Towards Parallel Non Serial Dynamic Programming for Solving Hard Weighted CSP

  • David Allouche
  • Simon de Givry
  • Thomas Schiex
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6308)

Abstract

We introduce a parallelized version of tree-decomposition based dynamic programming for solving difficult weighted CSP instances on many cores. A tree decomposition organizes cost functions in a tree of collection of functions called clusters. By processing the tree from the leaves up to the root, we solve each cluster concurrently, for each assignment of its separator, using a state-of-the-art exact sequential algorithm. The grain of parallelism obtained in this way is directly related to the tree decomposition used. We use a dedicated strategy for building suitable decompositions.

We present preliminary results of our prototype running on a cluster with hundreds of cores on different decomposable real problems. This implementation allowed us to solve the last open CELAR radio link frequency assignment instance to optimality.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Bertelé, U., Brioshi, F.: Nonserial Dynamic Programming. Academic Press, London (1972)MATHGoogle Scholar
  2. 2.
    Cabon, B., de Givry, S., Lobjois, L., Schiex, T., Warners, J.: Radio link frequency assignment. Constraints 4, 79–89 (1999)MATHCrossRefGoogle Scholar
  3. 3.
    Dechter, R.: Bucket elimination: A unifying framework for reasoning. Artificial Intelligence 113(1-2), 41–85 (1999)MATHCrossRefMathSciNetGoogle Scholar
  4. 4.
    Dechter, R., Pearl, J.: Tree clustering for constraint networks. Artificial Intelligence 38, 353–366 (1989)MATHCrossRefMathSciNetGoogle Scholar
  5. 5.
    Kask, K., Dechter, R., Larrosa, J., Dechter, A.: Unifying Cluster-Tree Decompositions for Reasoning in Graphical models. Artificial Intelligence 166(1-2), 165–193 (2005)MATHCrossRefMathSciNetGoogle Scholar
  6. 6.
    Larrosa, J., de Givry, S., Heras, F., Zytnicki, M.: Existential arc consistency: getting closer to full arc consistency in weighted CSPs. In: Proc. of the 19th IJCAI, Edinburgh, Scotland, pp. 84–89 (August 2005)Google Scholar
  7. 7.
    Lauritzen, S., Spiegelhalter, D.: Local computations with probabilities on graphical structures and their application to expert systems. Journal of the Royal Statistical Society – Series B 50, 157–224 (1988)MATHMathSciNetGoogle Scholar
  8. 8.
    Lecoutre, C., Saïs, L., Tabary, S., Vidal, V.: Reasoning from last conflict(s) in constraint programming. Artificial Intelligence 173, 1592–1614 (2009)MATHCrossRefMathSciNetGoogle Scholar
  9. 9.
    Otten, L., Dechter, R.: Towards parallel search for optimization in graphical models. In: Proc. of ISAIM 2010. Fort Lauderdale (FL), USA (2010)Google Scholar
  10. 10.
    Rose, D.: Tringulated graphs and the elimination process. Journal of Mathematical Analysis and its Applications 32 (1970)Google Scholar
  11. 11.
    Sanchez, M., Allouche, D., de Givry, S., Schiex, T.: Russian doll search with tree decomposition. In: Proc. IJCAI 2009, San Diego, CA, USA, pp. 603–608 (2009)Google Scholar
  12. 12.
    Schiex, T., Fargier, H., Verfaillie, G.: Valued constraint satisfaction problems: hard and easy problems. In: Proc. of the 14th IJCAI, Montréal, Canada, pp. 631–637 (August 1995)Google Scholar
  13. 13.
    Xia, Y., Prasanna, V.: Parallel Exact Inference. In: Bischof, C., Bücker, M., Gibbon, P. (eds.) Parallel Computing 2007. NIC, vol. 38, pp. 185–192. John von Neumann Institute for Computing (September 2007)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • David Allouche
    • 1
  • Simon de Givry
    • 1
  • Thomas Schiex
    • 1
  1. 1.Unité de Biométrie et Intelligence Artificielle, UR 875, INRACastanet TolosanFrance

Personalised recommendations