基于序列模式挖掘的变型设计知识推送  

Product design knowledge recommendation based on sequential pattern mining

作　　者：徐荣振 高琦[1,2] 王昊[1,2] 徐廷[1,2] 

机构地区：[1]山东大学机械工程学院,山东济南250061 [2]山东大学高效洁净机械制造教育部重点实验室,山东济南250061

出　　处：《计算机集成制造系统》2016年第5期1179-1186,共8页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(51375277);山东省自然科学基金资助项目(ZR2012GM015)~~

摘　　要：为提高产品变型设计中知识推送的自动化程度、减少人力投入,提出一种新的知识推送方法。利用频繁序列模式挖掘技术分析知识的历史使用数据,识别设计任务所对应的频繁知识序列;在产品设计过程中,结合设计师的知识使用行为、频繁知识序列的支持度和当前任务知识序列与频繁知识序列的相似度实现知识的推送;根据所提出的知识推送方法开发了基于产品数据管理系统的知识推送构件,并在某防爆电机生产企业中得到应用。应用后的统计数据表明,该构件提高了产品设计效率。To increase the automatieity of recommendation and reduce the labor costs in product variant design, a new method of knowledge recommendation was presented. The frequent knowledge sequence technology was used to ana- lyze historical knowledge usage data of designers with sequential pattern mining technique, and the corresponding frequent knowledge sequence of design task was recognized. In the process of product design, the knowledge recom- mendation was realized by combining the designer＇s knowledge usage behavior with the support degree of frequent knowledge sequences and the similarity between current knowledge sequence and frequent knowledge sequences. Ac- cording to the proposed knowledge recommending method, a component was developed based on Product Data Man- agement （PDM）, which was applied in a company which produces explosion-proof electric machine. The statistics showed that the component was helpful to increase the design efficiency.

关 键 词：变型设计 知识推送 序列模式挖掘 知识使用行为 

分 类 号：TH122[机械工程—机械设计及理论]