一种基于差分演化的K-medoids聚类算法

K-medoids clustering algorithm method based on differential evolution

针对传统的K-medoids聚类算法具有对初始聚类中心敏感、全局搜索能力差、易陷入局部最优、收敛速度缓慢等缺点,提出一种基于差分演化的K-medoids聚类算法。差分演化是一类基于种群的启发式全局搜索技术,有很强的鲁棒性。将差分演化的全局优化能力用于K-medoids聚类算法,有效地克服了K-medoids聚类算法的缺点,缩短了收敛时间,改善了聚类质量。通过仿真验证了此算法的稳定性和鲁棒性。

The traditional K-medoids clustering algorithm,because on the initial clustering center sensitive,the global search ability is poor,easily trapped into local optimal,slow convergent speed,and so on.Therefore,this paper proposed a kind of K-medoids clustering algorithm based on differential evolution.Differential evolution was a kind of heuristic global search technology population,had strong robustness.It combined with the global optimization ability of differential evolution using K-medoids clustering algorithm,effectively overcame K-medoids clustering algorithm,shortend convergence time,improved clustering quality.Finally,the simulation result shows that the algorithm is verified stability and robustness.