Memory-based immigrants for genetic algorithms in dynamic environments

Date

2005

Advisors

Journal Title

Journal ISSN

ISSN

DOI

Volume Title

Publisher

ACM Press

Type

Conference

Peer reviewed

Yes

Abstract

Investigating and enhancing the performance of genetic algorithms in dynamic environments have attracted a growing interest from the community of genetic algorithms in recent years. This trend reflects the fact that many real world problems are actually dynamic, which poses serious challenge to traditional genetic algorithms. Several approaches have been developed into genetic algorithms for dynamic optimization problems. Among these approches, random immigrants and memory schemes have shown to be beneficial in many dynamic problems. This paper proposes a hybrid memory and random immigrants scheme for genetic algorithms in dynamic environments. In the hybrid scheme, the best solution in memory is retrieved and acts as the base to create random immigrants to replace the worst individuals in the population. In this way, not only can diversity be maintained but it is done more efficiently to adapt the genetic algorithm to the changing environment. The experimental results based on a series of systematically constructed dynamic problems show that the proposed memory-based immigrants scheme efficiently improves the performance of genetic algorithms in dynamic environments.

Description

Keywords

Memory-based immigrants, Genetic algorithms, Dynamic optimisation problems

Citation

Yang, S. (2005) Memory-based immigrants for genetic algorithms in dynamic environments. Proceedings of the 2005 Genetic and Evolutionary Computation Conference, 2, pp. 1115-1122

Rights

Research Institute

Institute of Artificial Intelligence (IAI)