基于粒计算和序列比对的典型工艺路线发现方法  

Typical process route discovery method based on granular computing and sequence alignment

作　　者：周丹晨[1] 

机构地区：[1]中国工程物理研究院机械制造工艺研究所,四川绵阳621900

出　　处：《计算机集成制造系统》2015年第3期694-701,共8页Computer Integrated Manufacturing Systems

基　　金：中国工程物理研究院科学技术发展基金资助项目(2013B0203031)~~

摘　　要：针对企业工艺设计数据的继承性和重用性问题,提出一种基于粒计算和序列比对的典型工艺路线发现方法。首先采用生物信息学领域的序列比对技术,从整个工序序列的范围比较两条工艺路线,建立其最佳比对,并综合考虑正确匹配的工序数、连续正确匹配工序数和连续空位数来计算其相似度,构建工艺路线的模糊相似矩阵;然后依据粒计算理论模型之一的模糊商空间理论,创建具有分层递阶结构的工艺路线商空间族,通过聚类有效性评价,确定最优工艺信息粒层群;最终从构成不同粒层的工艺信息粒中,根据工艺路线两两最佳比对结果,选取完全一致的工序集作为典型工艺路线。通过应用实例验证了方法的可行性和有效性。Aiming at the problems of hereditability and reusability for process planning data in enterprise, a typical process route discovery method based on granular computing and sequence alignment was put forward. By applying sequence alignment technology in the field of bioinformatics, the optimal alignment between two process routes was established through their comparison from the range of the whole operation sequence, and their similarity was calcu- lated by comprehensive consideration of the number of correct matching operations, the number of continuous cor- rect matching operations and the number of continuous gaps. Therefore, a fuzzy similarity matrix of all process routes was constructed. According to fuzzy quotient space theory which was one of the theoretical models of granu- lar computing, the process route quotient space family with hierarchical structure was built, and the optimal process information granular layer group was determined by means of the evaluation of clustering validity. From the process information granules in different granular layers, the exactly same operation sets were selected as typical process routes according to the pairwise optimal alignment results of process routes. The feasibility and validity of proposed method was verified by an application example.

关 键 词：典型工艺路线 序列比对 粒计算 模糊商空间 工艺信息粒 粒层 相似度 

分 类 号：TH162[机械工程—机械制造及自动化] TP18[自动化与计算机技术—控制理论与控制工程]