一种基于信息传递的分布估计算法  

An Estimation of Distribution Algorithm Based on Information Transmission

作　　者：何小娟[1,2] 曾建潮[2] 王丽芳[2] 

机构地区：[1]兰州理工大学电信工程学院,甘肃兰州730050 [2]太原科技大学复杂系统与计算智能实验室,山西太原030024

出　　处：《电子学报》2011年第4期967-970,共4页Acta Electronica Sinica

基　　金：山西省青年科技基金(No.2010021017-2)

摘　　要：借鉴信息传递的概率模型,提出一种求解非数值优化问题的新的分布估计算法.首先根据进化过程中的优良信息建立一个不断更新的先验知识概率模型,以相邻符号出现的频率为基础建立条件传递概率模型,然后通过二者的结合建立了一种后验概率模型并用以指导产生新群体.针对旅行商问题进行的仿真试验表明本文算法可较好地改善分布估计算法的早熟收敛现象.Reference to the probability model of information transmission,a new estimation of distribution algorithm is proposed for non numerical optimization problems.Firstly,an updating model of a priori knowledge probability is built according to the superior information produced during evolution process,and the model of conditional transfer probability is also constructed based on the emerging frequencies of neighboring symbols.Secondly,the model of posterior probability is given by combining the above mentioned probability model to guide new population generating.Finally the presented approach is tested on TSP problems,and the results show that the proposed algorithm can improve the premature convergence of estimation of distribution algorithms.

关 键 词：分布估计算法 信息传递 后验概率 旅行商问题 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]