Decompositional analysis of Kronecker structured Markov chains
Author
Bao, Y.
Bozkur, I. N.
Dayar, T.
Sun, X.
Trivedi, K. S.
Date
2008Source Title
Electronic Transactions on Numerical Analysis
Print ISSN
1068-9613
Publisher
Kent State University
Volume
31
Pages
271 - 294
Language
English
Type
ArticleItem Usage Stats
77
views
views
21
downloads
downloads
Abstract
This contribution proposes a decompositional iterative method with low memory requirements for the steadystate analysis ofKronecker structured Markov chains. The Markovian system is formed by a composition of subsystems using the Kronecker sum operator for local transitions and the Kronecker product operator for synchronized transitions. Even though the interactions among subsystems, which are captured by synchronized transitions, need not be weak, numerical experiments indicate that the solver benefits considerably from weak interactions among subsystems, and is to be recommended specifically in this case. © 2008, Kent State University.
Keywords
AggregationDecomposition
Disaggregation
Iterative method
Kronecker representation
Markov chain
Multigrid
Kronecker product
Local transitions
Low memory
Markovian
Multi-grid
Numerical experiments
Steady-state analysis
Weak interactions
Markov processes