Detail publikace

Parallel Processing of Genetic Algorithms in Python Language

ŠKORPIL, V. OUJEZSKÝ, V. ČÍKA, P. TULEJA, M.

Originální název

Parallel Processing of Genetic Algorithms in Python Language

Anglický název

Parallel Processing of Genetic Algorithms in Python Language

Jazyk

en

Originální abstrakt

Modern genetic algorithms are derived from natural laws and phenomenons and belong to evolutionary algorithms. Genetic algorithms are, by their very nature, suitable for parallel processing that leads to increased speed and to optimization. The paper deals with selected ways of parallelization of genetic algorithms with subsequent implementation. Parallelization brings an increase in algorithm speed and load distribution, which is compared to a serial model. Python language is used for demonstration. Four Python modules have been selected to provide parallel processing. They are the Global One - Population Master-Slave Model, the One-Population Fine-Grained Model, the Multi-Population Coarse-Grained Model, and the Hierarchical Model.

Anglický abstrakt

Modern genetic algorithms are derived from natural laws and phenomenons and belong to evolutionary algorithms. Genetic algorithms are, by their very nature, suitable for parallel processing that leads to increased speed and to optimization. The paper deals with selected ways of parallelization of genetic algorithms with subsequent implementation. Parallelization brings an increase in algorithm speed and load distribution, which is compared to a serial model. Python language is used for demonstration. Four Python modules have been selected to provide parallel processing. They are the Global One - Population Master-Slave Model, the One-Population Fine-Grained Model, the Multi-Population Coarse-Grained Model, and the Hierarchical Model.

Dokumenty

BibTex


@inproceedings{BUT159755,
  author="Vladislav {Škorpil} and Václav {Oujezský} and Petr {Číka} and Martin {Tuleja}",
  title="Parallel Processing of Genetic Algorithms in Python Language",
  annote="Modern genetic algorithms are derived from natural laws and phenomenons and belong to evolutionary algorithms. Genetic algorithms are, by their very nature, suitable for parallel processing that leads to increased speed and to optimization. The paper deals with selected ways of parallelization of genetic algorithms with subsequent implementation. Parallelization brings an increase in algorithm speed and load distribution, which is compared to a serial model. Python language is used for demonstration. Four Python modules have been selected to provide parallel processing. They are the Global One - Population Master-Slave Model, the One-Population Fine-Grained Model, the Multi-Population Coarse-Grained Model, and the Hierarchical Model.",
  address="IEEE",
  booktitle="2019 Progress in Electomagnetics Research Symposium (PIERS - Rome)",
  chapter="159755",
  howpublished="online",
  institution="IEEE",
  year="2019",
  month="june",
  pages="1--5",
  publisher="IEEE",
  type="conference paper"
}