A REAL-TIME PROCEDURE FOR KNOWLEDGE PROCESSING

A real-time knowledge processing procedure is proposed for rule-based systems in general and control systems applications in particular. The procedure provides mechanisms that address important characteristics for real-time processing, including focus of attention, integration of symbolic/numeric processing, optimun use of environment and response time warranty. It also supports the trade-off needed in most real-time systems, for example between processing power, response time, data space and inattention. An application concerning supervisory group control of elevators is also included to show the usefulness of the proposed procedure.