Load balancing and parallel implementation of iterative algorithms for row-continuous Markov chains

Author

Colajanni, M. ; Angelaccio, M.

Author_Institution

Rome Univ., Italy

fYear

1992

fDate

26-29 Apr 1992

Firstpage

157

Lastpage

161

Abstract

Presents the first parallel algorithms for solving row-continuous or generalized birth-death (GBD) Markov chains on distributed memory MIMD multiprocessors. These systems are characterized by very large transition probability matrices, decomposable in heterogeneous tridiagonal blocks. The parallelization of three aggregation/disaggregation iterative methods is carried out by a unique framework that keeps into account the special matrix structure. Great effort has been also devoted to define a general algorithm for approximating the optimum workload. Various computational experiments show that Vantilborgh´s (1985) method is the fastest of the three algorithms on any data set dimension

Keywords

Markov processes; distributed memory systems; iterative methods; mathematics computing; matrix algebra; parallel algorithms; resource allocation; aggregation/disaggregation iterative methods; data set dimension; distributed memory MIMD multiprocessors; generalized birth-death Markov chains; heterogeneous tridiagonal blocks; iterative algorithms; load balancing; optimum workload; parallel algorithms; row-continuous Markov chains; transition probability matrices; Biological system modeling; Gaussian processes; Iterative algorithms; Iterative methods; Load management; Matrix decomposition; Power system modeling; Steady-state; Stochastic processes; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Scalable High Performance Computing Conference, 1992. SHPCC-92, Proceedings.

Conference_Location

Williamsburg, VA

Print_ISBN

0-8186-2775-1

Type

conf

DOI

10.1109/SHPCC.1992.232656

Filename

232656