Performance Evaluation and Population Reduction for a Self Adaptive Hybrid Genetic Algorithm (SAHGA)

Felipe P. Espinoza¹, Barbara S. Minsker¹, and David E. Goldberg²

¹University of Illinois,
Department of Civil and Environmental Engineering,
205 N. Matthews Ave,
Urbana IL 61801
{fespinoz,minsker}@uiuc.edu

²University of Illinois,
Department of General Engineering,
104 N. Matthews Ave,
Urbana IL 61801
deg@uiuc.edu

Abstract. This paper examines the effects of local search on hybrid genetic algorithm performance and population sizing. It compares the performance of a self-adaptive hybrid genetic algorithm (SAHGA) to a non-adaptive hybrid genetic algorithm (NAHGA) and the simple genetic algorithm (SGA) on eight different test functions, including unimodal, multimodal and constrained optimization problems. The results show that the hybrid genetic algorithm substantially reduces required population sizes because of the reduction in population variance. The adaptive nature of the SAHGA algorithm together with the reduction in population size allow for faster solution of the test problems without sacrificing solution quality.

LNCS 2723, p. 922 ff.

Full article in PDF

lncs@springer.de
© Springer-Verlag Berlin Heidelberg 2003