An Adaptive Co-evolutionary Algorithm Based on Genotypic Diversity Measure

Wide study and application exposes some problems of evolutionary algorithms such as premature convergence and poor performance in convergence. In order to overcome these issues, this paper proposes an adaptive co-evolutionary algorithm based on genotypic diversity measure, where adaptive selection, mutation and substitution operators are designed to realize cooperative search among operators and dynamic pairing among sub-populations. The proposed algorithm successfully avoids negative effect brought by a single operator, and takes full advantage of mutual information in sub-populations. Finally, experiments on two benchmark functions are carried out to show the effectiveness of the proposed algorithm.