Evolutionary optimization in dynamic environment

The behavior of standard evolutionary algorithm in the case of multi-modal optimization problems meets a major difficulty. It generally converges towards a single optimum point failing to maintain in the population the multiple optima of the problem under consideration. Various methods enrich the standard algorithm to obtain efficient techniques for solving multi-modal problems. These methods mainly consist of increasing the population diversity and of maintaining the promising areas in the search space in order to finally achieve convergence of the population towards the multiple optima. The present paper introduces mmEA, an evolutionary algorithm for multimodal optimization based on multidimensional exploration of the search space. This technique doesn't require any user defined parameter except those specific to standard evolutionary algorithm. Experiments and comparisons with similar techniques from literature, for static and dynamic environment, prove that mmEA technique is promising.