Abstract
In constraint satisfaction, pairwise consistency (PWC) is a well-known local consistency improving generalized arc consistency in theory but not often in practice. A popular approach to enforcing PWC enforces arc consistency on the dual encoding of the problem, allowing to reuse existing AC algorithms. In this paper, we explore the benefit of this simple approach in the optimization context of cost function networks and soft local consistencies. Using a dual encoding, we obtain an equivalent binary cost function network where enforcing virtual arc consistency achieves virtual PWC on the original problem. We experimentally observed that adding extra non-binary cost functions before the dual encoding results in even stronger bounds. Such supplementary cost functions may be produced by bounded variable elimination or by adding ternary zero-cost functions. Experiments on (probabilistic) graphical models, from the UAI 2022 competition benchmark, show a clear improvement when using our approach inside a branch-and-bound solver compared to the state-of-the-art.
This research was funded by the grants ANR-18-EURE-0021 and ANR-19-P3IA-0004. It receives support from the Genotoul (Toulouse) Bioinformatic platform.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
This corresponds to full PWC because unary constraints may appear in these pairs.
- 2.
This is unsurprising because the strongest bound that can be derived using EPTs is obtained using a linear program which includes pairwise consistency constraints [34].
- 3.
https://uaicompetition.github.io/uci-2022, see MPE and MMAP entries.
- 4.
- 5.
Options -A -P=1000 -T=1000 -vacint -vacthr -rasps -raspsini in toulbar2-vacint.
- 6.
Called double encoding in [26], it allows more flexibility to enforce various levels of consistency from GAC to PWC depending on the selected channeling constraints.
- 7.
- 8.
It is done only if the median degree in the original problem is less than 8, eliminating variables with a current degree less than or equal to the original median degree.
- 9.
With additional options -i -pils -p = -8 -O = −3 -t = 1. A triangle is defined by three variables involved in three binary cost functions. The score of a triangle is given by the average cost in the three functions. Triangles with the largest score are selected first. This approach allows to simulate soft path inverse consistency [22].
- 10.
See toulbar2-ipr results on the UAI 2022 Tuning Leader Board. Multiple runs of VNS with increasing floating-point precision were done with a total amount of time of \(\frac{1}{2}\)h. The remaining time is allocated to HBFS-VPWC. Each search procedure gives its best solution found to the next search procedure. On UAI 2022 tuning instances, this approach found 119 best solutions, ranking first among our 7 tested methods.
References
Allouche, D., et al.: Computational protein design as an optimization problem. Artif. Intell. 212, 59–79 (2014)
Bensana, E., Lemaître, M., Verfaillie, G.: Earth observation satellite management. Constraints 4(3), 293–299 (1999)
Beuvin, F., de Givry, S., Schiex, T., Verel, S., Simoncini, D.: Iterated local search with partition crossover for computational protein design. Proteins Struct. Funct. Bioinf. 87, 1522–1529 (2021)
Boussemart, F., Hemery, F., Lecoutre, C., Sais, L.: Boosting systematic search by weighting constraints. In: ECAI. vol. 16, p. 146 (2004)
Cabon, B., de Givry, S., Lobjois, L., Schiex, T., Warners, J.: Radio link frequency assignment. Constraints J. 4, 79–89 (1999)
Cooper, M., de Givry, S., Sanchez, M., Schiex, T., Zytnicki, M., Werner, T.: Soft arc consistency revisited. Artif. Intell. 174(7–8), 449–478 (2010)
Cooper, M.C.: High-order consistency in valued constraint satisfaction. Constraints 10, 283–305 (2005)
Cooper, M.C., de Givry, S., Schiex, T.: Graphical models: queries, complexity, algorithms (tutorial). In: 37th International Symposium on Theoretical Aspects of Computer Science (STACS-20). LIPIcs, vol. 154, pp. 4:1–4:22. Montpellier, France (2020)
Cooper, M.C., de Givry, S., Schiex, T.: Valued constraint satisfaction problems. In: Marquis, P., Papini, O., Prade, H. (eds.) A Guided Tour of Artificial Intelligence Research, pp. 185–207. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-06167-8_7
Dechter, R.: Bucket elimination: a unifying framework for reasoning. Artif. Intell. 113(1–2), 41–85 (1999)
Dechter, R., Rish, I.: Mini-buckets: a general scheme for bounded inference. J. ACM (JACM) 50(2), 107–153 (2003)
Demirović, E., Chu, G., Stuckey, P.J.: Solution-based phase saving for CP: a value-selection heuristic to simulate local search behavior in complete solvers. In: Hooker, J. (ed.) CP 2018. LNCS, vol. 11008, pp. 99–108. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-98334-9_7
