Ontology-Based Design Rationale Retrieval

In today’s world, the success of the enterprise increasingly relies on its own knowledge assets and the effective use of knowledge assets. Design rationale is very important product design knowledge which plays a key role in improving the efficiency and quality of product design, therefore, the research about design rationale has been paid much attention of industry and academia. However, due to various reasons, the management and use of design rationale in enterprises is not ideal, and there remains many problems to be solved. In view of the problem that it is hard for designers to precisely and conveniently retrieve what they want from large amounts of design rationale, this thesis studies on ontology-based design rationale representation and retrieval, aims to significantly increase the effectiveness of design rationale retrieval and to support design rationale reuse.The main contents of this thesis are presented as follows:· An extended design rationale representation based on ontology is proposed, which supports the effectively retrieval and reuse of design knowledge. By analyzing the design knowledge which needs to be retrieved and reused, an extended design rationale representation based on IBIS, combining with requirements, FBS model, artifact hierarchy model and design process is presented. A design rationale ontology and some semantic rules are constructed accordingly with the purpose of effectively supporting semantic-based retrieval and reuse of design knowledge.·An approach for design rationale retrieval using ontology-aided indexing is presented. Firstly, translate design rationale records into instance ontologies by ontology population. Secondly, use semantic rules to achieve ontology reasoning and get inferred instance on-tologies. Finally, build an index based on the explicit information and the inferred infor-mation of. the instance ontologies. All the needed information for design rationale retrieval is contained in the index through ontology-aided indexing method. In addition, three kind- s of query modes including keyword, natural language and design rationale record, are developed to fulfill users’different retrieval requirements.·An SPARQL-based design rationale retrieval approach is presented, based on which, users can input keywords, natural language questions or design rationale records to search what they want. First, initial SPARQL query is generated from the user input query through SPARQL template based method or ontology-based automatic generation method. Then, a complete SPARQL query is generated by extending and optimizing the keywords. Based on this approach, users can use ontology-based retrieval with normal queries, and thus the retrieval recall and precision are improved.· A prototype system called DRManager is developed. Meanwhile, some tests are conducted to validate the main ideas of the proposed approaches.