Improved Differential Evolution Algorithm Based On Elite Group
Xiaobo Gao, YouCai Wang, GuangZhao Yang · 2016
By introduce the information entropy and the average-distance-amongst-points to analysis the population distribution in the process of evolution, and figured out the cause of the DE/best/* premature convergence is the control function of the current optimal individual to decrease the population diversity of the algorithm.Based on the number of base vectors, improved the DE algorithm by setting up the elite group, the elite differential evolution algorithm is proposed.Finally, several typical test functions are used to test the performance.The results show that the elite differential evolution algorithm has a good performance in the search success rate and the global search capability.