Dlask, T., Werner, T., de Givry, S.: Bounds on weighted CSPs using constraint propagation and super-reparametrizations. In: Proceedings of CP-21. Montpellier, France (2021)
Favier, A., de Givry, S., Legarra, A., Schiex, T.: Pairwise decomposition for combinatorial optimization in graphical models. In: Proceedings of IJCAI-11. Barcelona, Spain (2011). http://www.inra.fr/mia/T/degivry/Favier11.mov
Hurley, B., et al.: Multi-language evaluation of exact solvers in graphical model discrete optimization. Constraints 21(3), 413–434 (2016)
Janssen, P., Jégou, P., Nouguier, B., Vilarem, M.C.: A filtering process for general constraint-satisfaction problems: achieving pairwise-consistency using an associated binary representation. In: IEEE International Workshop on Tools for Artificial Intelligence, pp. 420–421. IEEE Computer Society (1989)
Kappes, J.H., et al.: A comparative study of modern inference techniques for structured discrete energy minimization problems. Intl. J. of Comput. Vis. 115(2), 155–184 (2015)
Larrosa, J.: Boosting search with variable elimination. In: Dechter, R. (ed.) CP 2000. LNCS, vol. 1894, pp. 291–305. Springer, Heidelberg (2000). https://doi.org/10.1007/3-540-45349-0_22
Larrosa, J., Heras, F.: Resolution in Max-SAT and its relation to local consistency in weighted CSPs. In: Proceedings of IJCAI 2005, pp. 193–198. Edinburgh, Scotland (2005)
Lecoutre, C., Saïs, L., Tabary, S., Vidal, V.: Reasoning from last conflict(s) in constraint programming. Artif. Intell. 173, 1592–1614 (2009)
Neveu, B., Trombettoni, G., Glover, F.: ID Walk: a candidate list strategy with a simple diversification device. In: Proceedings of CP, pp. 423–437 (2004)
Nguyen, H., Bessiere, C., de Givry, S., Schiex, T.: Triangle-based consistencies for cost function networks. Constraints 22(2), 230–264 (2017)
Otten, L., Ihler, A., Kask, K., Dechter, R.: Winning the pascal 2011 map challenge with enhanced AND/OR branch-and-bound. In: DISCML 2012 Workshop, at NIPS 2012. Lake Tahoe, NV (2012)
Quali, A.: Variable neighborhood search for graphical model energy minimization. Artif. Intell. 278(103194), 22p (2020)
Rossi, F., Petrie, C.J., Dhar, V.: On the equivalence of constraint satisfaction problems. In: ECAI, vol. 90, pp. 550–556 (1990)
Samaras, N., Stergiou, K.: Binary encodings of non-binary constraint satisfaction problems: algorithms and experimental results. J. Artif. Intell. Res. 24, 641–684 (2005)
Sánchez, M., de Givry, S., Schiex, T.: Mendelian error detection in complex pedigrees using weighted constraint satisfaction techniques. Constraints 13(1), 130–154 (2008)
Savchynskyy, B.: Discrete graphical models - an optimization perspective. Found. Trends Comput. Graph. Vis. 11(3–4), 160–429 (2019)
Schneider, A., Choueiry, B.Y.: PW-AC: extending compact-table to enforce pairwise consistency on table constraints. In: Hooker, J. (ed.) CP 2018. LNCS, vol. 11008, pp. 345–361. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-98334-9_23
Trösser, F., de Givry, S., Katsirelos, G.: Relaxation-aware heuristics for exact optimization in graphical models. In: Proceedings of CP-AI-OR’2020, pp. 475–491. Vienna, Austria (2020)
Wainwright, M.J., Jordan, M.I., et al.: Graphical models, exponential families, and variational inference. Found. Trends Mach. Learn. 1(1–2), 1–305 (2008)
Wang, R., Yap, R.H.C.: Arc consistency revisited. In: Rousseau, L.-M., Stergiou, K. (eds.) CPAIOR 2019. LNCS, vol. 11494, pp. 599–615. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-19212-9_40
Wang, R., Yap, R.H.: Bipartite encoding: a new binary encoding for solving non-binary CSPs. In: Proceedings of the Twenty-Ninth International Conference on International Joint Conferences on Artificial Intelligence, pp. 1184–1191 (2021)
Werner, T.: Revisiting the linear programming relaxation approach to Gibbs energy minimization and weighted constraint satisfaction. IEEE Trans. Pattern Anal. Mach. Intell. 32(8), 1474–1488 (2010)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Montalbano, P., Allouche, D., de Givry, S., Katsirelos, G., Werner, T. (2023). Virtual Pairwise Consistency in Cost Function Networks. In: Cire, A.A. (eds) Integration of Constraint Programming, Artificial Intelligence, and Operations Research. CPAIOR 2023. Lecture Notes in Computer Science, vol 13884. Springer, Cham. https://doi.org/10.1007/978-3-031-33271-5_27
Download citation
DOI: https://doi.org/10.1007/978-3-031-33271-5_27
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-33270-8
Online ISBN: 978-3-031-33271-5
eBook Packages: Computer ScienceComputer Science (R0)