【作者】 王俊年； 申群太； 沈洪远； 周鲜成；

【Author】 WANG Jun-nian SHEN Qun-tai SHEN Hong-yuan ZHOU Xian-cheng(1.Institute of Information Science and Engineering,Central South University,Changsha 410083,China;2.Institute of Information and Electrical Engineering,Hunan University of Science and Technology,Xiangtan 411201,China)

【机构】 中南大学信息科学与工程学院；

【摘要】 在微粒群算法中引入"基于密度"的聚类算法,构建出一种改进的小生境微粒群算法.该算法组合了两种方法来实现小生境的思想:第一,采用多种群策略,初始化产生一个没有子微粒群区分的主微粒群D₀后,在对D₀迭代执行l-bestPSO算法的同时,允许其中动态产生不相同的子微粒群D_i（i≥1）;第二,子微粒群的产生采用一种"基于密度"的聚类算法,如果两个个体之间的距离小于一个给定的极值σ_dist,则将这两个个体联系起来归入一个聚类簇,当聚类簇中的个体数目达到规定的子微粒群最小规模时形成一个小生境.用这种算法能够产生大小和形状不同的小生境,与生物学中地理小生境具有多种形状的事卖相符合,也克服了NichePSO算法只能以某一微粒为中心产生圆形小生境的不足.对3个常用的基本测试函数的测试实验表明,这种改进的小生境微粒群算法在多峰函数寻优中性能优于标准PSO和NichePSO.更多还原

【Abstract】 A‘density based’clustering algorithm is used in PSO to construct a modified niching PSO algo- rithm.In the new algorithm,two methods are combined to introduce the biological concept of niche into PSO. Firstly,a muhi-swarm strategy is adopted.After initializing the main swarm D₀,the particles of D₀ are trained using l-best PSO and difform sub-swarms are created simultaneously.Secondly,we use a‘density based’clustering algorithm to form sub-swarms.Two particles are connected to classify a cluster when their distance is lower than a given threshold value σ_dist,and if the cluster size exceeds a given minimum niche size, a new sub-swarm is created.The new niching PSO algorithm allows niches of varying size and shape.It not only accords with the fact that the geographical microenvironment have varying size and shape,but also get rid of the disadvantage of NichePSO that creating of niche depends on a circular topological field.Testing of the algorithm by using three benchmark functions indicate that the modified niching PSO has higher performance than standard PSO and NichePSO in locating multiple optimal solutions.更多还原