“…Each particle in DE corresponds to a solution vector, and the main process is composed of three steps: mutation, crossover, and selection. DE algorithm has many advantages, such as robustness, reliability, simple algorithm structure and few control parameters, etc., and it has been widely applied in global optimization ( Zhao et al, 2020 ; Ge et al, 2017 ), artificial intelligence ( Zhao et al, 2021 ; Tang, Zhang & Hu, 2020 ), bioinformatics ( Zhang et al, 2020a ), signal processing ( Yin et al, 2020 ; Zhang et al, 2017 ), machine design ( Zhou et al, 2018 ), and other fields ( Ren et al, 2019 ; Tang et al, 2020b ). Generation strategy of trial vector and setting method of control parameters will greatly affect the performance of DE algorithm.…”