual left-shift, and let $\mu$ be the Gibbs measure for a H\"{o}lder continuous potential that is not cohomologous to a constant. In this paper we study recurrence rates for the dynamical system $(\Sigma_{A},\sigma)$ that hold $\mu$-almost surely. In particular, given a function $\psi:\mathbb{N}\to \mathbb{N}$ we are interested in the following set $$R_{\psi}=\{{\texttt i}\in \Sigma_{A}:i_{n+1}\ldots i_{n+\psi(n)+1}=i_1\ldots i_{\psi(n)}\textrm{ for infinitely many }n\in\mathbb{N}\}.$$ We provide sufficient conditions for $\mu(R_{\psi})=1$ and sufficient conditions for $\mu(R_{\psi})=0$. As a corollary of these results, we discover a new critical threshold where the measure of $R_{\psi}$ transitions from zero to one. This threshold was previously unknown even in the special case of a non-uniform Bernoulli measure defined on the full shift. The proofs of our results combine ideas from Probability Theory and Thermodynamic Formalism. In our final section we apply our results to the study of dynamics on self-similar sets.