The effect of population structure on the rate of convergence of genetic algorithms
Po Hsiang Chu, Susan A. Dudley · 1993
In this paper, we compare the performance of genetic algorithms using several different population structures: one large population of 200 individuals; 20 subpopulations of 10 individuals with or without migration; and 20 subpopulations of 10 individuals with periodic migration.We show that the performance of these population structures is dependent on the geome~of the given problem's search space.We apply these structures to several problems with different solution space structures.We address the following issues: (1) does population structure affect the rate of convergence to the solution?; and (2) if so, does the best population structure depend on the nature of the problem?