In order to make the differential evolution algorithm fully mine and use the historical data information in evolu⁃tion process,and improve the global searching ability and convergence rate,a differential evolution algorithm based on principal component PCADE is proposed in this paper. The population space is mapped to the principal component space in this algorithm to obtain a population space composed of principal component. In evolution process,the first m individuals composing of the principal component can access to the next generation evolution directly,and the individual with high fitness value in the residual N⁃m individuals is selected from the original population and principal component population space to access to the next genera⁃tion. The experimental results show that the improved algorithm can obtain better result in cluster analysis.%为了使微分进化算法在进化过程中充分挖掘和利用历史数据信息，提高它的全局搜索能力和收敛速度，提出了一种基于主成分的微分进化算法PCADE。该算法将种群空间映射到主成分空间从而得到一个由主成分构成的种群空间，在进化过程中前m个主成分构成的个体可以直接进入下一代的进化，而剩余的N-m个个体则从原种群和主成分种群空间中选择出适应度值较高的个体进入下一代。实验结果表明改进算法在聚类分析中取得了较好的结果。
展开▼
