A density clustering based niching genetic algorithm for multimodal optimization

A new density clustering based niching method for genetic algorithm is proposed in this paper, which is able to identify and track global and local optima for a multimodal function. To prevent the loss of diversity the global selection pressure within a single population is replaced by local selection of a multipopulation strategy. The subpopulations representing species specialized on niches are dynamically identified using density based clustering algorithm on a primordial population. Moreover, a new method is designed for automatically calculating clustering threshold. Finally, the presented algorithm is applied to the optimizations of typical multimodal functions compared with SH and DC algorithms, and the results reveal its efficiency and effectiveness.