基于空间收缩的种群灭亡差异演化算法

Differential Evolution Algorithms with Extinction Based on Space Contraction

为避免演化算法在求解多峰函数优化问题时对冗余空间的过度搜索 ,提高差异演化算法的搜索效率 ,提出一种新的基于空间收缩的种群灭亡差异演化算法 (DEESC) ,通过最优个体收缩可行空间 ,用均匀设计方法反复初始化种群 ,并且讨论了DEESC的主要参数敏感问题。

This paper proposes a novel approach differential evolution algorithms with extinction based on space contraction (DEESC) to increase the efficiency and to avoid too much searching in wrong space. This method contract the feasible searching space by optimal individual using uniform design tablet to regenerate initial feasible point. And the main parameter is discussed. Cases studies illustrate that DEESC for function optimization proposed can improve the global convergence speed in lower computation efforts and has the advantages of robustness and efficiency to such a certain extent.