Resolving Markov chains onto Bernoulli shifts via positive polynomials (Q2709429)

This memoir is divided into two separate yet related parts. In part A, \textit{Resolving Markov chains onto Bernoulli shifts}, the authors give necessary and sufficient conditions for a Markov shift to have an eventually Bernoulli factor corresponding to a right-closing map, and a Bernoulli factor corresponding to different types of maps, namely, a right-closing map, a right-closing map of degree 1 and a regular isomorphism. To achieve their results, they start by replacing the defining stochastic matrix of the Markov chain by a matrix whose entries are polynomials with positive coefficients in several variables. In this representation, a Bernoulli shift is given by a single polynomial \(p\) with positive coefficients. In this way, topological and measure theoretical coding problems transform into combinatorial problems, which are solved with the help of the second part of the memoir. In part B, \textit{On large powers of positive polynomials in several variables}, the authors study the structure of the set \(\text{Log}(p^n)\) for \(n\) large enough, where \(p= \sum_{w\in\mathbb{Z}^k} p_w x^w\), \(w= (w_1,\dots, w_k)\), \(x^w= \prod^k_{i=1} x^{w_i}_i\) and \(p_w\geq 0\) for all \(w\).