Adaptive differential evolution with coordinated crossover and diversity-based population
Chunmei Zhang, Zhicheng Zhao, Tiemei Yang, Bingyao Fan · 2016
As a population-based optimizer, the differential evolution (DE) algorithm has a very good reputation for its competence in global search and numerical robustness. DE retains the knowledge of good solutions in the current population. To achieving a desirable tradeoff between exploration and exploitation, a new adaptive differential evolution is introduced in this paper. The adaptive mechanism is employed to combine the coordinated crossover operator with the diversity-based population. Pertaining to the coordinated crossover, adaptive operator acts on binomial crossover and exponential crossover alternatively. With respect to the population, diversity-based adaptive strategy is applied. Experimental results show that the proposed algorithm adapts DE's adaptive behavior to different search stages, has made great improvement in convergence rate and solution quality.