Control the Diversity of Population with Mutation Strategy and Fuzzy Inference System for Differential Evolution Algorithm
Jing-Zhong Wang, Tsung-Ying Sun · 2019
This paper proposed a mutation strategy of differential evolution algorithm (DE) with a fuzzy inference system (FIS) to control the diversity of the population. Diversity is an essential issue when studying how to balance between exploitation and exploration in evolutionary duration for evolutionary computation. The proposed mutation strategy integrates conventional mutation strategies. It does not fix the base vector and controls the ratio of mutating toward the global best individual by FIS. The study measures the diversity of population with entropy and sets a decreasing linear function of diversity. From experimental results, the DE with well-controlled diversity of the population has superior performances in 6 out of 15 evaluated functions. The convergence and diversity curves confirm the control ability of FIS. This paper offers DE a promising direction for controlling the diversity of population by mutation strategy with FIS to get superior performance.