Abstract
Within the Constraints Satisfaction Problems (CSP) context, a methodology that has proven to be particularly performant consists in using a portfolio of different constraint solvers. Nevertheless, comparatively few studies and investigations have been done in the world of Constraint Optimization Problems (COP). In this work, we provide a generalization to COP as well as an empirical evaluation of different state of the art existing CSP portfolio approaches properly adapted to deal with COP. Experimental results confirm the effectiveness of portfolios even in the optimization field, and could give rise to some interesting future research.
Keywords
- Constraint Satisfaction Problem
- Satisfiability Modulo Theory
- Portfolio Size
- Portfolio Approach
- Constraint Optimization Problem
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
Work partially supported by Aeolus project, ANR-2010-SEGI-013-01.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsNotes
- 1.
Formally, \(\mathtt{MIN }(i) = \min V_i\) and \(\mathtt{MAX }(i) = \max V_i\) where \(V_i = \{ \mathtt{val }(s, i, T) \ . \ s \in U \}\). Note that a portfolio solver executing more than one solver for \(t < T\) seconds could produce a solution that is worse than \(\mathtt{MIN }(i)\). This however is very uncommon: in our experiments we noticed that the 0 score was assigned only to the solvers that did not find any solution.
- 2.
FlatZinc [6] is the low level language that each solver uses for solving a given MiniZinc instance. A key feature of FlatZinc is that, starting from a general MiniZinc model, every solver can produce a specialized FlatZinc by redefining the global constraints definitions. We noticed that, especially for huge instances, the time needed for extracting features was strongly dominated by the FlatZinc conversion. However, for the instances of the final dataset this time was in average 10.36 s, with a maximum value of 504 s and a median value of 3.17 s.
- 3.
Following [1] methodology, CPX won all the elections we simulated using different criteria, viz.: Borda, Approval, and Plurality.
- 4.
The objective function of the best approach considered was obtained by replacing that of the IP problem defined in [18] (we use the very same notation) with:
$$\begin{aligned} \max \left[ C_1 \sum _y y_i + C_2 \sum _{i, S, t} \mathtt{score }(S, i, t) \cdot x_{S, t} + C_3 \sum _{S,t} t \cdot x_{S, t} \right] \end{aligned}$$where \(C_1 = -C^2\), \(C_2 = C\), \(C_3 = -\frac{1}{C}\), and adding the constraint \(\sum _t x_{S,t} \le 1\), \(\forall S\).
- 5.
For more details, we defer the interested reader to [37].
- 6.
To conduct the experiments we used Intel Dual-Core 2.93 GHz computers with 3 MB of CPU cache, 2 GB of RAM, and Ubuntu 12.04 operating system. For keeping track of the solving times we considered the CPU time by exploiting Unix time command.
References
Amadini, R., Gabbrielli, M., Mauro, J.: An empirical evaluation of portfolios approaches for solving CSPs. In: Gomes, C., Sellmann, M. (eds.) CPAIOR 2013. LNCS, vol. 7874, pp. 316–324. Springer, Heidelberg (2013)
Amadini, R., Gabbrielli, M., Mauro, J.: SUNNY: a Simple and dynamic algorithm portfolio for solving CSPs. CoRR, abs/1311.3353 (2013)
Amadini, R., Gabbrielli, M., Jacopo M.: An Enhanced Features Extractor for a Portfolio of Constraint Solvers. In: SAC (2014)
Arlot, S., Celisse, A.: A survey of cross-validation procedures for model selection. Stat. Surv. 4, 40–79 (2010)
Baral, C.: Knowledge Representation Reasoning and Declarative Problem Solving. Cambridge University Press, Cambridge (2003)
Becket, R.: Specification of FlatZinc - Version 1.6. http://www.minizinc.org/downloads/doc-1.6/flatzinc-spec.pdf
Biere, A., Heule, M., van Maaren, H., Walsh, T. (eds.): Handbook of Satisfiability. IOS Press, Amsterdam (2009)
Borenstein, Y., Poli, R.: Kolmogorov complexity, Optimization and Hardness pp. 12–119. In Evolutionary Computation (2006)
Carchrae, T., Beck, J.C.: Applying machine learning to low-knowledge control of optimization algorithms. Comput. Intell. 21(4), 372–387 (2005)
CPX Discrete Optimiser. http://www.opturion.com/cpx.html
Third International CSP Solver Competition 2008. http://www.cril.univ-artois.fr/CPAI09/
Gomes, C.P., Selman, B.: Algorithm portfolios. Artif. Intell. 126(1–2), 43–62 (2001)
