Publication Details

Parallel Genetic Algorithm on the CUDA Architecture

POSPÍCHAL Petr, JAROŠ Jiří and SCHWARZ Josef. Parallel Genetic Algorithm on the CUDA Architecture. In: Applications of Evolutionary Computation. Lecture Notes in Computer Science, vol. 6024. Berlin Heidelberg: Springer Verlag, 2010, pp. 442-451. ISBN 978-3-642-12238-5.
Czech title
Paralelní genetický algoritmus na CUDA architektuře
Type
conference paper
Language
english
Authors
Pospíchal Petr, Ing. (DCSY FIT BUT)
Jaroš Jiří, doc. Ing., Ph.D. (DCSY FIT BUT)
Schwarz Josef, doc. Ing., CSc. (DCSY FIT BUT)
Keywords

massively parallel, genetic algorithm, island model, CUDA, migrations

Abstract

This paper deals with the mapping of the parallel island-based genetic algorithm with unidirectional ring migrations to nVidia CUDA software model. The proposed mapping is tested using Rosenbrock's, Griewank's and Michalewicz's benchmark functions. The obtained results indicate that our approach leads to speedups up to seven thousand times higher compared to one CPU thread while maintaining a reasonable results quality. This clearly shows that GPUs have a potential for acceleration of GAs and allow to solve much complex tasks.

Published
2010
Pages
442-451
Proceedings
Applications of Evolutionary Computation
Series
Lecture Notes in Computer Science
Volume
6024
Conference
European Conference on the Applications of Evolutionary Computation , Istanbul, TR
ISBN
978-3-642-12238-5
Publisher
Springer Verlag
Place
Berlin Heidelberg, DE
DOI
10.1007/978-3-642-12239-2_46
UT WoS
000279390200046
EID Scopus
2-s2.0-77952385474
BibTeX 
@INPROCEEDINGS{FITPUB9184,
   author = "Petr Posp\'{i}chal and Ji\v{r}\'{i} Jaro\v{s} and Josef Schwarz",
   title = "Parallel Genetic Algorithm on the CUDA Architecture",
   pages = "442--451",
   booktitle = "Applications of Evolutionary Computation",
   series = "Lecture Notes in Computer Science",
   volume = 6024,
   year = 2010,
   location = "Berlin Heidelberg, DE",
   publisher = "Springer Verlag",
   ISBN = "978-3-642-12238-5",
   doi = "10.1007/978-3-642-12239-2\_46",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/9184"
}
Files
Back to top