融合并行混沌萤火虫算法的K-调和均值聚类  

K-harmonic means clustering merged with parallel chaotic firefly algorithm

作　　者：朱书伟 周治平[1] 张道文[1] 

机构地区：[1]江南大学物联网工程学院,江苏无锡214122

出　　处：《智能系统学报》2015年第6期872-880,共9页CAAI Transactions on Intelligent Systems

基　　金：江苏省产学研联合创新资金-前瞻性联合研究基金资助项目(BY2013015-33)

摘　　要：针对K-调和均值算法易陷于局部最优的缺点,提出一种基于改进萤火虫算法(firefly algorithm,FA)的K-调和均值聚类算法。将基于FA的粗搜索与基于并行混沌优化FA的精细搜索相结合,其中精细搜索部分首先通过FA搜索到当前最优解及次优解,然后通过改进的logistic映射与并行混沌优化策略产生混沌序列在其附近直接搜索,以增强算法的寻优性能。最终,将这种改进的FA用于K-调和均值算法聚类中心的优化。实验结果表明:该算法不但对几种测试函数具有更高的搜索精度,而且对6种数据集的聚类结果均有一定的改善,有效地抑制了K-调和均值算法陷于局部最优的问题,提高了聚类准确性和稳定性。The K-harmonic means algorithm （KHM） has the disadvantage of easily falling into a local optimum. To solve this problem, we propose a hybrid KHM based on an improved firefly algorithm （FA）. In this paper, we com- bined raw FA-based searching with parallel chaotic FA-based elaborate searching. In the elaborate searching, we found the current best and second-best solutions using the FA, then we used an improved logistic map model com- bined with parallel chaotic optimization to search this area in order to enhance the searching ability of the algorithm. Finally, we used the improved FA to optimize the cluster centers obtained by the KHM. Experimental results dem- onstrate that the proposed algorithm not only had higher search precision for several test functions, but also improved the clustering accuracy and stability of six datasets, effectively avoiding being trapped into a local optimum.

关 键 词：K-调和均值 局部最优 萤火虫算法 聚类 并行混沌优化 混沌局部搜索 映射模型 种群多样性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]