多目标云分布估计算法

为增强多目标分布估计算法(MEDA)的局部搜索能力,将云模型引入到多目标分布估计算法中,提出一种多目标云分布估计算法(CMEDA).该算法一方面利用分布估计的采样操作对进化种群进行搜索,另一方面利用云滴具有随机性、稳定倾向性等特点,进行外部档案搜索,实现群体间信息交换,从而提高多目标分布估计算法的全局搜索能力.数值实验选取6个常用测试函数,并与NSGA-Ⅱ和MEDA算法进行比较,结果表明,CMEDA算法在收敛性和多样性两方面都有较好的性能.

Cloud model based on multi-objective estimation of distribution algorithm

In order to enhance the local search capability of multi-objective estimation of distribution algorithm(MEDA),a cloud model was introduced into this algorithm.A cloud model based on multi-objective estimation of distribution(CMEDA)was proposed.In this algorithm,the evolution population was searched with sampling operation of estimation of distribution on the one hand,and on the other hand the outer population file was searched by using the feature of cloud drop such as its randomness and tendency to stabilize.Therefore,the information exchange between the populations was realized and the global searching ability with estimation of distribution algorithm was subsequently improved.In numerical experiment six common test functions were chosen and the experiment result was compared with that of both multi-objective algorithms NSGA-Ⅱ and MEDA.The result showed that both the convergency and diversity of the CMEDA exhibited more superiority.