Abstract
With the rapid development of artificial intelligence in recent years, applying various learning techniques to solve mixed-integer linear programming (MILP) problems has emerged as a burgeoning research domain. Apart from constructing end-to-end models directly, integrating learning approaches with some modules in the traditional methods for solving MILPs is also a promising direction. The cutting plane method is one of the fundamental algorithms used in modern MILP solvers, and the selection of appropriate cuts from the candidate cuts subset is crucial for enhancing efficiency. Due to the reliance on expert knowledge and problem-specific heuristics, classical cut selection methods are not always transferable and often limit the scalability and generalizability of the cutting plane method. To provide a more efficient and generalizable strategy, we propose a reinforcement learning (RL) framework to enhance cut selection in the solving process of MILPs. Firstly, we design feature vectors to incorporate the inherent properties of MILP and computational information from the solver and represent MILP instances as bipartite graphs. Secondly, we choose the weighted metrics to approximate the proximity of feasible solutions to the convex hull and utilize the learning method to determine the weights assigned to each metric. Thirdly, a graph convolutional neural network is adopted with a self-attention mechanism to predict the value of weighting factors. Finally, we transform the cut selection process into a Markov decision process and utilize RL method to train the model. Extensive experiments are conducted based on a leading open-source MILP solver SCIP. Results on both general and specific datasets validate the effectiveness and efficiency of our proposed approach.
Article PDF
Similar content being viewed by others
Avoid common mistakes on your manuscript.
References
Achterberg T. Constraint integer programming. PhD Thesis. Berlin: Technische Universität Berlin, 2007
Achterberg T, Wunderling R. Mixed integer programming: Analyzing 12 years of progress. In: Jünger M, Reinelt G, eds. Facets of Combinatorial Optimization: Festschrift for Martin Grötschel. Berlin-Heidelberg: Springer, 2013, 449–481
Alvarez A M, Louveaux Q, Wehenkel L. A machine learning-based approximation of strong branching. INFORMS J Comput, 2017, 29: 185–195
Anderson L, Turner M, Koch T. Generative deep learning for decision making in gas networks. Math Methods Oper Res, 2022, 95: 503–532
Andreello G, Caprara A, Fischetti M. Embedding {0, 1/2}-cuts in a branch-and-cut framework: A computational study. INFORMS J Comput, 2007, 19: 229–238
Balcan M F, Dick T, Sandholm T, et al. Learning to branch. In: Proceedings of International Conference on Machine Learning. Ann Arbor: PMLR, 2018, 344–353
Balcan M F, Prasad S, Sandholm T, et al. Sample complexity of tree search configuration: Cutting planes and beyond. In: Proceedings of the 35th International Conference on Neural Information Processing Systems. San Francisco: Curran Associates, 2021, 4015–4027
Baltean-Lugojan R, Bonami P, Misener R, et al. Scoring positive semidefinite cutting planes for quadratic optimization via trained neural networks. https://optimization-online.org/2018/11/6943/, 2019
Berthold T, Francobaldi M, Hendel G. Learning to use local cuts. ar**v:2206.11618, 2022
Bestuzheva K, Besançon M, Chen W K, et al. The SCIP Optimization Suite 8.0. Berlin: Zuse Institute Berlin, 2021, 21–41
Blieklú C, Bonami P, Lodi A. Solving mixed-integer quadratic programming problems with IBM-CPLEX: A progress report. In: Proceedings of the twenty-sixth RAMP Symposium. Montreal: ACM Press, 2014, 16–17
Bonami P, Cornuéjols G, Dash S, et al. Projected Chvatal-Gomory cuts for mixed integer linear programs. Math Program, 2008, 113: 241–257
Chen W K, Chen L, Dai Y H. Lifting for the integer knapsack cover polyhedron. J Global Optim, 2022: 1–45
Chvátal V. Edmonds polytopes and a hierarchy of combinatorial problems. Discrete Math, 1973, 4: 305–337
Dey S S, Molinaro M. Theoretical challenges towards cutting-plane selection. Math Program, 2018, 170: 237–266
Ding J Y, Zhang C, Shen L, et al. Accelerating primal solution findings for mixed integer programs based on solution prediction. AAAI J, 2020, 34: 1452–1459
Gasse M, Chételat D, Ferroni N, et al. Exact combinatorial optimization with graph convolutional neural networks. In: Proceedings of the 32nd International Conference on Neural Information Processing Systems. San Francisco: Curran Associates, 2019, 15580–15592
Ge D, Huangfu Q, Wang Z, et al. Cardinal Optimizer (COPT) User Guide. ar**v:2208.14314, 2022
Gleixner A, Hendel G, Gamrath G, et al. MIPLIB 2017: Data-driven compilation of the 6th mixed-integer programming library. Math Program Comput, 2021, 13: 443–490
