An Expert System Shell for Aerospace Applications
IEEE Expert: Intelligent Systems and Their Applications
Comparison of three algorithms for ensuring serializable executions in parallel production systems
AAAI'92 Proceedings of the tenth national conference on Artificial intelligence
| Hi-index | 0.00 |
This report describes Version 2.0 of the experimental Parallel OPS5 developed at University of Massachusetts. This is a lisp-based rule-based system which supports parallelism at three levels, rule-, action-, and matched-level, on a shared-memory multiprocessor. The parallel OPS5 has been augmented with a queue-based scheduler which supports asynchronous rule firing, non-conflict-set-based heuristic control mechanisms, and a locking mechanism for preventing certain classes of rule conflicts during parallel rule firing. The User''s manual describes the language features and syntactic modifications which are specific to the UMASS Parallel OPS5 (UMPOPS). This is followed by a brief tutorial on programming OPS5 programs in parallel accompanied by a discussion of two benchmark programs and an analysis of the parallelism achieved in these bench-marks. The final chapter of this report is devoted to a discussion of implementation issues