Dominance learning in diploid genetic algorithms for dynamic optimization problems

This paper proposes an adaptive dominance mechanism for diploidy genetic algorithms in dynamic environments. In this scheme, the genotype to phenotype mapping in each gene locus is controlled by a dominance probability, which is learned adaptively during the searching progress and hence is adapted to the dynamic environment. Using a series of dynamic test problems, the proposed dominance scheme is compared to two other dominance schemes for diploidy genetic algorithms. The experimental results validate the efficiency of the proposed dominance learning scheme.