Implementation of multiple rule firing production systems on hypercube

Authors:
Steve Kuo;Dan Moldovan
Affiliations:
Department of Electrical Engineering - Systems, University of Southern California, Los Angeles, California;Department of Electrical Engineering - Systems, University of Southern California, Los Angeles, California
Venue:
AAAI'91 Proceedings of the ninth National conference on Artificial intelligence - Volume 1
Year:
1991

Citing 3
Cited 0

Parallel program design: a foundation

Parallel program design: a foundation
Transforming rule-based programs: from the sequential to the parallel

IEA/AIE '90 Proceedings of the 3rd international conference on Industrial and engineering applications of artificial intelligence and expert systems - Volume 2
A parallel asynchronous distributed production system

AAAI'90 Proceedings of the eighth National conference on Artificial intelligence - Volume 1

Quantified Score

Hi-index	0.00

Visualization

Abstract

The performance of production programs can be improved by firing multiple rules in a production cycle. In this paper, we present the multiple-contexts-multiple-rules (MCMR) model which speeds up production program execution by firing multiple rule concurrently and guarantees the correctness of the solution. The MCMR model is implemented using the RUBIC parallel inference model on the Intel iPSC/2 hypercube. The Intel iPSC/2 hypercube is chosen because it is a cost-effective solution to large-scale application. To avoid unnecessary synchronization and improve performance, rules are executed asynchronously and messages are used to update the database. Preliminary implementation results for the RUBIC parallel inference environment on the Intel iPSC/2 hypercube are reported.