基于蜜蜂繁殖机理的蜂群聚类算法  

A Clustering Algorithm Based on Marriage Mechanism in Honeybee Optimization

作　　者：曹永春[1] 纪金水[1] 邓涛[1] 蔡正琦[1] CAO Yong-chun;JI Jin-shui;DENG Tao;CAI Zheng-qi(College of Mathematics and Computing Science,Northwest Minzu University,Lanzhou,730030,China)

机构地区：[1]西北民族大学数学与计算机科学学院,甘肃兰州730030

出　　处：《西北民族大学学报（自然科学版）》2018年第3期1-6,共6页Journal of Northwest Minzu University(Natural Science)

基　　金：国家自然科学基金资助项目(31560256);中央高校科研项目(31920180114;31920170142)

摘　　要：将基于蜜蜂繁殖机理的蜂群算法应用于聚类问题,提出了一种新的蜂群聚类算法.对随机生成的初始蜂后在选优的基础上进一步优化,提高了算法收敛速度和聚类结果的稳定性.结合蜂后染色体的编码方式和雄蜂精子的单倍体特征,设计了产生幼蜂的交叉操作.充分利用蜂后良好基因信息对幼蜂展开邻域搜索来改进幼蜂质量.通过与其他聚类算法的对比实验,表明该算法具有良好的聚类效果和稳定性.Based on the marriage mechanism of honeybee optimization,the authors proposed a new clustering algorithm,which was applied to clustering problem.The algorithm chose the best one after a set of initial queens were generated randomly,and applied local search to get a good queen.This process improved convergence rate and the stability of the algorithm.Under considering the encoding method of the queen and haploid characteristics of drones,a crossover operation was designed to produce broods.Neighborhood search for broods could improve the quality of broods by using the good genetic information of queens.The results showed that the algorithm had good clustering effect and stability in comparison with the experiments in other clustering algorithms.

关 键 词：蜜蜂繁殖机理 聚类分析 邻域搜索 模拟退火 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]