基于单点变异算法的单元分组问题的研究  

Research on the cell partition problem based on single point mutation algorithm

作　　者：韩毅[1,2] 王德志[1] 林华珍[1] 顾冰[1] 

机构地区：[1]浙江工业大学经贸管理学院,浙江杭州310023 [2]浙江工业大学浙江省技术创新与企业国际化研究中心,浙江杭州310023

出　　处：《浙江工业大学学报》2016年第2期202-206,共5页Journal of Zhejiang University of Technology

基　　金：国家自然科学基金资助项目(71301147;71301148;71302051);教育部人文社科基金资助项目(12YZCZH065)

摘　　要：针对单元制造中单元构建问题所涉及到的单元分组数问题,结合零件—设备关联矩阵的特点,提出一种划分单元数的新颖算法.以直接聚类算法(DCA)的计算结果为基础,根据4种不同的扩张路径形成决策序列.对决策序列进行单点变异,利用单元成组效率评价指标对决策序列进行评价.算例结果表明:对于复杂的初始关联矩阵,所提算法可以提高单元划分的成组效率,得到较满意的结果.Aiming at cell partition problem in the cellular manufacturing system,a novel algorithm,combining with the characteristics of the parts and equipment association matrix for partitioning cells is presented.During the process of cell formation,the result from the direct clustering algorithm（DCA）is adopted as an initial solution.A decision sequence is then reached according to four different expansion directions.Then,the single point mutation on the decision sequence is carried out,and the efficiency evaluation index of cell formation is used as the objective function to evaluate the algorithm.The results show that the proposed algorithm can improve the efficiency of cell partition and get better satisfactory results for those problems with complicated initial correlation matrix.

关 键 词：单元制造系统 直接聚类算法 单元分组数 单元构建 

分 类 号：TH163[机械工程—机械制造及自动化]