机构地区: 江南大学物联网工程学院
出 处: 《控制与决策》 2011年第10期1463-1468,共6页
摘 要: 为了改善二进制量子行为粒子群优化(BQPSO)算法的收敛性能,提出了一种基于完全学习策略的改进BQPSO优化(CLBQPSO)算法,并由此设计了一种新的数据聚类方法.该算法在4个测试数据集上与其他一些聚类算法进行了聚类实验比较,实验结果表明,基于CLBQPSO的聚类算法不仅收敛速度快,而且有较好的全局收敛性,收敛精度优于其他聚类算法,聚类效果更好. A binary quantum-behaved particle swarm optimization(BQPSO) algorithm based on comprehensive learning strategy is proposed for improving the performance of convergence.Then a new data clustering method is designed according to comphensive learing BQPSO(CLBQPSO).The new clustering algorithm is compared with some other clustering algorithms on four testing data sets in clustering experiment.The experiment results show that the CLBQPSO clustering algorithm not only converges faster but also owns the better global convergence.Contrast to other clustering algorithms,the better convergence accuracy and clustering solution are obtained.
关 键 词: 量子行为粒子群优化 二进制编码 完全学习策略 聚类
领 域: [自动化与计算机技术] [自动化与计算机技术]