Abstract
Parallel constraint solving is a promising way to enhance the performance of constraint programming. Yet, current solutions for parallel constraint solving ignore the importance of hypergraph decomposition when mapping constraints onto cores. This paper explains why and how the hypergraph decomposition can be employed to relatively evenly distribute workload in parallel constraint solving. We present our dedicated hypergraph decomposition method det-k-CP for parallel constraint solving. The result of det-k-CP, which conforms with four conditions of hypertree decomposition, can be used to allocate constraints of a given constraint network to cores for parallel constraint solving. Our benchmark evaluations have shown that det-k-CP can relatively evenly decompose a hypergraph for specific scale of constraint networks. Besides, we obtained competitive execution time as long as the hypergraphs are sufficiently simple.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
The names of the constraints are consistent with the names of constraints used in the Choco Solver [10].
- 2.
Please note that the verb “eliminate” does not mean an edge is deleted, it means that we can ignore the join selection for the nodes connected by this edge.
- 3.
The term permutation is explained in Algorithm 1.
- 4.
We observed all the instances can be successfully decomposed by det-k-decomp when the width is 2.
References
Gottlob, G., Samer, M.: A backtracking-based algorithm for hypertree decomposition. J. Exp. Algorithmics 13, 1 (2009)
Gottlob, G., Leone, N., Scarcello, F.: Advanced parallel algorithms for acyclic conjunctive queries. Technical Report DBAI-TR-98/18 (1998). http://www.dbai.tuwien.ac.at/staff/gottlob/parallel.ps
Gottlob, G., Leone, N., Scarcello, F.: The complexity of acyclic conjunctive queries. J. ACM (JACM) 48(3), 431–498 (2001)
Gottlob, G., Greco, G., Leone, N., Scarcello, F.: Hypertree decompositions: questions and answers. In: Proceedings of the 35th ACM SIGMOD-SIGACT-SIGAI Symposium on Principles of Database Systems, pp. 57–74. ACM (2016)
Dechter, R.: Constraint Processing. Morgan Kaufmann, Burlington (2003)
Dechter, R., Pearl, J.: The cycle-cutset method for improving search performance in AI applications. In: Third IEEE Conference on AI Applications, pp. 224–230. IEEE (1987)
Gyssens, M., Paredaens, J.: A decomposition methodology for cyclic databases. In: Gallaire, H., Nicolas, J., Minker, J. (eds.) Advances in Data Base Theory, vol. 2, pp. 85–122. Plemum Press, New York (1984). https://doi.org/10.1007/978-1-4615-9385-0_4
Gottlob, G., Leone, N., Scarcello, F.: Hypertree decompositions and tractable queries. In: Proceedings of the Eighteenth ACM SIGMOD-SIGACT-SIGART Symposium on Principles of Database Systems, pp. 21–32. ACM (1999)
Gottlob, G., Leone, N., Scarcello, F.: A comparison of structural CSP decomposition methods. Artif. Intell. 124(2), 243–282 (2000)
Prud’homme, C., Fages, J.G., Lorca, X.: Choco Documentation. TASC, INRIA Rennes, LINA CNRS UMR 6241, COSLING S.A.S. (2016)
Gottlob, G., Leone, N., Scarcello, F.: Robbers, marshals, and guards: game theoretic and logical characterizations of hypertree width. J. Comput. Syst. Sci. 66(4), 775–808 (2003)
Acknowledgments
We would like to express our special thanks to Georg Gottlob and Wolfgang Fischl for their source code of det-k-decomp, especially the benchmark suite for hypertree decomposition.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer Nature Switzerland AG
About this paper
Cite this paper
Liu, K., Löffler, S., Hofstedt, P. (2018). Hypertree Decomposition: The First Step Towards Parallel Constraint Solving. In: Seipel, D., Hanus, M., Abreu, S. (eds) Declarative Programming and Knowledge Management. WFLP WLP INAP 2017 2017 2017. Lecture Notes in Computer Science(), vol 10997. Springer, Cham. https://doi.org/10.1007/978-3-030-00801-7_6
Download citation
DOI: https://doi.org/10.1007/978-3-030-00801-7_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-00800-0
Online ISBN: 978-3-030-00801-7
eBook Packages: Computer ScienceComputer Science (R0)