Adaptive Multi-subpopulation based Differential Evolution for Global Optimization

Properly configuring mutation strategies and their associated parameters in DE is inherently a difficult issue. In this paper, an adaptive multi-subpopulation based differential evolution has been proposed and employed for global optimization. In the proposed method, the entire population is firstly adaptively divided at each generation according to a devised population division strategy, which try to partition the population into multiple subpopulations according to the potential of individuals. Then, a suitable mutation strategy along with an appropriate parameter control scheme is introduced and assigned to each subpopulation for evolution, with the purpose of delivering a balanced evolution. The performance of proposed algorithm has been evaluated on CEC'2015 benchmark functions and compared with related methods. The results show that our method can outperform related methods to be compared.