Application of Parallel Mixed-Integer Evolution Strategies with Mutation Rate Pooling

Two parallel evolution strategies (PESs) for mixed-integer optimization are presented. Based on the well-known migrationand neighboorhoodmodel and using a mutation rate pool for strategy parameters, both algorithms are applied to the synthesis of multilayer optical coatings (MOCs), especially for the design of an antire ection coating for germanium infrared optics. Even though PESs are no tailored heuristics, i.e., need no optical knowledge or a special starting design, the optimization results con rm theoretical considerations that predict a relationship between the optical thickness and the best achievable mean reectance of the MOC under consideration. Conrming theoretical results and beating a number of other solutions presented in literature clearly demonstrate the robustness of the presented alogrithms.