A hybrid genetic algorithm for task allocation in multicomputers

Abstract

A hybrid genetic algorithm for the task allocation problem (HGATA) in multicomputers is presented. It minimizes the possibility of premature convergence and finds good solutions in a reasonable time. HGATA includes elitist ranking selection, variable rates for the genetic operators, the inversion operator and hill-climbing of individuals. Hill-climbing is done by a simple heuristic procedure tailored to the task allocation problem. HGATA also makes use of problem-specific information to evade some computational costs and to reinforce favorable aspects of the genetic search at some appropriate points. The experimental results on realistic test cases support the HGATA approach for task allocation.