基于粒子群优化的模糊核聚类方法.pdfVIP

43 6 上 海 交 通 大 学 学 报 Vol. 43 No. 6 2009 6 JOU RNAL OF SH ANGHAI JIAOT ONG UNIVERSITY Jun . 2009 : 006-2467( 2009) 06-0935-05 杨广全, 朱昌明2 ( . , 0008 ; 2. , 200240) : 针对模糊核聚类对初始值敏感易陷入局部最 的缺点, 提出了基于粒子群 化的模糊核 聚类方法. 该方法根据聚类准则设计适应度函数, 利用粒子群 化算法对聚类中心进行 化, 在迭 代 化过程中设计了梯度下降法加快算法的收敛速度, 并引入变异机制增强粒子群的多样性. 仿 真实验及在水轮机转轮叶片裂纹源定位中的应用验证了算法的可行性和有效性. : 聚类分析; 模糊核聚类; 粒子群 化算法; 梯度下降法 : TP 39 . 4 : A Part icle Swarm Opt imizat ion Algorithm Based Fuzzy Kernel Clustering Met hod 2 YA N G Guang-quan , ZH U Chang-ming ( . T ransport Economics Research Institute, China Academy of Railw ay Sciences, Beijing 0008 , China; 2. School of Mechanical Engineering, Shanghai Jiaotong U niversity, Shanghai 200240, China) Abstract: T o overcome the draw backs of fuzzy kernel clustering method sensitive to the initial cluster cen- ters and easy to be trapped into local minima, the fuzzy kernel clustering method based on particle swarm optimization w as proposed. The proposed method designs the fitness function according to the clustering principle and utilizes the particle sw arm optimization to optimize the cluster centers. During the procedure of the iterations, the gradient descent operator is designed to accelerate the convergent speed and the muta- tion operator to improve the diversity of particle sw arm . T he simulation experiments and the crack source location in turbine blades verify the feasibility and effectiveness of the proposed method. Key words: cluster analysis; fuzzy kernel clustering; particle sw arm optimization; gradient descent ; º ; » , , . [ , 2] C-( K