Towards Interval Type-3 Fuzzy Parameter Adaptation in Metaheuristics

In this paper, fuzzy parameter adaptation in metaheuristics using type-3 fuzzy logic is presented. Previous works in this area have shown the success of utilizing type-2 fuzzy logic for adaptation in several metaheuristics, such as harmony search, cuckoo search, differential evolution and others. However, in this paper an extension for using type-3 fuzzy logic is put forward with the idea of handling in a better way the intrinsic uncertainty in solving search and optimization problems with metaheuristics. The originality resides in the utilization of a type-3 fuzzy system for dynamically obtaining the main parameter values of a particular metaheuristic, with the goal of enhancing convergence and accuracy of the algorithm. The proposed approach is illustrated with several algorithms, but could be extended to other ones in the future.