Control Knowledge in Expert Systems: Relaxing Restrictive Assumptions
Stephen Fickas, David Novick
Committee:
Technical Report(Dec 1969)
Keywords:

Many expert systems use fixed, domain-independent control strategies. These systems often embody assumptions about control and control knowledge which limit the utility of application programs. This paper shows that effective control often requires domain-specific knowledge. We present the ORBS model and implementation of a rule-based systems environment which attempts to expose, and give users access to, hidden control points. ORBS provides a language tor defining control strategies, and supports the interactive development of these strategies. We discuss the process or explicating control knowledge through a process or refining an application rule.