Scaling effects in the CSP phase transition

  • Ian P. Gent
  • Ewan MacIntyre
  • Patrick Prosser
  • Toby Walsh
Constraint Satisfaction Problems 1
Part of the Lecture Notes in Computer Science book series (LNCS, volume 976)


Phase transitions in constraint satisfaction problems (CSP's) are the subject of intense study. We identify a control parameter for random binary CSP's. There is a rapid transition in the probability of a CSP having a solution at a critical value of this parameter. This parameter allows different phase transition behaviour to be compared in an uniform manner, for example CSP's generated under different regimes. We then show that within classes, the scaling of behaviour can be modelled by a technique called “finite size scaling”. This applies not only to probability of solubility, as has been observed before in other NP-problems, but also to search cost. Furthermore, the technique applies with equal validity to several different methods of varying problem size. As well as contributing to the understanding of phase transitions, we contribute by allowing much finer grained comparison of algorithms, and for accurate empirical extrapolations of behaviour.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    M.N. Barber. Finite-size scaling. In Phase Transitions and Critical Phenomena, Volume 8, pages 145–266. Academic Press, 1983.Google Scholar
  2. 2.
    P. Cheeseman, B. Kanefsky, and W.M. Taylor. Where the really hard problems are. In Proceedings of the 12th IJCAI, pages 331–337. International Joint Conference on Artificial Intelligence, 1991.Google Scholar
  3. 3.
    R. Dechter, Constraint Networks, in Encyclopedia of Artificial Intelligence, Wiley, New York, 2nd ed., 276–286, 1992.Google Scholar
  4. 4.
    R. Dechter and I. Meiri, Experimental evaluation of preprocessing algorithms for constraint satisfaction problems, Artif. Intell. 68(2) (1994) 211–242.Google Scholar
  5. 5.
    D. Frost and R. Dechter, In search of the best search: an empirical evaluation, Proceedings AAAI-94, Seattle, WA (1994) 301–306.Google Scholar
  6. 6.
    J. Gaschnig, A general backtracking algorithm that eliminates most redundant tests, Proceedings IJCAI-77, Cambridge, MA (1977) 457.Google Scholar
  7. 7.
    J. Gaschnig, Performance measurement and analysis of certain search algorithms, Tech. Rept. CMU-CS-79-124, Carnegie-Mellon University, Pittsburgh, PA (1979).Google Scholar
  8. 8.
    I. P. Gent and T. Walsh. The SAT phase transition. In Proceedings of ECAI-94, pages 105–109, 1994.Google Scholar
  9. 9.
    I.P. Gent and T. Walsh. The satisfiability constraint gap. To appear in Artificial Intelligence.Google Scholar
  10. 10.
    I.P. Gent and T. Walsh. The TSP phase transition. Research report 95-178, Department of Computer Science, University of Strathclyde, 1995. Presented at First Workshop on AI and OR, Timberline, Oregon.Google Scholar
  11. 11.
    S.W. Golomb and L.D. Baumert, Backtrack programming. JACM 12 (1965) 516–524.CrossRefGoogle Scholar
  12. 12.
    R.M. Haralick and G.L. Elliott, Increasing Tree Search Efficiency for Constraint Satisfaction Problems, Artif. Intell. 14 (1980) 263–313.CrossRefGoogle Scholar
  13. 13.
    T. Hogg and C. Williams. The hardest constraint problems: A double phase transition. Artificial Intelligence, 69:359–377, 1994.CrossRefGoogle Scholar
  14. 14.
    J.N. Hooker, Needed: An empirical science of algorithms, Operations Research 42 (2) (1994) 201–212.Google Scholar
  15. 15.
    S. Kirkpatrick, G. Györgyi, N. Tishby, and L. Troyansky. The statistical mechanics of k-satisfaction. In Advances in Neural Information Processing Systems 6, pages 439–446. Morgan Kaufmann, 1994.Google Scholar
  16. 16.
    S. Kirkpatrick and B. Selman. Critical behavior in the satisfiability of random boolean expressions. Science, 264:1297–1301, May 27 1994.Google Scholar
  17. 17.
    V. Kumar, Algorithms for constraint satisfaction problems: a survey, AI magazine 13(1) (1992) 32–44.Google Scholar
  18. 18.
    E. MacIntyre, Really hard problems, Final Year Report for the B.Sc. degree, Department of Computer Science, University of Strathclyde, Scotland, 1994.Google Scholar
  19. 19.
    D. Mitchell, B. Selman, and Hector Levesque. Hard and easy distributions of SAT problems. In Proceedings, 10th National Conference on Artificial Intelligence. AAAI Press/The MIT Press, pages 459–465, 1992.Google Scholar
  20. 20.
    P. Prosser. Hybrid algorithms for the constraint satisfaction problem. Computational Intelligence, 9:268–299, 1993.Google Scholar
  21. 21.
    P. Prosser. Binary constraint satisfaction problems: Some are harder than others. In Proceedings of ECAI-94, pages 95–99, 1994.Google Scholar
  22. 22.
    P. Prosser. An empirical study of phase transitions in binary constraint satisfaction problems. To appear in Artificial Intelligence.Google Scholar
  23. 23.
    P.W. Purdom, Search rearrangement backtracking and polynomial average time, Artif. Intell. 21 (1983) 117–133.Google Scholar
  24. 24.
    D. Sabin and E.C. Freuder, Contradicting conventional wisdom in constraint satisfaction, Proceedings ECAI-94, Amsterdam, The Netherlands (1994) 125–129.Google Scholar
  25. 25.
    B. Selman and S. Kirkpatrick. Critical behaviour in the computational cost of satisfiability testing. To appear in Artificial Intelligence.Google Scholar
  26. 26.
    B.M Smith. Phase transition and the mushy region in constraint satisfaction problems. In Proceedings of ECAI-94, pages 100–104, 1994.Google Scholar
  27. 27.
    B. M. Smith and M.E. Dyer. Locating the phase transition in binary constraint satisfaction problems. To appear in Artificial Intelligence.Google Scholar
  28. 28.
    E.P.K. Tsang, Foundations of Constraint Satisfaction, Academic Press, 1993.Google Scholar
  29. 29.
    E.P.K. Tsang, J. Borrett, and A.C.M. Kwan, An attempt to map the performance of a range of algorithm and heuristic combinations, In Proceedings AISB-95, pages 203–216, Ed. J. Hallam, IOS Press, Amsterdam, 1995.Google Scholar
  30. 30.
    C.P. Williams and T. Hogg. Exploiting the deep structure of constraint problems. Artificial Intelligence, 70:73–117, 1994.CrossRefGoogle Scholar
  31. 31.
    K.G. Wilson. Problems in physics with many scales of length. Scientific American, 241:140–157, 1979.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1995

Authors and Affiliations

  • Ian P. Gent
    • 1
  • Ewan MacIntyre
    • 1
  • Patrick Prosser
    • 1
  • Toby Walsh
    • 2
    • 3
  1. 1.Department of Computer ScienceUniversity of StrathclydeGlasgowUK
  2. 2.Mechanized Reasoning GroupIRSTTrento
  3. 3.DISTUniversity of GenoaItaly

Personalised recommendations