Gomes, C.P., Selman, B., Crato, N.: Heavy-tailed distributions in combinatorial search. In: Smolka, G. (ed.) CP 1997. LNCS, vol. 1330. Springer, Heidelberg (1997)
Guo, H., Hsu, W.H.: A machine learning approach to algorithm selection for NP-hard optimization problems: a case study on the MPE problem. Ann. Oper. Res. 156(1), 61–82 (2007)
Hall, M., Frank, E., Holmes, G., Pfahringer, B., Reutemann, P., Witten, I.H.: The WEKA data mining software: an update. SIGKDD Explor. Newsl. 11(1), 10–18 (2009)
Hoos, H.H., Kaufmann, B., Schaub, T., Schneider, M.: Robust benchmark set selection for boolean constraint solvers. In: Nicosia, G., Pardalos, P. (eds.) LION 7. LNCS, vol. 7997, pp. 138–152. Springer, Heidelberg (2013)
Hutter, F., Xu, L., Hoos, H.H., Leyton-Brown, K.: Algorithm Runtime Prediction: The State of the Art. CoRR, abs/1211.0906 (2012)
Kadioglu, S., Malitsky, Y., Sabharwal, A., Samulowitz, H., Sellmann, M.: Algorithm selection and scheduling. In: Lee, J. (ed.) CP 2011. LNCS, vol. 6876, pp. 454–469. Springer, Heidelberg (2011)
Kadioglu, S., Malitsky, Y., Sellmann, M., Kevin T.: ISAC - Instance-specific algorithm configuration. In: ECAI (2010)
Knowles, J.D., Corne, D.W: Towards landscape analyses to inform the design of hybrid local search for the multiobjective quadratic assignment problem. In: HIS (2002)
Kotthoff, L.: Algorithm Selection for Combinatorial Search Problems: A Survey. CoRR, abs/1210.7959 (2012)
Kroer, C., Malitsky, Y.: Feature filtering for instance-specific algorithm configuration. In: ICTAI (2011)
Leyton-Brown, K., Nudelman, E., Shoham, Y.: The case of combinatorial auctions. In: CP, Learning the Empirical Hardness of Optimization Problems (2002)
Lobjois, L., Lemaître, M.: Branch and bound algorithm selection by performance prediction. In: AAAI/IAAI (1998)
Mackworth, A.K.: Consistency in networks of relations. Artif. Intell. 8(1), 99–118 (1977)
Malitsky, Y., Sabharwal, A., Samulowitz, H., Sellmann, M.: Algorithm portfolios based on cost-sensitive hierarchical clustering. In: IJCAI (2013)
Max-SAT 2013. http://maxsat.ia.udl.cat/introduction/
Merz, P.: Advanced fitness landscape analysis and the performance of memetic algorithms. Evol. Comput. 12(3), 303–325 (2004)
Minizinc version 1.6. http://www.minizinc.org/download.html
MiniZinc Challenge. http://www.minizinc.org/challenge2012/results2012.html
O’Mahony, E., Hebrard, E., Holland, A., Nugent, C., O’Sullivan, B.: Using case-based reasoning in an algorithm portfolio for constraint solving. In: AICS 08 (2009)
Rice, J.R.: The algorithm selection problem. Adv. Comput. 15, 65–118 (1976)
SAT Challenge 2012. http://baldur.iti.kit.edu/SAT-Challenge-2012/
Smith-Miles, K.A.: Cross-disciplinary perspectives on meta-learning for algorithm selection. ACM Comput. Surv. 41(1), 1–25 (2008)
Smith-Miles, K.A.: Towards insightful algorithm selection for optimisation using meta-learning concepts. In: IJCNN (2008)
Telelis, O., Stamatopoulos, P.: Combinatorial optimization through statistical instance-based learning. In: ICTAI (2001)
Xu, L., Hutter, F., Shen, J., Hoos, H., Leyton-Brown, K.: SATzilla2012: improved algorithm selection based on cost-sensitive classification models. In: SAT Challenge, Solver description (2012)
Xu, L., Hutter, F., Hoos, H.H., Leyton-Brown, K.: SATzilla-07: The design and analysis of an algorithm portfolio for SAT. In: Bessière, C. (ed.) CP 2007. LNCS, vol. 4741, pp. 712–727. Springer, Heidelberg (2007)
Xu, L., Hutter, F., Hoos, H.H., Leyton-brown, K.: Hydra-MIP: automated algorithm configuration and selection for mixed integer programming. In: RCRA workshop on Experimental Evaluation of Algorithms for Solving Problems with Combinatorial Explosion (2011)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Amadini, R., Gabbrielli, M., Mauro, J. (2014). Portfolio Approaches for Constraint Optimization Problems. In: Pardalos, P., Resende, M., Vogiatzis, C., Walteros, J. (eds) Learning and Intelligent Optimization. LION 2014. Lecture Notes in Computer Science(), vol 8426. Springer, Cham. https://doi.org/10.1007/978-3-319-09584-4_3
Download citation
DOI: https://doi.org/10.1007/978-3-319-09584-4_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-09583-7
Online ISBN: 978-3-319-09584-4
eBook Packages: Computer ScienceComputer Science (R0)