织物变形舒适性的双群粒子群聚类研究.pdfVIP

第 31 卷第4 期 纺织学报 VoJ. 3 1, No.4 2010 年 4 月 Journal of Textile R传.earch Apr. , 20 10 文章编号:0253-972 1 (2010)伽1 - ∞60-05 织物变形舒适性的双群粒子群聚类研究 王永林，王东 云 ( 中~工学院 电子信息学院，河南郑州 45ω07) 搞 要 钊对传统蒙类方法存在的一些问题，提出了一种具有个体交换策略的双群幸辈子群聚类方法，裤 1 根据E 个基于滔应度的非线性公式动态更新惯性权重.群2 采用固定仪重.每进化一代，2 .!1羊交换部分个体。 将该方法应 用于织物变形舒适性~类中，以褒类中心作为被子位置，通过校子群优化算法获得是优聚类中心，采用最小距离准 则对样本进行Jt类e 最后与模糊缓类做 f 简单比较，结果表明该方法结论合理，便于应用，为逃得服装面料和评价 织物性能提供了一种新手段。 关键词 变形舒适性;双指控子在宇算法;模糊聚类:织物评价 中图分类号:T5 101; TP 18 1 文献标志码:A Clustering study of fabric deformation comfort using bi-swarm PSO algorithm WANG Yonglin , WANG Dongyu n ( Dtpartm ent of EleclTonic and lnformalion , Zhongyul1t UnilJlJT$ùy of Techllology , Zhen古zhou ， Henon 450ω7 ， China) Abstract Aiming al solving the problems of some traditional clusler melhods , an improved clusler melhod based on bi-sw缸m particle swarm opt imizalion (PSQ) algorithm wilh exchanging particles strategy wa民 proposed_ Qne swarm dynamically updated its inertia weight by a new nonlinear updating fonnu la , and the olher employed a conslant inertia weigh t. The two swarms exc hanged some particles after each iteralion . T he method was applied to Ih e c1usler of fabric deformation comfort a nd took the c1 us ler center as the po制tion of lhe particle. The oplimal clu ster center was obta ined by PSO algorithm opt imizing and Ihe sample data were clustered using lhe rn inimum dislance criterion . The comparison resuhs between fuzzy cluster results and PSO-based cluster shows Ihal Ihe .proposed method can get Ihe proper clu st