Parallelizing a global optimization method in a distributed-memory environment

Authors:
Zdzisław Szczerbiński;Stanisław Kowalik
Affiliations:
Polish Academy of Sciences, Institute for Theoretical and Applied Computer Science, Gliwice, Poland;Silesian Technical University, Faculty of Mining and Geology, Institute for Management and Restructuring in Mining, Gliwice, Poland
Venue:
EURO-PDP'00 Proceedings of the 8th Euromicro conference on Parallel and distributed processing
Year:
2000

Citing 4
Cited 0

A global optimization algorithm using stochastic differential equations

ACM Transactions on Mathematical Software (TOMS)
Genetic algorithms + data structures = evolution programs (3rd ed.)

Genetic algorithms + data structures = evolution programs (3rd ed.)
A high-performance, portable implementation of the MPI message passing interface standard

Parallel Computing
Global Optimisation by Evolutionary Algorithms

PAS '97 Proceedings of the 2nd AIZU International Symposium on Parallel Algorithms / Architecture Synthesis

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present research into parallelizing the zone-parallel method of global optimization. The method belongs to the class of genetic algorithms and is briefly described in the paper. Upon introduction to genetic algorithms, parallelization models for genetic algorithms are presented. The subsequent part of the paper is devoted to the global optimization problem of finding sources of tremors in coal mines. First, a short description of the S-P method for localizing hypocenters of tremors is given; the method requires minimizing the error function for hypocenter location. Next, a practical coal-mining example is given where data on a tremor are collected by seismometers and the location of the hypocenter is found by employing the zone parallel method. Experimental results are presented which were obtained from implementing both the sequential and parallel versions of the zone-parallel method in a local area network of Sun Ultra workstations. The results show suitability of the island model of parallelization for this optimization method, as well as disproving usefulness of the master-slave model.