Advertisement

Efficient Access To Large Prolog Knowledge Bases

  • Christos Garidis
  • Stefan Böttcher
Conference paper

Abstract

One requirement for language systems for knowledge based applications is to handle large knowledge bases efficiently. Large knowledge bases written in Prolog have a large load time from secondary storage. We describe how a Prolog system can be supported with a clustering concept for minimizing both the load time and the loaded code of large Prolog knowledge bases, which additionally enables an efficient cluster buffer managment, if knowledge base size exceeds the available main memory. Clusters are knowledge base partitions of equal size which are generated at compile time and contain semantically related clauses. The paper focusses on a comparative performance evaluation of a Prolog system supported with various cluster replacement strategies and compares “intelligent” cluster replacement strategies with conventional replacement strategies known from operating systems. The result of the performance evaluation is that the load time for Prolog knowledge bases can be reduced by using “intelligent” instead of conventional cluster replacement strategies.

Keywords

Buffer Size Cluster State Choice Point Active Cluster Load Time 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Be166.
    Belady, L.A.: A study of replacement algorithms for virtual storage computers. IBM Syst. J. 5, 2 (1966).CrossRefGoogle Scholar
  2. Boc90.
    Bocca, J.: Compilation of Logic Programs to Implement Very Large Knowledge Base Systems - A Case Study: Educe’. Proc. of the 6th Int. Conf. on Data Engineering, Los Angeles, USA, Feb. 90.Google Scholar
  3. BôBe89.
    Böttcher, S. Beierle, C.: Database Support for the PROTOSL System. Microprocessing and Microprogramming, Vol. 27, 1989.Google Scholar
  4. Den68.
    Denning, P.G.: The working set model for program behavior. Commun. ACM 11, 5 (1968), 323–333.MathSciNetMATHCrossRefGoogle Scholar
  5. EfHä84.
    Effelsberg, W. Härder, T.: Principles of Database Buffer Managment. ACM TODS, Vol. 9, No. 4, Dec. 84.Google Scholar
  6. Gar90.
    Garidis, C.: Clustering-Konzepte für optimalen Zugriff auf große und datenbankresidente Wissensbasen. PhD Thesis, Uni. Stuttgart, Dezember 90.Google Scholar
  7. Herz86.
    Herzog O. et. al: LILOG - Linguistic and Logic Methods for the Computational Understanding of German. TR lb, IBM Germany, 1986.Google Scholar
  8. RoMa84.
    Ross M.L. McMahon A.G.: Memory Managment of a Sequential Prolog Interpreter. Technical Report 846, RMIT, 1984.Google Scholar
  9. RoRa86.
    Ross M.L. Ramamohanarao K.: Paging Strategy for Prolog based dynamic virtual Memory. Symp. on Logic Programming, Salt Lake City, Utah, Sep. 86.Google Scholar
  10. SaSc86.
    Sacco, G.M. Schkolnick, M.: Buffer Management in Relational Database Systems. ACM TODS, Vol. 11, No. 4, Dec. 86.Google Scholar
  11. Ston81.
    Stonebraker, M.: Operating system support for database managment systems. Commun. ACM 24, 7 (1981), 412–418.CrossRefGoogle Scholar
  12. VMP.
    VM/Programming in Logic, Program Description and Operations Manual. Program Number 5785-ABH, SB11–63740, IBM, 1985.Google Scholar
  13. WaTz88.
    Walker, A. Tzoar, D.: The SYLLOG Expert Database System. Notes for Users, Version 0.7, IBM T.J. Watson Research Center, 1988.Google Scholar
  14. War83.
    Warren, H.D.: An Abstract Prolog Instruction Set. SRI Tchnical Note 309, 1983.Google Scholar
  15. Zin89.
    Zink, J.: Entwurf und Implementierung eines Meßssystems zur Evaluierung des Clustering-Konzeptes des PROTOS-L- Systems. Diplomarbeit 607, Universität Stuttgart, Dec. 1989.Google Scholar

Copyright information

© Springer-Verlag Wien 1991

Authors and Affiliations

  • Christos Garidis
    • 1
  • Stefan Böttcher
    • 2
  1. 1.Institute for Parallel and Distributed High Performance SystemsStuttgart UniversityStuttgart 1Germany
  2. 2.IBM Germany Scientific CenterInstitute for Knowledge Based SystemsStuttgart 1Germany

Personalised recommendations