Modified genetic algorithms based on disparity theory of evolution

Generally, it has been a problem that the searching performance in the initial and final stages of a genetic algorithm (GA) search is not always good because the genetic parameters are normally constant. We apply the disparity theory of evolution proposed by M. Furusawa et al. (1992) to GA searches. In this research, we propose a parity-and-disparity model of GAs, which uses the mutation algorithm of this theory. We also report some simulation results using a knapsack problem.