Parallelism in production systems: the sources and the unexpected speed up
Aman Gupta · 1986
Production systems (or rule-based systems) are widely used in Artificial Intelligence for modeling intelligent behavior and building expert systems. On the surface production systems appear to be capable of using large amounts of parallelism—it is possible to perform match for each production in parallel. Initial measurements and simulations, however, show that the speed-up available from such use of parallelism is quite small. The limited speed-up available from the obvious sources has led us to explore other sources of parallelism. This paper represents an initial attempt to identify the various sources of parallelism in production system programs and to characterize them, that is, to determine the potential speed-up offered by each source and the overheads associated with it. The paper also addresses some implementation issues related to using the various sources of parallelism. This research was sponsored by the Defense Advanced Research Projects Agency (DOD), ARPA Order No. 3597, monitored by the Air Force Avionics Laboratory under Contract F33615-81-K-1539. ITie views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied of the Defense Advanced Research Projects Agency or the US Government