Adaptive genetic algorithm based on density distribution of population

The control parameters in evolutionary algorithms (EAs) have significant effects on the behavior and performance of the algorithm. Most existing parameter control mechanisms are based on either individual fitness or positional distribution of population. This paper proposes a parameter adaptation strategy which aims at evaluating the density distribution of population as well as both the fitness values comprehensively, and adapting the parameters accordingly. The proposed method partitions the individuals into clusters according to their positional distribution. In order to depict the density distribution of population, a variable termed relative cluster density is proposed. Rules are used to modify the values of px and pm based on the relative cluster density and the relative sizes of clusters containing the best and the worst individuals. Experiments are conducted on a set of benchmark functions to investigate the performance and behavior of the proposed method, and the results show that the strategy is promising.