A LISP-based shell for model structuring in system design

Abstract

This thesis builds a knowledge-based, computer-aided decision making shell, written in LISP, for assistance in generic engineering system design problems. The theoretical framework presented in the thesis places system design processes in the environment of multifaceted modelling methodology and artificial intelligence techniques. A new reliable and efficient knowledge representation scheme--FRASES is introduced into the knowledge base design. The scheme combines system entity structure and frame and production rule system, and allows us to easily acquire, represent, and infer knowledge and information about the system being designed. In the design of the inference engine, multiple inference algorithms are supported in the shell. They infer a set of desired system configurations with respect to the designer's objectives and requirements. In comparison, top-down reasoning with depth-first offers the most efficient reasoning algorithm when using the FRASES knowledge representation scheme.