一种基于混合差分策略的改进差分进化算法  

A Modified Differential Evolution Algorithm Based on Hybrid Differential Strategies

作　　者：蔡亮[1] 杨启文[1] 岳兴汉[1] 

机构地区：[1]河海大学计算机及信息工学院,江苏常州213022

出　　处：《河海大学常州分校学报》2007年第4期9-12,共4页Journal of Hohai University Changzhou

摘　　要：对启发式优化算法中的差分进化算法进行改进,在进化过程中并行交叉采用DE/rand/1/exp和DE/best/1/exp差分策略,应用聚集度因子进行种群重构,缩小了种群重构后的搜索范围,有效避免了种群重构的随机性.仿真结果表明,改进算法与使用单一差分策略的差分进化算法及PSO算法相比,寻优能力得到了显著提高.To improve the differential algorithm, a new kind of heuristic optimization method and a modified differential evolution （MDE） scheme are presented, where DE / rand / 1 / exp and DE / best / 1 / exp run currently in evolution process. With the usage of current aggregation degree, the search range of re- initialization of the MDE is reduced and the randomicity of re- initialization is avoided efficiently. The experimental results show that the search ability of MDE is improved remarkably compared to DE with single differential strategy and PSO.

关 键 词：差分进化 差分策略 种群重构 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]