Real-Time Knowledge-Based Systems

Abstract

Knowledge-based systems can model a system, assess its performance, detect abnormalities in its operation, and make suggestions for its safe, stable, and optimal operation by capturing quantitative and heuristic information about a system and using rule-based reasoning to make inferences. A knowledge-based system can interface algorithms and quantitative information with heuristics and rules to yield a powerful environment that can accommodate nonlinearities, uncertainties in information, and rapid shifts in process operation. This chapter introduces knowledge-based and agent-based systems. The latter relies on distributed artificial intelligence and is well-suited for systems with distributed or discrete elements. Three applications are presented to illustrate the capabilities of knowledge-based and agent-based systems: supervision of penicillin fermentation systems, development of a distributed process supervision and fault diagnosis system, and modeling of a mammalian cell bioreactor.