Data-Driven Parallel Production Systems

Jean Luc Gaudiot; Andrew Sohn

doi:10.1109/32.48936

Data-Driven Parallel Production Systems

Jean Luc Gaudiot, Andrew Sohn

Research output: Contribution to journal › Article › peer-review

7 Scopus citations

Abstract

Much effort has been expended on developing special architectures dedicated to the efficient execution of production systems. While data-flow principles of execution offer the promise of high programmability for numerical computations, we demonstrate here that the data driven principles can also be applied to symbolic computations. In particular, we consider a mapping of the RETE match algorithm along the line of production systems. Bottlenecks of the RETE match algorithm in multiprocessor environment are identified, based on which the algorithm is parallelized. The modifications to the actor set as well as the program graph design are shown for execution on the Tagged-Token Data-flow Computer. The results of a deterministic simulation of this multiprocessor architecture demonstrate that artificial intelligence production systems can be efficiently mapped on data-driven architectures.

Original language	English (US)
Pages (from-to)	281-293
Number of pages	13
Journal	IEEE Transactions on Software Engineering
Volume	16
Issue number	3
DOIs	https://doi.org/10.1109/32.48936
State	Published - Mar 1990
Externally published	Yes

All Science Journal Classification (ASJC) codes

Software

Keywords

Data-flow principles of execution
expert systems
parallel processing
pattern matching
production systems

Access to Document

10.1109/32.48936

Cite this

@article{4ed4849bbcd04a7a828820a3fc78bcd3,

title = "Data-Driven Parallel Production Systems",

abstract = "Much effort has been expended on developing special architectures dedicated to the efficient execution of production systems. While data-flow principles of execution offer the promise of high programmability for numerical computations, we demonstrate here that the data driven principles can also be applied to symbolic computations. In particular, we consider a mapping of the RETE match algorithm along the line of production systems. Bottlenecks of the RETE match algorithm in multiprocessor environment are identified, based on which the algorithm is parallelized. The modifications to the actor set as well as the program graph design are shown for execution on the Tagged-Token Data-flow Computer. The results of a deterministic simulation of this multiprocessor architecture demonstrate that artificial intelligence production systems can be efficiently mapped on data-driven architectures.",

keywords = "Data-flow principles of execution, expert systems, parallel processing, pattern matching, production systems",

author = "Gaudiot, {Jean Luc} and Andrew Sohn",

note = "Funding Information: Manuscript received May I. 1987: revised September 29. 1989. This work was supported in part by the National Science Foundation under Grenl CCR-8603772 and by the USC Faculty Research and Innovation Fund. The authors are with the Department of Electrical Engineering-Sys-terns. University of Southern California. Los Angeles. CA 90089. lEEE Log Number 8933197.",

year = "1990",

month = mar,

doi = "10.1109/32.48936",

language = "English (US)",

volume = "16",

pages = "281--293",

journal = "IEEE Transactions on Software Engineering",

issn = "0098-5589",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "3",

}

TY - JOUR

T1 - Data-Driven Parallel Production Systems

AU - Gaudiot, Jean Luc

AU - Sohn, Andrew

N1 - Funding Information: Manuscript received May I. 1987: revised September 29. 1989. This work was supported in part by the National Science Foundation under Grenl CCR-8603772 and by the USC Faculty Research and Innovation Fund. The authors are with the Department of Electrical Engineering-Sys-terns. University of Southern California. Los Angeles. CA 90089. lEEE Log Number 8933197.

PY - 1990/3

Y1 - 1990/3

N2 - Much effort has been expended on developing special architectures dedicated to the efficient execution of production systems. While data-flow principles of execution offer the promise of high programmability for numerical computations, we demonstrate here that the data driven principles can also be applied to symbolic computations. In particular, we consider a mapping of the RETE match algorithm along the line of production systems. Bottlenecks of the RETE match algorithm in multiprocessor environment are identified, based on which the algorithm is parallelized. The modifications to the actor set as well as the program graph design are shown for execution on the Tagged-Token Data-flow Computer. The results of a deterministic simulation of this multiprocessor architecture demonstrate that artificial intelligence production systems can be efficiently mapped on data-driven architectures.

AB - Much effort has been expended on developing special architectures dedicated to the efficient execution of production systems. While data-flow principles of execution offer the promise of high programmability for numerical computations, we demonstrate here that the data driven principles can also be applied to symbolic computations. In particular, we consider a mapping of the RETE match algorithm along the line of production systems. Bottlenecks of the RETE match algorithm in multiprocessor environment are identified, based on which the algorithm is parallelized. The modifications to the actor set as well as the program graph design are shown for execution on the Tagged-Token Data-flow Computer. The results of a deterministic simulation of this multiprocessor architecture demonstrate that artificial intelligence production systems can be efficiently mapped on data-driven architectures.

KW - Data-flow principles of execution

KW - expert systems

KW - parallel processing

KW - pattern matching

KW - production systems

UR - http://www.scopus.com/inward/record.url?scp=0025401692&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0025401692&partnerID=8YFLogxK

U2 - 10.1109/32.48936

DO - 10.1109/32.48936

M3 - Article

AN - SCOPUS:0025401692

SN - 0098-5589

VL - 16

SP - 281

EP - 293

JO - IEEE Transactions on Software Engineering

JF - IEEE Transactions on Software Engineering

IS - 3

ER -

Data-Driven Parallel Production Systems

Abstract

All Science Journal Classification (ASJC) codes

Keywords

Access to Document

Other files and links

Fingerprint

Cite this