Gomory R. An algorithm for the mixed integer problem. RAND CORP SANTA MONICA CA, 1960
Gupta P, Gasse M, Khalil E, et al. Hybrid models for learning to branch. In: Proceedings of the 33rd International Conference on Neural Information Processing Systems. San Francisco: Curran Associates, 2020, 18087–18097
Gurobi Optimization L. Gurobi Optimizer Reference Manual. https://www.gurobi.com, 2023
Huang Z, Wang K, Liu F, et al. Learning to select cuts for efficient mixed-integer programming. Pattern Recogn, 2022, 123: 108353
Jia H, Shen S. Benders cut classification via support vector machines for solving two-stage stochastic programs. INFORMS J Optim, 2021, 3: 278–297
Kallrath J. Modeling Languages in Mathematical Optimization. New York: Springer, 2004
Khalil E, Dilkina B, Nemhauser GL, et al. Learning to run heuristics in tree search. In: Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence. IJCAI, 2017, 659–666
Khalil E, Le Bodic P, Song L, et al. Learning to branch in mixed integer programming. AAAI J, 2016, 30: 724–731
Kingma DP, Ba J. Adam: A method for stochastic optimization. ar**v:1412.6980, 2014
Kool W, van Hoof H, Welling M. Attention, learn to solve routing problems! In: Proceedings of the International Conference on Learning Representations. ICLR, 2019, 3: 5057–5069
Land AH, Doig AG. An automatic method of solving discrete programming problems. Econometrica, 1960, 28: 497–520
Lodi A, Tramontani A. Performance variability in mixed-integer programming. In: Theory Driven by Influential Applications. Annapolis: INFORMS, 2013, 1–12
Marchand H, Martin A, Weismantel R, et al. Cutting planes in integer and mixed integer programming. Discrete Appl Math, 2002, 123: 397–446
Nair V, Bartunov S, Gimeno F, et al. Solving mixed integer programs using neural networks. ar**v:2012.13349, 2020
Nazari M, Oroojlooy A, Snyder L, et al. Reinforcement learning for solving the vehicle routing problem. In: Proceedings of the International Conference on Neural Information Processing Systems. San Francisco: Curran Associates, 2018, 94–99
Nickel S, Steinhardt C, Schlenker H, et al. IBM ILOG CPLEX Optimization Studio: Modellierung von Planungs-und Entscheidungsproblemen des Operations Research mit OPL, 2021, 9–23
Paszke A, Gross S, Massa F, et al. Pytorch: An imperative style, high-performance deep learning library. In: Proceedings of the 33rd International Conference on Neural Information Processing Systems. San Francisco: Curran Associates, 2019, 8026–8037
Paulus MB, Zarpellon G, Krause A, et al. Learning to cut by looking ahead: Cutting plane selection via imitation learning. In: Proceedings of International Conference on Machine Learning. Ann Arbor: PMLR, 2022, 17584–17600
Shen Y, Sun Y, Eberhard A, et al. Learning primal heuristics for mixed integer programs. In: Proceedings of the International Joint Conference on Neural Networks. San Francisco: IEEE, 2021, 1–8
Sutton R S, McAllester D, Singh S, et al. Policy gradient methods for reinforcement learning with function approximation. Adv Neural Inform Process Syst, 2000, 12: 1057–1063
Tang Y, Agrawal S, Faenza Y. Reinforcement learning for integer programming: Learning to cut. In: Proceedings of International Conference on Machine Learning. Ann Arbor: PMLR, 2020, 9367–9376
Turner M, Koch T, Serrano F, et al. Adaptive cut selection in mixed-integer linear programming. ar**v:2202.10962, 2022
Vasilyev I, Boccia M, Hanafi S. An implementation of exact knapsack separation. J Global Optim, 2016, 66: 127–150
Vinyals O, Fortunato M, Jaitly N. Pointer networks. In: Proceedings of the 28th International Conference on Neural Information Processing Systems. San Francisco: Curran Associates, 2015, 2692–2700
Wang Z, Li X, Wang J, et al. Learning cut selection for mixed-integer linear programming via hierarchical sequence model. ar**v:2302.00244, 2023
Wesselmann F, Stuhl U. Implementing cutting plane management and selection techniques. Technical Report. Paderborn: University of Paderborn, 2012
Wolsey LA. Integer Programming. New York: John Wiley & Sons, 2020
Yilmaz K, Yorke-Smith N. Learning efficient search approximation in mixed integer branch and bound. http://www.resolvertudelftnl/uuid:bce72457-108f-4215-b7e4-599866ba52aa, 2020
Acknowledgements
This work was supported by the National Key R&D Program of China (Grant No. 2022YFB2403400) and National Natural Science Foundation of China (Grant Nos. 11991021 and 12021001).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Wang, S., Chen, L., Niu, L. et al. Enhancing cut selection through reinforcement learning. Sci. China Math. 67, 1377–1394 (2024). https://doi.org/10.1007/s11425-023-2294-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11425-023-2294-3