A multipopulation cultural algorithm using fuzzy clustering

In this paper, we propose a new model for multimodal function optimization, based on the concepts of cultural algorithms and fuzzy clustering. Fuzzy clustering is used in order to partition a same population into communicating subpopulations which evolve in parallel, whilst cultural exchange ensures preserving population diversity. The ability of this model to identify global/local optima without loss of diversity is demonstrated using several multimodal test functions, including a deceptive example.