Improvement of IMODE - a Differential Evolution Algorithm - by Replacing the Third Mutation Strategy
Jinghan Hu, Xuanyu Zhu, Chenxi Ye · 2023
In the field of intelligent computing, real parameter single objective optimization is an essential research direction. Recently, a new setting is employed for real parameter single objective optimization. IMODE is a DE algorithm with three mutation strategies suitable for the new setting. Although IMODE has good performance, it can be further improved. In this paper, we replace the third mutation strategy in IMODE - DE/weighted- rand-to-ϕbest/1 - by DE/rand/1. Compared with the original third mutation strategy, it has much lower probability leading to structurally bias in searching. Based on the replacement, we obtain IMODE-II. Experimental results show that our proposed algorithm is very competitive.