Abstract: In order to fully utilize the advantages of Particle Swarm Optimization(PSO) and Genetic Algorithm(GA) respectively, this paper proposes a new collaborative algorithm based on PSO and GA which is applied to clustering analysis. By constructing two mutual competitive populations, the algorithm produces the optimal individual in a bootstrapping process using relative fitness criteria instead of absolute fitness criteria. Experimental results on real world datasets show that the new algorithm is superior than Genetic Algorithm(GA) based clustering method and basic PSO clustering algorithm since it has higher convergence accuracy.

Key words: clustering algorithm, collaborative algorithm, Particle Swarm Optimization(PSO), Genetic Algorithm(GA), dual population

摘要： 为充分发挥粒子群优化算法和遗传算法各自的优势，提出一种新的基于粒子群和遗传算法的协同进化算法，并将其应用于聚类分析。通过构建2个相互竞争的种群，采用相对适应度度量方法，在一个纯自举的过程中产生最优竞争个体。在现实世界数据集上的仿真实验表明，该算法在收敛精度方面优于基于遗传算法的聚类方法和基本粒子群优化聚类算法。

关键词: 聚类算法, 协同算法, 粒子群优化, 遗传算法, 双种群

CLC Number: 

• TP301.6