Differential evolution method integrated with a fuzzy decision-making mechanism and Virtual Mutant agent: Theory and application

Abstract

The current study deals with upgrading the Differential Evolution (DE) method by equipping it with a novel fuzzy decision-making strategy and a new auxiliary agent called Virtual Mutant. The new algorithm is called Fuzzy Differential Evolution incorporated Virtual Mutant (FDEVM) The presented fuzzy strategy employs two nine-rule mapping mechanisms to adjust the internal parameters of the algorithm considering governing conditions of the current problem. Also, the Virtual Mutant providing reasonable attractive and compulsive effects on the other agents, plays important role in increasing the efficiency of the algorithm. These two new enhancements (i.e., adding the fuzzy module and Virtual Mutant) converts the conventional DE algorithm to a self-adaptive and parameter-free approach which can dynamically adjust its search behavior during the optimization process. Consequently, the search performance of the presented FDEVM is tested on a suite of unconstrained mathematical functions and constrained engineering problems. The outcomes are reported and compared with those obtained by six other well-established search techniques. The acquired results indicate that the proposed method could improve the search process in the terms of computational cost, stability and accuracy.