EUREKA-II: a programming tool for knowledge-based real time control systems

Abstract

A knowledge engineering (KE) tool called EUREKA-II that is intended to support the building of knowledge-based real-time control systems is outlined. Its requirements are: the knowledge representation scheme and the inference method must correspond to those of the operators: easy connection to conventional control systems: efficient knowledge processing capability. To meet these requirements, EUREKA-II provides representation of five types of knowledge, i.e. metaknowledge, rule knowledge, factual knowledge, procedural knowledge, and hypothetical knowledge. These types of knowledge are processes by a high-performance inference mechanism consisting of a supervisor, a rule-network processor, a frame-base manager, and a view-note manager. This architecture allows high-level knowledge processing and flexible integration with existing application programs. EUREKA-II has successfully been applied to blast furnace control and is being applied to several other fields.<>