Differential Evolution with Graph-Based Speciation by Competitive Hebbian Rules
Tetsuyuki Takahama, Setsuko Sakai · 2012
Differential evolution (DE) is an evolutionary algorithm and has been successfully applied to optimization problems including non-linear, non-differentiable, non-convex and multimodal functions. However, it is still difficult to solve hard problems such as multimodal problems and problems with ridge structures. in this study, we propose a new speciation method "graph-based speciation" to keep the diversity of the search points and realize the global search. Also, we utilize the species-best strategy that can realize the global search using speciation and the local search around the seeds of species. It is expected that the efficiency and the robustness of DE can be improved by using the strategy. the advantage of the proposed method is shown by solving some benchmark problems including multimodal problems and problems with ridge structures.