Abstract rewriting approach to solve datalog programs

Fernando Tarin Morales, Fuyuki Isihikawa, Shinichi Honiden

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Over the past decade, we have seen a resurgence in the Datalog language in different computing areas for solving a number of nontrivial problems. In this paper we introduce a novel resolution approach to solve Datalog programs. We present a version of the technique that works on plain Datalog programs. We have developed an abstract rewriting formalism to create a functional resolution process for Datalog. The resolution approach translates the Datalog resolution strategy into a fix-point abstract rewriting equation system. Being an abstract rewriting formalism, every equation of the system can be viewed as a function. Based on this fact, we also developed an optimization process that improves the initial rewriting equation system. The optimization process generates an equation system that computes the solutions much more efficiently. Well known optimizations such as strength reduction or memoization have been used. We also developed a prototype compiler that encodes the optimized equation system into a solver. Experimental results obtained with the solver suggest execution times several orders of magnitude better than modern Prolog solvers like YAP or XSB and usually one order of magnitude faster than state-of-the-art Datalog solvers such as BDDBDDB and DLV.

Original languageEnglish
Title of host publicationDBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages
PublisherAssociation for Computing Machinery, Inc
Pages29-36
Number of pages8
ISBN (Electronic)9781450339025
DOIs
Publication statusPublished - 2015 Oct 27
Externally publishedYes
Event15th Symposium on Database Programming Languages, DBPL 2015 - Pittsburgh, United States
Duration: 2015 Oct 27 → …

Other

Other15th Symposium on Database Programming Languages, DBPL 2015
CountryUnited States
CityPittsburgh
Period15/10/27 → …

Keywords

  • Abstract Rewriting Systems
  • Datalog
  • Deductive Databases
  • In-memory Databases

ASJC Scopus subject areas

  • Computer Science Applications

Cite this

Morales, F. T., Isihikawa, F., & Honiden, S. (2015). Abstract rewriting approach to solve datalog programs. In DBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages (pp. 29-36). Association for Computing Machinery, Inc. https://doi.org/10.1145/2815072.2815076

Abstract rewriting approach to solve datalog programs. / Morales, Fernando Tarin; Isihikawa, Fuyuki; Honiden, Shinichi.

DBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages. Association for Computing Machinery, Inc, 2015. p. 29-36.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Morales, FT, Isihikawa, F & Honiden, S 2015, Abstract rewriting approach to solve datalog programs. in DBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages. Association for Computing Machinery, Inc, pp. 29-36, 15th Symposium on Database Programming Languages, DBPL 2015, Pittsburgh, United States, 15/10/27. https://doi.org/10.1145/2815072.2815076
Morales FT, Isihikawa F, Honiden S. Abstract rewriting approach to solve datalog programs. In DBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages. Association for Computing Machinery, Inc. 2015. p. 29-36 https://doi.org/10.1145/2815072.2815076
Morales, Fernando Tarin ; Isihikawa, Fuyuki ; Honiden, Shinichi. / Abstract rewriting approach to solve datalog programs. DBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages. Association for Computing Machinery, Inc, 2015. pp. 29-36
@inproceedings{365c5a41cc1e4f3ea4cd65b1c0296ba8,
title = "Abstract rewriting approach to solve datalog programs",
abstract = "Over the past decade, we have seen a resurgence in the Datalog language in different computing areas for solving a number of nontrivial problems. In this paper we introduce a novel resolution approach to solve Datalog programs. We present a version of the technique that works on plain Datalog programs. We have developed an abstract rewriting formalism to create a functional resolution process for Datalog. The resolution approach translates the Datalog resolution strategy into a fix-point abstract rewriting equation system. Being an abstract rewriting formalism, every equation of the system can be viewed as a function. Based on this fact, we also developed an optimization process that improves the initial rewriting equation system. The optimization process generates an equation system that computes the solutions much more efficiently. Well known optimizations such as strength reduction or memoization have been used. We also developed a prototype compiler that encodes the optimized equation system into a solver. Experimental results obtained with the solver suggest execution times several orders of magnitude better than modern Prolog solvers like YAP or XSB and usually one order of magnitude faster than state-of-the-art Datalog solvers such as BDDBDDB and DLV.",
keywords = "Abstract Rewriting Systems, Datalog, Deductive Databases, In-memory Databases",
author = "Morales, {Fernando Tarin} and Fuyuki Isihikawa and Shinichi Honiden",
year = "2015",
month = "10",
day = "27",
doi = "10.1145/2815072.2815076",
language = "English",
pages = "29--36",
booktitle = "DBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - Abstract rewriting approach to solve datalog programs

AU - Morales, Fernando Tarin

AU - Isihikawa, Fuyuki

AU - Honiden, Shinichi

PY - 2015/10/27

Y1 - 2015/10/27

N2 - Over the past decade, we have seen a resurgence in the Datalog language in different computing areas for solving a number of nontrivial problems. In this paper we introduce a novel resolution approach to solve Datalog programs. We present a version of the technique that works on plain Datalog programs. We have developed an abstract rewriting formalism to create a functional resolution process for Datalog. The resolution approach translates the Datalog resolution strategy into a fix-point abstract rewriting equation system. Being an abstract rewriting formalism, every equation of the system can be viewed as a function. Based on this fact, we also developed an optimization process that improves the initial rewriting equation system. The optimization process generates an equation system that computes the solutions much more efficiently. Well known optimizations such as strength reduction or memoization have been used. We also developed a prototype compiler that encodes the optimized equation system into a solver. Experimental results obtained with the solver suggest execution times several orders of magnitude better than modern Prolog solvers like YAP or XSB and usually one order of magnitude faster than state-of-the-art Datalog solvers such as BDDBDDB and DLV.

AB - Over the past decade, we have seen a resurgence in the Datalog language in different computing areas for solving a number of nontrivial problems. In this paper we introduce a novel resolution approach to solve Datalog programs. We present a version of the technique that works on plain Datalog programs. We have developed an abstract rewriting formalism to create a functional resolution process for Datalog. The resolution approach translates the Datalog resolution strategy into a fix-point abstract rewriting equation system. Being an abstract rewriting formalism, every equation of the system can be viewed as a function. Based on this fact, we also developed an optimization process that improves the initial rewriting equation system. The optimization process generates an equation system that computes the solutions much more efficiently. Well known optimizations such as strength reduction or memoization have been used. We also developed a prototype compiler that encodes the optimized equation system into a solver. Experimental results obtained with the solver suggest execution times several orders of magnitude better than modern Prolog solvers like YAP or XSB and usually one order of magnitude faster than state-of-the-art Datalog solvers such as BDDBDDB and DLV.

KW - Abstract Rewriting Systems

KW - Datalog

KW - Deductive Databases

KW - In-memory Databases

UR - http://www.scopus.com/inward/record.url?scp=84956987548&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84956987548&partnerID=8YFLogxK

U2 - 10.1145/2815072.2815076

DO - 10.1145/2815072.2815076

M3 - Conference contribution

SP - 29

EP - 36

BT - DBPL 2015 - Proceedings of the 15th Symposium on Database Programming Languages

PB - Association for Computing Machinery, Inc

ER -