Skip to Main Content
Computational characteristics of real-time expert systems have limited its integration in real-time control and monitoring environments. The computation time required to complete inferences carried out by expert systems present high variability, which usually leads to severe under-utilization of resources when the schedule of inferences is based on their worst computation times. Moreover, the event-based aperiodic activation of inferences increases the risk of transient overloads, as during critical conditions of the controlled or monitored environment the arrival rate of events increases. The dynamic scheduling technique presented in this article obtains statistical bounds of the time required to complete inferences on-line, and uses these bounds to schedule inferences achieving highly effective utilization of resources. In addition, this technique handles transient overloads using a robust approach. During overloads this technique completes nearly as many inferences as other dynamic scheduling techniques, but shows significantly better effective utilization of resources. The specific real-time architecture presented in this work, based on component object model (COM) technology, completes our approach to the problem of building efficient real-time rule-based systems suitable for controlling or monitoring complex industrial processes.