Differential Evolution with strategy of improved population diversity
Zhao Li, Chao-jiao Sun, Xian-chi Huang, Bing-xu Zhou · 2016
Differential Evolution (DE) algorithm is well known as a simple and efficient scheme for global optimization over continuous spaces. In order to ameliorate the population diversity, an improved differential evolution (IDE) algorithm is proposed in this paper. The idea is to vary the assembling positions of the premature individuals by mutation operation. It is proved in theory that the direction pointed to the center individual is the reasonable one to improve the diversity. Then the adaptive disturbance scheme after population premature is designed. The example based on the standard function shows that the IDE with good diversity has much better searching capacity and calculating precision in the whole evolution than that of DE.