Study on two-layer knowledge retrieval technology in conceptual design

Abstract

This paper studies on knowledge retrieval technology in product conceptual design. In this study, a two-layer retrieval technology is proposed. One is to get the similarities of function units through matching function attribute values which is weighted. The concepts of similarity vector are proposed for searching similar function units or module. Another, extended-rule and restriction matrix is proposed to depict the restrictions and constraints of problem-case and historic-cases. The similarity of restrictions between problem-case and historic-case is accounted by restriction matrix. A conceptual design knowledge retrieval framework is constructed to retrieve the design knowledge of the similar historic-case to problem-case. The framework comprises of requirement analysis, function decomposing, retrieving module of relation design knowledge and knowledge support module. Finally, an application instance for mobile telephone is shown to explain the process of knowledge support process based on case and integrated rule.