Using Abstraction for Planning in Sokoban

  • Adi Botea
  • Martin Müller
  • Jonathan Schaeffer
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2883)


Heuristic search has been successful for games like chess and checkers, but seems to be of limited value in games such as Go and shogi, and puzzles such as Sokoban. Other techniques are necessary to approach the performance that humans achieve in these hard domains. This paper explores using planning as an alternative problem-solving framework for Sokoban. Previous attempts to express Sokoban as a planning application led to poor performance results. Abstract Sokoban is introduced as a new planning formulation of the domain. The approach abstracts a Sokoban problem into rooms and tunnels. This allows for the decomposition of the hard initial problem into several simpler sub-problems, each of which can be solved efficiently. The experimental results show that the abstraction has the potential for an exponential reduction in the size of the search space explored.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Junghanns, A.: Pushing the Limits: New Developments in Single-Agent Search. PhD thesis, Department of Computing Science, University of Alberta (1999)Google Scholar
  2. 2.
    Bacchus, F.: AIPS’00 planning competition. AI Magazine, 47–56 (2001)Google Scholar
  3. 3.
    Shapiro, A.: Structured Induction in Expert Systems. Turing Institute Press. Addison-Wesley (1987)Google Scholar
  4. 4.
    Wilkins, D.: Using knowledge to control tree searching. Artificial Intelligence 18, 1–51 (1982)CrossRefMathSciNetGoogle Scholar
  5. 5.
    Korf, R.: Macro-operators: A weak method for learning. Artificial Intelligence 26(1), 35–77 (1985)zbMATHCrossRefMathSciNetGoogle Scholar
  6. 6.
    Culberson, J.: Sokoban is PSPACE-complete. Technical report, Department of Computing Science, University of Alberta, Edmonton, Alberta, Canada (1997)
  7. 7.
    Junghanns, A., Schaeffer, J.: Sokoban: Enhancing single-agent search using domain knowledge. Artificial Intelligence 129, 219–251 (2001)zbMATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    McDermott, D.: Using regression-match graphs to control search in planning (1997)
  9. 9.
    Holte, R., Perez, M., Zimmer, R., MacDonald, A.: Hierarchical A*: Searching abstraction hierarchies efficiently. Technical report, University of Ottawa, TR-95-18 (1995)Google Scholar
  10. 10.
    Bacchus, F., Kabanza, F.: Using temporal logics to express search control knowledge for planning. Artificial Intelligence 16, 123–191 (2000)CrossRefMathSciNetGoogle Scholar
  11. 11.
    Junghanns, A., Schaeffer, J.: Domain-dependent single-agent search enhancements. In: Sixteenth International Joint Conference on Artificial Intelligence (IJCAI 1999), pp. 570–575. Morgan Kaufmann Publishers, San Francisco (1999)Google Scholar
  12. 12.
    Junghanns, A., Schaeffer, J.: Single-agent search in the presence of deadlock. In: Fifteenth National Conference of the American Association for Artificial Intelligence (AAAI 1998), pp. 419–424. AAAI Press, Menlo Park (1998)Google Scholar
  13. 13.
    Bacchus, F.: Personal communication (2002)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Adi Botea
    • 1
  • Martin Müller
    • 1
  • Jonathan Schaeffer
    • 1
  1. 1.Department of Computing ScienceUniversity of AlbertaCanada

Personalised recommendations