Cadeias de Markov: tempo de mistura, cuttoff e redes

Abstract

This work deals with Markov chains in discrete time and finite state space. We treat the convergence of these objects, and define the total variation distance and studied some of their properties , as well as ways of estimating the mixing time, for example, using the eigenvalues of the transition matrix. Present is the one by one relationship between reversible Markov chains and graphs, and how the network theory can help in Markov Chains context. We also define and show some results concerning the so called Cutoff phenomenon, concluding by exhibiting a counter-example due to Aldous.