Autonomous local search algorithms with island representation

  • Authors:
  • Adrien Goëffon;Frédéric Lardeux

  • Affiliations:
  • LERIA, University of Angers, France;LERIA, University of Angers, France

  • Venue:
  • LION'12 Proceedings of the 6th international conference on Learning and Intelligent Optimization
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

The aim of this work is to use this dynamic island model to autonomously select local search operators within a classical evolutionary algorithm. In order to assess the relevance of this approach, we will use the model considering a population-based local search algorithm, with no crossover and where each island is associated to a particular local search operator. Here, contrary to recent works [6], the goal is not to forecast the most promising crossovers between individuals like in classical island models, but to detect at each time of the search the most relevant LS operators. This application constitutes an original approach in defining autonomous algorithms.