Get access

Restricted additive Schwarz methods for Markov chains


Michele Benzi, Department of Mathematics and Computer Science, Emory University, Atlanta, GA 30322, USA



The restricted additive Schwarz method is adapted to the problem of computing the stationary probability distribution vector of large, sparse, and irreducible stochastic matrices. Inexact and two-level variants are also considered as well as acceleration by Krylov subspace methods. The convergence properties are analyzed, and extensive numerical experiments aimed at assessing the effect of varying the number of subdomains and the amount of overlap are discussed. Copyright © 2011 John Wiley & Sons, Ltd.