Hierarchical distributed genetic algorithms

Genetic algorithm behavior is determined by the exploration/exploitation ba lance kept throughout the run. When this balance is disproportionate,;the p remature convergence problem will probably appear, causing a drop in the ge netic-algorithm's efficacy. One approach presented for dealing with this pr oblem is the distributed genetic algorithm model. Its basic idea is to keep , in parallel, several subpopulations that are processed by genetic algorit hms, with each one being independent from the others Furthermore, a migrati on operator produces a chromosome exchange between the subpopulations. Maki ng distinctions between the subpopulations of a distributed: genetic algori thm by applying,genetic algorithms with different configurations, we obtain the so-called heterogeneous distributed genetic algorithms. In this paper, we present a hierarchical model of distributed genetic algorithms in which a higher level distributed:genetic algorithm joins different simple distri buted genetic algorithms. Furthermore, with the union of the hierarchical s tructure presented and the idea of the heterogeneous distributed genetic al gorithms, we propose a type of heterogeneous hierarchical distributed genet ic algorithms, the hierarchical gradual distributed genetic algorithms. Exp erimental results show that the proposals consistently outperform equivalen t sequential genetic algorithms and simple distributed genetic algorithms. (C) 1999 John Wiley & Sons, Inc.