Detail publikace

GPU-Based Acceleration of the Genetic Algorithm

Originální název

GPU-Based Acceleration of the Genetic Algorithm

Anglický název

GPU-Based Acceleration of the Genetic Algorithm

Jazyk

en

Originální abstrakt

Genetic algorithm, a robust, stochastic optimization technique, is effective in solving manypractical problems in science, engineering, and business domains. Unfortunatelly, executionusually takes long time. In this paper, we study a possibility of utilization consumer-levelgraphics cards for acceleration of GAs. We have designed a mapping of the parallel islandgenetic algorithm to the CUDA software model and tested our implementation on GeForce8800GTX and GTX285 GPUs using a Rosenbrock's, Griewank's and Michalewicz's benchmarkfunctions. Results indicates that our optimization leads to speedups up to seven thousand timescompared to single CPU thread while maintaing reasonable results quality.

Anglický abstrakt

Genetic algorithm, a robust, stochastic optimization technique, is effective in solving manypractical problems in science, engineering, and business domains. Unfortunatelly, executionusually takes long time. In this paper, we study a possibility of utilization consumer-levelgraphics cards for acceleration of GAs. We have designed a mapping of the parallel islandgenetic algorithm to the CUDA software model and tested our implementation on GeForce8800GTX and GTX285 GPUs using a Rosenbrock's, Griewank's and Michalewicz's benchmarkfunctions. Results indicates that our optimization leads to speedups up to seven thousand timescompared to single CPU thread while maintaing reasonable results quality.

BibTex


@inproceedings{BUT35531,
  author="Petr {Pospíchal}",
  title="GPU-Based Acceleration of the Genetic Algorithm",
  annote="Genetic algorithm, a robust, stochastic optimization technique, is effective in
solving manypractical problems in science, engineering, and business domains.
Unfortunatelly, executionusually takes long time. In this paper, we study
a possibility of utilization consumer-levelgraphics cards for acceleration of
GAs. We have designed a mapping of the parallel islandgenetic algorithm to the
CUDA software model and tested our implementation on GeForce8800GTX and GTX285
GPUs using a Rosenbrock's, Griewank's and Michalewicz's benchmarkfunctions.
Results indicates that our optimization leads to speedups up to seven thousand
timescompared to single CPU thread while maintaing reasonable results quality.",
  address="Faculty of Information Technology BUT",
  booktitle="Proceedings of the 16th Conference Student EEICT 2010 Volume 5",
  chapter="35531",
  edition="NEUVEDEN",
  howpublished="print",
  institution="Faculty of Information Technology BUT",
  year="2010",
  month="april",
  pages="234--238",
  publisher="Faculty of Information Technology BUT",
  type="conference paper"
}