Enhancing the scalability of metaheuristics by cooperative coevolution

  • Authors:
  • Ciprian Crăciun;Monica Nicoară;Daniela Zaharie

  • Affiliations:
  • Department of Computer Science, West University of Timişoara, Romania;Department of Computer Science, West University of Timişoara, Romania;Department of Computer Science, West University of Timişoara, Romania

  • Venue:
  • LSSC'09 Proceedings of the 7th international conference on Large-Scale Scientific Computing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The aim of this paper is to analyze the ability of cooperative coevolution to improve the scalability of population based metaheuristics An extensive set of experiments on high dimensional optimization problems has been conducted in order to study the particularities and effectiveness of some elements involved in the design of cooperative coevolutionary algorithms: groupings of variables into components, choice of the context based on which the components are evaluated, length of evolution for each component Scalability improvements have been obtained in the case of both analyzed metaheuristics: differential evolution and harmony search.