基于K-means的改进差分进化聚类算法

K-means聚类算法简单,收敛速度快,但是聚类算法的结果很容易受到初始聚类种群的影响,往往导致局部最优。差分进化算法具有很强的全局收敛能力和鲁棒性,但其收敛速度较慢。为此,将K-means聚类算法和差分进化算法相结合,提出一种基于K-means的改进差分进化聚类算法。该算法设置在一定范围内随迭代次数动态增加的交叉算子,以使算法在迭代过程中先进行全局搜索,再进行局部搜索,这样有助于平衡算法的全局寻优和局部搜索能力,并且加快了算法的收敛速度。最后,通过实验测试了算法的有效性。

Modified Differential Evolution Clustering Algorithm Based on K-means

K-means clustering algorithm is simple and converge rapid,but the result of clustering algorithm is vulnerable to the influence of initial cluster population,which often leads to local optimum.Differential evolution algorithm has strong global convergence ability and robustness,but it’s convergence volocity is slow.For this reason,K-means clustering algorithm is combined with the differential evolution algorithm,then a modified differential evolution clustering algorithm based on K-means is proposed.Within a certain range of the algorithm,the crossover operators that increased dynamically with iterative number are set,so that the algorithm carries out global search first and local search second in the iterative process,which can help to balance the global search capability and local search capability of algorithm,and accelerate the convergence speed of algorithm.Finally,the experiments have tested the effectiveness of algorithm.