基于领域本体和CBR的案例知识检索方法  

Case retrieve methodology based on domain ontology and case-based reasoning

作　　者：文家富[1] 郭伟[1,2] 邵宏宇[1,2] 

机构地区：[1]天津大学机械工程学院,天津300072 [2]天津大学装备设计与制造技术天津市重点实验室,天津300072

出　　处：《计算机集成制造系统》2017年第7期1377-1385,共9页Computer Integrated Manufacturing Systems

基　　金：国家863计划资助项目(2015BAF32B03)~~

摘　　要：为解决汽车覆盖件模具设计经验依赖性强、隐形设计知识重用效率低、案例知识检索条件要求高、精确度低等问题,针对结构设计知识多源、异构、动态的特点及现有CBR方法的不足,建立了覆盖件模具设计知识的领域本体模型,构建了各本体概念之间的逻辑关系,并提出一种新的案例设计知识检索重用方法。该方法引入基于语义距离、节点深度和节点密度的数学模型,利用语义相似度计算和CBR原理结合(S-CBR)的方法,实现了设计案例知识的双层检索和快速重用,提高了检索效率和准确率,并开发了模具设计支撑系统。最后,以模具结构设计为例验证了所提方法的可行性和有效性。To solve the problems occurred in the automobile panel die design process included strong dependency of design experience,low tacit design knowledge reuse efficiency and low retrieve accuracy,according to the multisource,heterogeneous and dynamic characteristics of die design case knowledge,the domain ontology model was created,and the logic relationship between ontology concepts was established.The hybrid method of semantic similarity and CBR mechanism named Semantic-CRB（S-CBR）included the node depth and node density was applied to fulfill the target of knowledge reuse and reduce the lead time based on double layer retrieve mechanism.The prototype system was built and an application was given to prove the validation of the proposed methodology.

关 键 词：模具设计 领域本体 语义相似度 基于案例的推理 汽车覆盖件 产品设计 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程] TH122[自动化与计算机技术—控制科学与工程]