28607 articles – 22094 Notices  [english version]

inria-00075444, version 1

Numerical methods in Markov chain modeling

Bernard Philippe a, 1, Yousef Saad () 2, William J. Stewart 3

N° RR-1115 (1989)

Résumé : This paper describes and compares several methods for computing stationary probability distributions of Markov chains. The main linear algebra problem consists of computing an eigenvector of a sparse, usually non-symmetric, matrix associated with a known eigenvalue. It can be also be cast as a problem of solving a homogeneous, singular linear system. We present several methods based on combinations of Krylov subspace techniques, single vector power iteration and relaxation procedures, and acceleration techniques. We compare the performance of these methods on some realistic problems.

  • a –  INRIA
  • 1 :  CALCPAR (INRIA - IRISA)
  • INRIA – CNRS : URA 227 – Institut National des Sciences Appliquées (INSA) - Rennes – Université de Rennes 1
  • 2 :  NASA Ames Research Center (NASA - ARC)
  • NASA
  • 3 :  Department of Computer Science (NCSU)
  • North Carolina State University
  • Domaine : Informatique/Autre
  • Référence interne : RR-1115
 
  • inria-00075444, version 1
  • http://hal.inria.fr/inria-00075444
  • oai:hal.inria.fr:inria-00075444
  • Contributeur : Rapport De Recherche Inria <>
  • Soumis le : Mercredi 24 Mai 2006, 18:13:46
  • Dernière modification le : Lundi 18 Avril 2011, 13:21:49