A new parallel algorithm for solving large-scale Markov chains

Research output: Contribution to journalArticle

Abstract

In this paper, we propose a new parallel sparse iterative method (PPSIA) for computing the stationary distribution of large-scale Markov chains. The PPSIA method is based on Markov chain state isolation and aggregation techniques. The parallel method conserves as much as possible the benefits of aggregation, and Gauss - Seidel effects contained in the sequential algorithm (SIA) using a pipelined technique. Both SIA and PPSIA exploit sparse matrix representation in order to solve large-scale Markov chains. Some Markov chains have been tested to compare the performance of SIA, PPSIA algorithms with other techniques such as the power method, and the generalized minimal residual GMRES method. In all the tested models, PPSIA outperforms the other methods and shows a super-linear speed-up.

Original languageEnglish
Pages (from-to)239-253
Number of pages15
JournalJournal of Supercomputing
Volume67
Issue number1
DOIs
Publication statusPublished - Jan 2014

Fingerprint

Parallel algorithms
Parallel Algorithms
Markov processes
Sequential Algorithm
Markov chain
Aggregation
Agglomeration
Minimal Residual Method
GMRES Method
Power Method
Gauss-Seidel
Parallel Methods
Conserve
Matrix Representation
Sparse matrix
Stationary Distribution
Iterative methods
Isolation
Speedup
Iteration

Keywords

  • Aggregation techniques
  • Iterative methods
  • Markov chains
  • Performance evaluation

ASJC Scopus subject areas

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

Cite this

A new parallel algorithm for solving large-scale Markov chains. / Touzene, Abderezak.

In: Journal of Supercomputing, Vol. 67, No. 1, 01.2014, p. 239-253.

Research output: Contribution to journalArticle

@article{97f4caae3cc145029b36b6329b2fc6b6,
title = "A new parallel algorithm for solving large-scale Markov chains",
abstract = "In this paper, we propose a new parallel sparse iterative method (PPSIA) for computing the stationary distribution of large-scale Markov chains. The PPSIA method is based on Markov chain state isolation and aggregation techniques. The parallel method conserves as much as possible the benefits of aggregation, and Gauss - Seidel effects contained in the sequential algorithm (SIA) using a pipelined technique. Both SIA and PPSIA exploit sparse matrix representation in order to solve large-scale Markov chains. Some Markov chains have been tested to compare the performance of SIA, PPSIA algorithms with other techniques such as the power method, and the generalized minimal residual GMRES method. In all the tested models, PPSIA outperforms the other methods and shows a super-linear speed-up.",
keywords = "Aggregation techniques, Iterative methods, Markov chains, Performance evaluation",
author = "Abderezak Touzene",
year = "2014",
month = "1",
doi = "10.1007/s11227-013-0997-5",
language = "English",
volume = "67",
pages = "239--253",
journal = "Journal of Supercomputing",
issn = "0920-8542",
publisher = "Springer Netherlands",
number = "1",

}

TY - JOUR

T1 - A new parallel algorithm for solving large-scale Markov chains

AU - Touzene, Abderezak

PY - 2014/1

Y1 - 2014/1

N2 - In this paper, we propose a new parallel sparse iterative method (PPSIA) for computing the stationary distribution of large-scale Markov chains. The PPSIA method is based on Markov chain state isolation and aggregation techniques. The parallel method conserves as much as possible the benefits of aggregation, and Gauss - Seidel effects contained in the sequential algorithm (SIA) using a pipelined technique. Both SIA and PPSIA exploit sparse matrix representation in order to solve large-scale Markov chains. Some Markov chains have been tested to compare the performance of SIA, PPSIA algorithms with other techniques such as the power method, and the generalized minimal residual GMRES method. In all the tested models, PPSIA outperforms the other methods and shows a super-linear speed-up.

AB - In this paper, we propose a new parallel sparse iterative method (PPSIA) for computing the stationary distribution of large-scale Markov chains. The PPSIA method is based on Markov chain state isolation and aggregation techniques. The parallel method conserves as much as possible the benefits of aggregation, and Gauss - Seidel effects contained in the sequential algorithm (SIA) using a pipelined technique. Both SIA and PPSIA exploit sparse matrix representation in order to solve large-scale Markov chains. Some Markov chains have been tested to compare the performance of SIA, PPSIA algorithms with other techniques such as the power method, and the generalized minimal residual GMRES method. In all the tested models, PPSIA outperforms the other methods and shows a super-linear speed-up.

KW - Aggregation techniques

KW - Iterative methods

KW - Markov chains

KW - Performance evaluation

UR - http://www.scopus.com/inward/record.url?scp=84893655171&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84893655171&partnerID=8YFLogxK

U2 - 10.1007/s11227-013-0997-5

DO - 10.1007/s11227-013-0997-5

M3 - Article

AN - SCOPUS:84893655171

VL - 67

SP - 239

EP - 253

JO - Journal of Supercomputing

JF - Journal of Supercomputing

SN - 0920-8542

IS - 1

ER -