Parallel execution of real-time rule-based systems

Abstract

When rule-based expert systems are used to monitor and control real-time systems, the ability of these expert systems to met stringent response time constraints is as important as their ability to produce correct results to react to input. This paper explores parallel execution as an approach to achieve higher execution speed in rule-based systems in domains requiring high performance and real-time response. In particular, it shows how rule-firing parallelism can be automatically extracted from a real-time rule-based system via static analysis of the system source code. To demonstrate the practicality of this approach, the proposed technique is applied to reduce the execution time of two NASA expert systems.<>