Coupling methods for the convergence rate of Markov processes

schedule le lundi 13 janvier 2020 de 17h00 à 18h00

Organisé par : F. Bechtold, W. Da Silva , A. Fermanian, S. Has, Y. Yu

Intervenant : Armand Bernou (LPSM)
Lieu : Jussieu, Salle Paul Lévy (16-26-209)

Sujet : Coupling methods for the convergence rate of Markov processes

Résumé :

Abstract: One of the key questions arising in the study of Markov processes is the computation of the convergence
rate towards the invariant distribution, when such distribution exists. Theories were first developed for the
simpler case of Markov chains. One of the possible approach is the so-called coupling method, in which
one considers two instances of the chain and show that, after some random time which can be studied
precisely, the two chains are equal. I will briefly introduce the strategy and then present an application to
the Markov process describing the dynamic of a collisionless gas enclosed in a vessel with boundary
effects.