基于粗糙集的模块聚类方法及应用  

Modular Clustering Method Based on Rough Set and Application

作　　者：李玉鹏[1] 褚学宁[1] 耿秀丽[1] 程辉[1] 程晓猷 全嘉青 

机构地区：[1]上海交通大学机械与动力工程学院,上海200240 [2]上海电气股份有限公司人民电器厂技术中心,上海200042

出　　处：《机械设计与研究》2011年第4期1-5,12,共6页Machine Design And Research

基　　金：国家自然科学基金资助项目(51075261);上海市科技创新行动计划资助项目(09dz1124600;10dz1121600);国家863计划资助项目(2007AA04Z140);教育部高等学校博士点基金资助项目(20070248020)

摘　　要：产品设计及客户化定制程度是影产品全生命周期的瓶颈;针对传统模块识别方法建模困难,计算复杂和无法有效处理评价数据的模糊性和不确定性等问题,提出了基于粗糙集的产品模块聚类方法。由于无需提供任何主观先验信息,粗糙集理论处理数据更为客观;上下限(集合)形式的粗糙数可使模块聚类结果更为精确。将评价数据转化成粗糙数形式进行模糊聚类分析,在聚类图中根据选取的上下限阈值实现模块划分。以某企业低压电器产品RMM2-250塑料外壳断路器为例验证了方法的有效性。The product design and customization level were indicated as the bottleneck of product life cycle.To deal with the difficulty of modeling,complexity of computing,uncertainty and fuzziness of evaluation data in the conventional module recognition method,a module clustering recognition method based on rough set was proposed.The computing process of evaluation data based on roughset theory is more objective because it doesn＇t need any subjective priori information;the data form of upper and lower limit（set） can make clustering result more precision.Express evaluation data as rough numbers for clustering analysis,the module recognition was achieved based on the upper and lower limit threshold in the clustering diagram.A case study of RMM2-250 molded case circuit breaker（MCCB） of a certain low-voltage apparatus producer was presented to illustrate the effectiveness of the proposed method.

关 键 词：产品设计 模块识别 粗糙集 模糊聚类 

分 类 号：TH122[机械工程—机械设计及理论] TP391.7[自动化与计算机技术—计算机应用技术]