Families of KBS Approaches

With the view that a KBS interpreter is a method for mapping from input data in the form of intermediate symbolic state descriptions to labels of interest, four families of approaches are described here, each offering inference mechanisms and related knowledge representations that can be used to solve interpretation problems namely, model-based approaches, digraphs, fault trees, and tables. These methods have been heavily used 

A large number of model-based systems use either qualitative or quantitative simulation, such as FAULTFINDER (Kelly and Lees, 1986) or EA-GOL (Roth, Woods, and Pople, 1992). These systems simulate normal behavior and compare the simulation results with observations, they simulate faults and compare simulation results with detected symptoms, or they interleave simulation with observation comparing the two to dynamically track normal and abnormal states. It is computationally very expensive to 

Widely used in rule-based systems, input-output matching approaches are used in conventional expert systems going back to the earliest systems. 

Overall, each of these KBS approaches has its place, depending on the needs of the problem and the characteristics of the process. However, the advantages, limitations, and roles of the various methods must be clarified so that problem-solving efficiency can be achieved along with the ability to reach correct conclusions despite challenging circumstances knowledge that is incomplete data that are incomplete, unreliable, or extremely plentiful or situations that are untested or completely unanticipated. 

Low Octane Low Olefin Content High CokeMake High H/C Ratio Low Conversion 

Faul tree. [Drawn from G. Stephanopoulos, Ed., Knowledge-Based Systems in Process Engineering, Vol. 1, Cache Case Studies Series, CATDEX, Cache Corporation, Austin, TX (1988). 

---