A Novel Bi-Group Evolutionary Programming

Premature convergence is the fatal shortcoming of traditional evolutionary programming. Based on the analysis of premature convergence of traditional evolutionary programming, a novel Bi-group Evolutionary Programming(BEP) algorithm is proposed. In this algorithm, evolution of two sub-groups is parallel performed with different mutation strategies, and then the group can explore the solution space separately and search the local part in detail all together. Information is exchanged when sub-groups are reorganized. Performance of this algorithm is analyzed in theory. Simulations based on benchmarks confirm that BEP algorithm is better than classical evolutionary programming algorithm in the aspects of global optimization, convergence speed and the robustness.