| This research aims at developing an unique knowledge-based system methodology and associated techniques to tackle the problems occurring in the early design stages of mechanical systems.; Domain knowledge transformation (DKT) methodology unveils our vision of the early stages in engineering design process and the applicability of knowledge-based design systems. As the cornerstone of computerizing the abstract design knowledge, our method of the symbolic representation of design knowledge, namely, conceptual graphs (CG) paired with AI language PROLOG, exhibits its expressive power and flexibility. To elaborate the design specifications, we offer a new scheme that incorporates the techniques of domain knowledge guided structured interview (DKGSI), question and answer (Q&A) interface and dynamic knowledge base, etc. Then we demonstrate a new systematic approach of using case-based reasoning (CBR) to generate design solution concepts. A novel architecture of DKT knowledge-based design system, called DKT-PS Designer, is proposed in the light of the conceptual design of the positioning systems. The real life design cases run through all the discussions for illustrative purposes. |
