一种基于改进差分进化的K-均值聚类算法研究

Research on a K-Means Clustering Algorithm Based on Improved Differential Evolution

针对K-均值算法的差异与缺点,对初始值敏感,易于落入局部最优解,差异进化算法具有强大的全局收敛能力和鲁棒性,但其收敛速度较慢。鉴于上述问题和缺陷,论文首先详细介绍了进化算法关键操作和差分进化算法的步骤和具体流程。然后,阐述了基于差分进化的K-均值聚类算法的描述,步骤和具体流程。最后,提出基于改进差分进化的K均值聚类算法,详细介绍改进方案,改进算法的步骤和具体流程。基于差分进化和改进算法的K均值聚类算法进行仿真实验,实验结果表明,该算法具有较好的搜索能力,算法收敛速度更快,鲁棒性更强。

The difference and the shortcomings of K-means algorithm are sensitive to the initial value and easily fall into the local optimal solution. The difference evolution algorithm has strong global convergence ability and robustness,but its convergence rate is slow. In view of the above problems and defects,this paper first introduces the steps and the concrete process of evolutionary algorithm key operation and differential evolution algorithm. Then,the description,steps and concrete flow of K-means clustering algorithm based on differential evolution are described. Finally,a K-means clustering algorithm based on improved differential evo. lution is proposed,and the improvement scheme,the steps and the concrete flow of the algorithm are introduced in detail. The K-means clustering algorithm based on differential evolution and improved algorithm is used to simulate the experiment. The experi. mental results show that the algorithm has better searching ability,and the algorithm is faster and more robust.