HyKET(Hybrid Knowledge Engineering Tool) is an expert system development tool under development at KAIST. It provides a unique developmental environment supporting multi-paradigm knowledge representation schemes such as Common LISP programming, rule-based system, object-oriented programming, window system, and graphichs user interface with Hangul support.
The Flavor System, a lisp-based object-orinted language which has been used intensively in AI Community, is chosen as the base environment of HyKET. The Flavor System has a powerful inheritance and method combination mechanisms to support parsimonious knowledge representation. as a thesis research, we implemented the Flavor System on a Common LISP environment running on IBM PC/AT. In doing this, the features of the Flavor System is extended to include selective inheritance of variables and methods.
We also designed and implemented a forward-chaining rule-based system with explanation capability, called KAPS, on the top of the Flavor system. Thus object-oriented representation as well as production rule representation may be used to construct a knowledge base. KAPS can express the same knowledge more naturally with a smaller number of rules than other plain production systems can. As a consequence, KAPS runs much fast in solving a particular problem.
