基于局部搜索能力强、收敛速度快的特点,首先初始化一个没有子种群的全局种群,再在全局种群中采用迭代搜索,并对其中的个体进行聚类,当聚类簇中的个体数目达到规定的最小规模时形成一个子种群,然后在各子种群中进行迭代搜索并重新进行聚类,从而提高进化过程中种群的多样性,增强算法跳出局部最优的能力.该算法基于weka，用于weka拓展功能，需要 weka算法包支持。-Based on the local search ability, the characteristics of fast convergence, first initialize a sub-population of the overall population, then the overall population in the iterative search, and clustering of the individuals, when the clustering of individual cluster achieve the required minimum number of the scale of the formation of a subset of the population, and then in the sub-populations in the iterative search and re-clustering to improve the evolutionary process of population diversity, enhancement algorithm' s ability to jump out of local optimum.