Capturing Designer Expertise: the CGEN System

Knowledge-based systems are becoming pervasive in the computer-aided design area. For these systems to achieve satisfactory levels of performance large amounts of knowledge are necessary. However, the acquisition of knowledge is a difficult and tedious task. Automated knowledge-acquisition tools (AKAT) provide capabilities for quickly building and maintaining knowledge-bases. This paper describes the CGEN AKAT, which allows hardware designers, unfamiliar with artificial intelligence programming techniques, to deposit their expertise into a synthesis tool's knowledge-base. A set of experiments which tested CGEN's capabilities are presented. The experiments show that with CGEN hardware designers can produce high quality knowledge-base