An approach for parallelizing OPS5 production systems and a faster match, C based indexing scheme for hypercube machines

Author

Shrivastava, Rajendra K. ; Iyengar, S.S. ; Carver, Doris

Author_Institution

Dept. of Comput. Sci., Southern Univ., Baton Rouge, LA, USA

fYear

1992

fDate

12-15 Apr 1992

Firstpage

489

Abstract

An algorithm for reducing the amount of time spent by a production system in its MATCH phase is presented. The algorithm utilizes C language´s arithmetic operations, which are not efficiently computed in standard production system languages, and also constricts the activity of the MATCH phase as performed in the OPS5 programming language. The parallelization is achieved by distributing the search amongst the nodes of a hypercube. The MATCH phase is an effort to satisfy the LHS (left hand side) of a PR (production rule) by the contents of a WM (working memory). C programming language was used for the implementation of the monkey-and-banana problem written in OPS5

Keywords

knowledge representation; parallel programming; pattern recognition; search problems; C based indexing; MATCH phase; OPS5 programming language; arithmetic operations; hypercube machines; monkey-and-banana problem; standard production system languages; Arithmetic; Computer languages; Computer science; Engines; Fires; Hypercubes; Indexing; Production systems; Robotic assembly; Robotics and automation;

fLanguage

English

Publisher

ieee

Conference_Titel

Southeastcon '92, Proceedings., IEEE

Conference_Location

Birmingham, AL

Print_ISBN

0-7803-0494-2

Type

conf

DOI

10.1109/SECON.1992.202399

Filename

202399