Adaptive differential evolution for high-dimension multimodal optimization problems
Yuanjing Feng · Control theory & applications · 2008
Based on the theoretcal analysis of selective variance in mutation operator of original differential evolution (DE) algorithm, we proposed an adaptive differential evolution (ADE) algorithm to tackle the high-dimension multimodal optimization problems. In order to make a good tradeoff between the exploration and exploitation, ADE algorithm adopts an adaptive weighted centroid mutation strategy. Furthermore, modifications in mutation and crossover rule are suggested to the original DE algorithm to intensify the search around the global minima. These modifications intend to exploit the information derived from the previous function evaluations to improve the efficiency of the algorithm in the local search, without deteriorating the behavior of the original DE algorithm in the global search. Numerical experiments indicate that the resulting algorithm is considerably better and more efficient than the DE, DERL and DERB algorithms. Finally, a numerical study is carried out using a set of 30 test problems, many of which are inspired by practical applications.