Abstract
Branch-and-bound is the workhorse of all state-of-the-art mixed integer linear programming (MILP) solvers. These implementations of branch-and-bound typically use variable branching, that is, the child nodes are obtained via disjunctions of the form \(x_j \le \lfloor \bar{x}_j \rfloor \vee x_j \ge \lceil \bar{x}_j \rceil \), where \(\bar{x}\) is an optimal solution to the LP corresponding to the parent node. Even though modern MILP solvers are able to solve very large-scale instances efficiently, relatively little attention has been given to understanding why the underlying branch-and-bound algorithm performs so well. In this paper, our goal is to theoretically analyze the performance of the standard variable branching based branch-and-bound algorithm. In order to avoid the exponential worst-case lower bounds, we follow the common idea of considering random instances. More precisely, we consider random integer programs where the entries of the coefficient matrix and the objective function are randomly sampled. Our main result is that with good probability branch-and-bound with variable branching explores only a polynomial number of nodes to solve these instances, for a fixed number of constraints. To the best of our knowledge this is the first known such result for a standard version of branch-and-bound. We believe that this result provides an indication as to why branch-and-bound with variable branching works so well in practice.
Similar content being viewed by others
Notes
We assume throughout that the \(C/2^\ell \)’s are integral to simplify the notation, but it can be easily checked that using \(\lceil C/2^\ell \rceil \) instead does not change the results.
References
Santanu, S.D., Yatharth, D., Marco, M.: Branch-and-bound solves random binary ips in polytime. In: Proceedings of the 2021 ACM-SIAM Symposium on Discrete Algorithms (SODA), pp. 579–591. SIAM (2021)
Alisa, H.L., Alison, G.D.: An automatic method of solving discrete programming problems. Econometrica 28, 497–520 (1960)
George, L.: Integer and Combinatorial Optimization. In: Nemhauser and laurence a wolsey, vol. 55. Wiley, New Jersey (1999)
Michele, C., Gérard, C., Giacomo, Z., et al.: Integer Programming, vol. 271. Springer, New York (2014)
Hendrik, W.L., Jr.: Integer programming with a fixed number of variables. Math. Oper. Res. 8(4), 538–548 (1983)
Hendrik, W.L., Arjen, K.L., László, L., et al.: Factoring polynomials with rational coeficients. Math. Ann. 261, 515–534 (1982)
Gábor, P., Mustafa, T., Erick, B.W. Basis reduction and the complexity of branch-and-bound. In: Proceedings of the Twenty-first Annual ACM-SIAM Symposium on Discrete Algorithms, pp. 1254–1261. SIAM (2010)
Karen, A., Robert, E.B., Cor, A.J.H., Arjen, K.L., Job, W.S.: Market split and basis reduction: towards a solution of the cornuéjols-dawande instances. INFORMS J. Comput. 12(3), 192–202 (2000)
Bixby, R., Rothberg, E.: Progress in computational mixed integer programming-a look back from the other side of the tipping point. Ann. Oper. Res. 149(1), 37 (2007)
Thomas F.C. (ed.).: Chapter 3: Large Sparse Linear Programming, pp. 35–46. Springer, Berlin Heidelberg, 1984. https://doi.org/10.1007/3-540-12914-6_3
Matthias, W.: Sparsity of lift-and-project cutting planes. In: Operations Research Proceedings 2012, pp. 9–14. Springer, New York (2014)
Santanu, S.D., Marco, M., Qianyi, W.: Approximating polyhedra with sparse inequalities. Math. Program. 154(1–2), 329–352 (2015)
Santanu, S.D., Marco, M., Qianyi, W.: Analysis of sparse cutting planes for sparse milps with applications to stochastic milps. Math. Oper. Res. 43(1), 304–332 (2018)
Chvátal, Vasek: Hard knapsack problems. Oper. Res. 28(6), 1402–1411 (1980)
Robert, G.J.: Trivial integer programs unsolvable by branch-and-bound. Math. Program. 6(1), 105–109 (1974)
