基于分层多子群的混沌粒子群优化算法  

Chaotic particle swarm optimization algorithm based on hierarchical multi-subpopulation

作　　者：王维博[1,2] 冯全源[2] 

机构地区：[1]西华大学电气信息学院,成都610039 [2]西南交通大学信息科学与技术学院,成都610031

出　　处：《控制与决策》2010年第11期1663-1668,共6页Control and Decision

基　　金：国家自然科学基金-中国工程物理研究院联合基金项目(10876029);国家863计划项目(2009AA01Z230)

摘　　要：在分层多子群结构模型的基础上,提出一种混沌粒子群优化算法(HCPSO).该算法对非线性递减的惯性权重进行混沌变异,并采用了混沌搜索方法.在更新全局历史最优位置每一维分量时,选取不同的若干个体作为学习对象,并计算它们的平均位置.混沌搜索区域半径可根据粒子个体最优位置与上述平均位置间的距离自适应地调整.通过对几种典型函数的测试结果表明,该算法具有较好的全局搜索和局部搜索能力,可有效避免早熟收敛问题.A chaotic particle swarm optimization algorithm（HCPSO） is proposed based on the model of hierarchical multi-subpopulation structure,which uses chaotic mutation in nonlinear and decreasing inertia weight.For reinforcing the local searching ability,every particle has its own new global best position by using chaotic searching strategy.The new global best position is the average position of several individuals which are picked out as exemplars when the new global best position is updated in each dimension.The radius of the chaotic searching region is adaptively adjusted according to the distance between the particle＇s personal best position and the average position.The simulation results show that HCPSO is more effective to keep a balance of global and local searching ability,and to overcome the slow convergence and prematurity.

关 键 词：粒子群优化算法 混沌搜索 惯性权重 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]