An adaptive mutation method for GA based on relative importance

A new adaptive mutation method, which uses the information of relative importance of chromosomes and alleles, is proposed for genetic algorithm(GA). In each generation, suitable chromosomes for mutation are automatically choosed based on cumulative distribution function of chromosomes fitness, without requiring to set the mutation probability anymore. After selecting chromosomes for mutation, the standard deviations of allele of all loci are used to decide the number of mutational loci and choose appropriate loci to undergo mutation. To protect good genes on a fit chromosome, the fitness information is included into loci selection by weighting the standard deviations of allele with fitness probability. The results of benchmark testing show the mutation method always has better results and higher speed than traditional mutation method.