Sanjeeb, D.: An exponential lower bound on the length of some classes of branch-and-cut proofs. In: International Conference on Integer Programming and Combinatorial Optimization, pp. 145–160. Springer (2002)
Kevin K.H., Cheung, A.M.G., Daniel E.S.: Verifying integer programming results. In: Friedrich, E., Jochen, K. (eds.) Integer Programming and Combinatorial Optimization - 19th International Conference, IPCO 2017, Waterloo, ON, Canada, June 26-28, 2017, Proceedings, Volume 10328 of Lecture Notes in Computer Science, pp. 148–160. Springer, 2017. https://doi.org/10.1007/978-3-319-59250-3_13
Daniel, A.S., Shang-Hua, T.: Smoothed analysis of algorithms: why the simplex algorithm usually takes polynomial time. J. ACM (JACM) 51(3), 385–463 (2004)
Jeff, T.L., Martin, W.P.S.: A computational study of search strategies for mixed integer programming. INFORMS J. Comput. 11(2), 173–187 (1999)
Achterberg, T., Koch, T., Martin, A.: Branching rules revisited. Oper. Res. Lett. 33(1), 42–54 (2005)
Lueker, G.S.: On the average difference between the solutions to linear and integer knapsack problems. In: Applied Probability – Computer Science, The Interface, vol. 1. Birkhäuser (1982)
Andrew, V.G., Alberto, M.-S.: On finding the exact solution of a zero-one knapsack problem. In: Richard, A.D. (ed.). STOC, pp. 359–368. ACM (1984). URL http://dblp.uni-trier.de/db/conf/stoc/stoc84.html#GoldbergM84
René, B., Berthold, V.: Random knapsack in expected polynomial time. In: Lawrence L.L., Michel, X.G. (eds.) STOC, pp. 232–241. ACM (2003). URL http://dblp.uni-trier.de/db/conf/stoc/stoc2003.html#BeierV03
René, B., Berthold V.: Probabilistic analysis of knapsack core algorithms. In: Ian Munro, J. (ed.) SODA, pp. 468–477. SIAM (2004). URL http://dblp.uni-trier.de/db/conf/soda/soda2004.html#BeierV04
Dyer, M.E., Frieze, A.M.: Probabilistic analysis of the multidimensional knapsack problem. Math. Oper. Res. 14(1), 162–176 (1989)
Sander, B., Daniel, D., Sophie, H., Samarth, T.: On the integrality gap of binary integer programs with gaussian data. In: IPCO, pp. 427–442 (2021)
Alan, M.F.: On the expected efficiency of branch and bound for the asymmetric tsp. (2020)
Vazirani, V.: Approximation Algorithms. Springer, New Year (2001)
Aharon, B.-T., Arkadi, N.: Lectures on Modern Convex Optimization: Analysis, Algorithms, and Engineering Applications. SIAM (2001)
Roman, V.: High-Dimensional Probability: An Introduction with Applications in Data Science. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, Cambridge (2018). https://doi.org/10.1017/9781108231596
Keith, B.: Cube slicing in \(\mathbb{R} ^n\). Proc. Am. Math. Soc. 97(3), 465–473 (1986)
Koltchinskii, V.: Oracle Inequalities in Empirical Risk Minimization and Sparse Recovery Problems. Springer-Verlag, New York (2011)
Acknowledgements
We are extremely grateful to Daniel Dadush for suggesting several improvements to the paper. Santanu S. Dey would like to gratefully acknowledge the support given by the Airforce office of scientific research, award number FA9550-22-1-0052. Marco Molinaro was supported in part by the Coordenaccão de Aperfeiccoamento de Pessoal de Nível Superior (CAPES, Brasil) - Finance Code 001, by Bolsa de Produtividade em Pesquisa \(\#3\)12751/2021-4 from CNPq, FAPERJ grant “Jovem Cientista do Nosso Estado”, and by the CAPES-PrInt program.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
All authors declare that they have no conflicts of interest.
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Dey, S.S., Dubey, Y. & Molinaro, M. Branch-and-bound solves random binary IPs in poly(n)-time. Math. Program. 200, 569–587 (2023). https://doi.org/10.1007/s10107-022-01895-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10107-022-01895-4