Computer Engineering and Applications ›› 2013, Vol. 49 ›› Issue (8): 198-202.
Previous Articles Next Articles
WANG Sha, GAO Maoting
Online:
Published:
王 莎,高茂庭
Abstract: Because particle swarm algorithm easily gets into local optimal solution, crossover and mutation of genetic algorithm will be introduced into particle swarm algorithm. According to different situation and the characters of crossover and mutation to use these two operators, this paper gives an improved particle swarm algorithm which can not only guard against into local optimal solution, but can jump out of local optimal solution. By using this improved algorithm into projection pursuit dynamic clustering model, a best projection pursuit dynamic clustering model can be found. Compared with original algorithm, experiments show that this method can make particle swarm algorithm avoid getting into local optimal solution.
Key words: mixed particle swarm algorithm, projection pursuit, dynamic clustering
摘要: 针对粒子群算法容易陷入局部最优解,将遗传算法的交叉和变异引入到粒子群算法中。根据不同的收敛情况及交叉和变异的特点使用两种算子,提出一种既能预防陷入局部最优解又能跳出局部最优解的混合粒子群算法,将该算法应用到投影寻踪动态聚类模型中来优化投影方向,得到近似最好的投影寻踪动态聚类模型。实验证明,相对于原始粒子群算法,该方法可以有效地避免陷入局部最优解,而且投影效果也更好。
关键词: 混合粒子群算法, 投影寻踪, 动态聚类
WANG Sha, GAO Maoting. New projection pursuit dynamic clustering model based on mixed PSO[J]. Computer Engineering and Applications, 2013, 49(8): 198-202.
王 莎,高茂庭. 一种基于混合PSO的投影寻踪动态聚类模型[J]. 计算机工程与应用, 2013, 49(8): 198-202.
0 / Recommend
Add to citation manager EndNote|Ris|BibTeX
URL: http://cea.ceaj.org/EN/
http://cea.ceaj.org/EN/Y2013/V49/I8/198