改进的混合粒子群算法  

Improved hybrid particle swarm optimization algorithm

作　　者：李鹏[1] 全惠云[2] 

机构地区：[1]长沙职业技术学院信息技术系,湖南浏阳410300 [2]湖南师范大学数学与计算机科学学院,长沙410081

出　　处：《计算机工程与应用》2010年第11期29-31,45,共4页Computer Engineering and Applications

摘　　要：从研究分析粒子群算法和郭涛算法的特点出发,提出一种综合两算法优点的混合算法。新算法改变了粒子的更新方式,以子空间搜索和串行搜索相结合的多点并行搜索,扩大了算法的搜索范围,减少了粒子对初值的依赖,增强了算法跳出局部最优的能力;通过后代较优个体变异产生子群,提高了算法局部寻优能力;实验证明,该算法正确高效。This essay starts with the analysis of the characteristics of particle swarm optimization and Guotao algorithm,and raises a hybrid algorithm with their advantages.The new algorithm changes the particles＇updating ways,searches with the combination of subspace and serial searches,enlarges the algorithm searching scope,reduces the particle＇s dependency to the initial value,strengthens its ability out of the partial optimalization.It also produces subgroups by the better individuals＇variation of the descendants,improves its partial optimization ability.Experiments show that the new algorithm is high-efficiency.

关 键 词：子群 粒子群算法 郭涛算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]