ABSTRACT
Recall from Chapter 7 that an expert system is composed of a problem solver, which contains the method for reaching a solution, and a knowledge base, which contains the specific knowledge needed to produce a solution. Generally, the problem solver and the knowledge base are kept logically separate; a distinction is made between the knowledge used by each. Such separation is essential for portability, the transfer of an expert system to a new problem area. An important question in expert system research is determining the scope of a system; that is, determining the range of problems to which the problem solver is applicable [11]. In this chapter we describe two experiments in which Ml and CGEN were applied to new problem areas, automated sequencing of tasks in a design environment and mechanical design.
