一种新的改进粒子群优化方法  

New improved particle swarm optimizer

作　　者：刘怀亮[1] 高鹰[1] 许若宁[2] 苏瑞娟[1] 

机构地区：[1]广州大学计算机科学与教育软件学院,广州510006 [2]广州大学数学与信息科学学院,广州510006

出　　处：《计算机工程与应用》2010年第12期38-41,45,共5页Computer Engineering and Applications

基　　金：广东省自然科学基金No.8451009101001040~~

摘　　要：为解决粒子群优化算法易于陷入局部最优问题,提出了两种新方法并行修改粒子群优化算法惯性权重:对好于或等于整体适应度平均值的粒子,用动态非线性方程调整惯性权重,在保存相对有利环境的基础上逐步向全局最优处收敛;对比平均值差的粒子,用动态Logistic混沌映射公式调整惯性权重,在复杂多变的环境中逐步摆脱局部最优,动态寻找全局最优值。两种方法前后相辅相成、动态协调,使两个动态种群相互协作、协同进化。实验结果证实:该算法在不同情况下都超越了同类著名改进算法。To solve the local-optima convergence problem of particle swarm optimization,two new methods are introduced to modify the Particle Swarm Optimization inertia weight in parallel：When particles＇fitness values are better than or equal to the average,the introduced dynamic nonlinear equations are employed to modify the inertia weight,which can make particles retain favorable conditions and converge to the global optima continually;on the contrary,when fitness values are worse than the average,the dynamic Logistic chaotic map formula is introduced to modify the inertia weight,which can make particles break away from the local optima and search the global optima dynamically.Two methods coordinate dynamically,and make two dynamic sub-swarms cooperate to evolve.Experimental results demonstrate that the new introduced algorithm outperforms many other famous improved Particle Swarm Optimization algorithms on many well-known benchmark problems.

关 键 词：粒子群优化算法 惯性权重 动态非线性方程 动态Logistic混沌映射公式 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]