A new parallel block aggregated algorithm for solving Markov chains

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

In this paper, we propose a new scalable parallel block aggregated iterative method (PBA) for computing the stationary distribution of a Markov chain. The PBA technique is based on aggregation of groups (block) of Markov chain states. Scalability of the PBA algorithm depends on varying the number of blocks and their size, assigned to each processor. PBA solves the aggregated blocks very efficiently using a modified LU factorization technique. Some Markov chains have been tested to compare the performance of PBA algorithm with other block techniques such as parallel block Jacobi and block Gauss-Seidel. In all the tested models PBA outperforms the other parallel block methods.

Original languageEnglish
Pages (from-to)573-587
Number of pages15
JournalJournal of Supercomputing
Volume62
Issue number1
DOIs
Publication statusPublished - Oct 2012

Fingerprint

Block Algorithm
Parallel Algorithms
Markov processes
Markov chain
Block Method
LU Factorization
Gauss-Seidel
Parallel Methods
Stationary Distribution
Iterative methods
Factorization
Jacobi
Scalability
Aggregation
Agglomeration
Iteration
Computing
Model

Keywords

  • Markov chains
  • Parallel block methods
  • Performance evaluation

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software
  • Information Systems
  • Theoretical Computer Science

Cite this

A new parallel block aggregated algorithm for solving Markov chains. / Touzene, Abderezak.

In: Journal of Supercomputing, Vol. 62, No. 1, 10.2012, p. 573-587.

Research output: Contribution to journalArticle

@article{c262d441b4b94aaab836a416e1182165,
title = "A new parallel block aggregated algorithm for solving Markov chains",
abstract = "In this paper, we propose a new scalable parallel block aggregated iterative method (PBA) for computing the stationary distribution of a Markov chain. The PBA technique is based on aggregation of groups (block) of Markov chain states. Scalability of the PBA algorithm depends on varying the number of blocks and their size, assigned to each processor. PBA solves the aggregated blocks very efficiently using a modified LU factorization technique. Some Markov chains have been tested to compare the performance of PBA algorithm with other block techniques such as parallel block Jacobi and block Gauss-Seidel. In all the tested models PBA outperforms the other parallel block methods.",
keywords = "Markov chains, Parallel block methods, Performance evaluation",
author = "Abderezak Touzene",
year = "2012",
month = "10",
doi = "10.1007/s11227-011-0737-7",
language = "English",
volume = "62",
pages = "573--587",
journal = "Journal of Supercomputing",
issn = "0920-8542",
publisher = "Springer Netherlands",
number = "1",

}

TY - JOUR

T1 - A new parallel block aggregated algorithm for solving Markov chains

AU - Touzene, Abderezak

PY - 2012/10

Y1 - 2012/10

N2 - In this paper, we propose a new scalable parallel block aggregated iterative method (PBA) for computing the stationary distribution of a Markov chain. The PBA technique is based on aggregation of groups (block) of Markov chain states. Scalability of the PBA algorithm depends on varying the number of blocks and their size, assigned to each processor. PBA solves the aggregated blocks very efficiently using a modified LU factorization technique. Some Markov chains have been tested to compare the performance of PBA algorithm with other block techniques such as parallel block Jacobi and block Gauss-Seidel. In all the tested models PBA outperforms the other parallel block methods.

AB - In this paper, we propose a new scalable parallel block aggregated iterative method (PBA) for computing the stationary distribution of a Markov chain. The PBA technique is based on aggregation of groups (block) of Markov chain states. Scalability of the PBA algorithm depends on varying the number of blocks and their size, assigned to each processor. PBA solves the aggregated blocks very efficiently using a modified LU factorization technique. Some Markov chains have been tested to compare the performance of PBA algorithm with other block techniques such as parallel block Jacobi and block Gauss-Seidel. In all the tested models PBA outperforms the other parallel block methods.

KW - Markov chains

KW - Parallel block methods

KW - Performance evaluation

UR - http://www.scopus.com/inward/record.url?scp=84865616711&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84865616711&partnerID=8YFLogxK

U2 - 10.1007/s11227-011-0737-7

DO - 10.1007/s11227-011-0737-7

M3 - Article

VL - 62

SP - 573

EP - 587

JO - Journal of Supercomputing

JF - Journal of Supercomputing

SN - 0920-8542

IS - 1

ER -