Abstract : This article illustrates how reward backward coupling improves simulation complexity for the estimation of stationary rewards. Bounds on the coupling time for M/M/1/C are given and experimental results on a large queueing network validate the practical interest of such an approach.
Type de document :
Communication dans un congrès
International Workshop on Applied Probability, 2008, Compiègne, France. 2008
https://hal.inria.fr/hal-00953625
Contributeur : Arnaud Legrand
<>
Soumis le : lundi 10 mars 2014 - 16:44:58
Dernière modification le : mercredi 7 octobre 2015 - 01:16:32
Document(s) archivé(s) le : mardi 10 juin 2014 - 10:37:30
Jean-Marc Vincent. Perfect sampling of stationary rewards of Markov chains. International Workshop on Applied Probability, 2008, Compiègne, France. 2008. 〈hal-00953625〉
