A quantitative model-based expert system for LAN development
M.E. Ulug · 2003
A quantitative model-based expert system, or QMB for short, is described. A QMB is a rule-based system coupled to an analytic model. The QMBs provide another level of refinement over rule-based systems by giving quantitative answers to problems. He discusses the difficulties associated with the design of a shell for a QMB and proposes the use of a semishell, which assists the knowledge engineers in writing the logic, uncertainty and inference modules of the QMB expert system program. The design and coding of the modules that deal with the numerical data is left to the knowledge engineer. The operation of a QMB having 43 rules is described. A software architecture is presented that can dynamically change its operation. The purpose of the architecture is to produce an expert system that can think more like human beings.>