透镜成像反学习策略在粒子群算法中的应用  

The Application of a Novel OBL Based on Lens Imaging Principle in PSO

作　　者：喻飞[1] 李元香[1] 魏波[1] 徐星[2] 赵志勇[1] 

机构地区：[1]武汉大学软件工程国家重点实验室、武汉大学计算机学院,湖北武汉430072 [2]景德镇陶瓷学院信息工程学院,江西景德镇333403

出　　处：《电子学报》2014年第2期230-235,共6页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61070009);国家科技支撑计划(No.2012BAH25F02);江西省自然(青年)科学基金(No.2012BAB211036)

摘　　要：在PSO中引入反向学习策略(Opposite-Based Learning)可使粒子在搜寻过程中总能找到当前解的反向位置,增加了接近全局最优解的机会.然而,OBL仅在演化初期作用显著,在演化后期则需通过变异等手段来提高其"开发"能力.针对该问题,基于透镜成像原理,引入缩放因子和搜索半径两个可调参数进一步平衡了算法的"探索"和"开发"能力.实验表明该策略能够提高种群多样性和收敛性能.By introducing opposition-based learning （OBL） in PSO ,particles are enabled to find an opposite position that is closer to the global optimization solution .However ,OBL only makes a good performance on the initial phrase of the evolution , while at later stages it needs to be combined with other techniques （e .g .Cauchy mutation ） to improve its ability of“exploration” . In this paper ,a novel OBL based on the principle of lens imaging is proposed .It uses two parameters （i .e .,zoom factor and the factor of search radius ） ,which will achieve a better balance between PSO’s“exploration”and“exploitation”abilities .The simula-tion shows that the novel OBL possesses better convergence rate and convergence effect .

关 键 词：反向学习 粒子群算法 透镜成像 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]