服务型制造中基于知识的产品服务配置  

Knowledge-based product service configuration in servitization

作　　者：沈瑾[1] 王丽亚[1] 隆惠君[2] 吴明兴[1] 江志斌[1] 

机构地区：[1]上海交通大学机械与动力工程学院工业工程与物流工程系,上海200240 [2]上海交通大学中美物流研究院,上海200030

出　　处：《计算机集成制造系统》2013年第10期2633-2643,共11页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金重点资助项目(70932004/G0209)~~

摘　　要：为了完成服务型制造中知识密集的产品服务配置活动,基于本体和神经网络规则抽取技术,从配置知识的表达、获取和推理等关键功能出发,提出一套完整的产品服务配置方法,包括基于网络本体语言和语义网规则语言完成对结构知识和规则知识的表达;通过改进的骨架法实现基于专家经验的结构知识获取,通过局部聚集神经网络和规则抽取算法的联合实施实现基于历史数据的规则知识提炼;基于专家系统完成对配置方案的推理求解。以楼宇控制产品服务为例,验证了该方法的有效性。To realize knowledge-intensive product service eoniiguration in servltlzatlon, based on the extraction tech- nique of ontology and neural networks, a complete product service configuration method was proposed from the view of knowledge representation, acquisition and reasoning. In this method, structural knowledge and rule knowledge were represented by Web ontology language and semantic Web rule language respectively. An improved skeleton method was proposed to acquire structural knowledge from expert experience, and an approach combining Local Cluster （LC） neural network and rule extraction algorithm was presented to extract rule knowledge from historical data. Configuration reasoning processes were performed with the support of Java Expert System Shell （JESS）. The effectiveness of proposed method was validated by services of building control product.

关 键 词：产品服务 配置 本体 神经网络 

分 类 号：C939[经济管理—管理学] TP3-05[自动化与计算机技术—计算机科学与技术]