Abstract
Building on the recent success of bet-and-run approaches for restarted local search solvers, we introduce the idea of learning online adaptive restart strategies. Universal restart strategies deploy a fixed schedule that runs with utter disregard of the characteristics that each individual run exhibits. Whether a run looks promising or abysmal, it gets run exactly until the predetermined limit is reached. Bet-and-run strategies are at least slightly less ignorant as they decide which trial to use for a long run based on the performance achieved so far. We introduce the idea of learning fully adaptive restart strategies for black-box solvers, whereby the learning is performed by a parameter tuner. Numerical results show that adaptive strategies can be learned effectively and that these significantly outperform bet-and-run strategies.
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.
We say black-box because we do not need to know anything about the inner workings of the solver. However, we make two assumptions. First, that we can set a time limit to the solver where it stops, and that we can add more time and continue the interrupted computation later. Second, that whenever the solver stops it returns information when it found the first solution, when it found the best solution so far, and what the quality of the best solution found so far is.
- 2.
This best possible solution is the best solution provided within the given time limit by any of the 10,000 runs we conducted.
References
Gomes, C.P., Selman, B., Crato, N., Kautz, H.A.: Heavy-tailed phenomena in satisfiability and constraint satisfaction problems. J. Autom. Reason. 24(1), 67–100 (2000)
Hoos, H.H.: Stochastic local search - methods, models, applications. Ph.D. thesis, TU Darmstadt (1998)
Luby, M., Sinclair, A., Zuckerman, D.: Optimal speedup of Las Vegas algorithms. Inf. Process. Lett. 47(4), 173–180 (1993)
Wu, H., van Beek, P.: On universal restart strategies for backtracking search. In: Bessière, C. (ed.) CP 2007. LNCS, vol. 4741, pp. 681–695. Springer, Heidelberg (2007). doi:10.1007/978-3-540-74970-7_48
Audemard, G., Simon, L.: Refining restarts strategies for SAT and UNSAT. In: Milano, M. (ed.) CP 2012. LNCS, vol. 7514, pp. 118–126. Springer, Heidelberg (2012). doi:10.1007/978-3-642-33558-7_11
Marti, R.: Multi-start methods. In: Glover, F., Kochenberger, G.A. (eds.) Handbook of Metaheuristics, pp. 355–368 (2003)
Lourenço, H.R., Martin, O.C., Stützle, T.: Iterated local search: framework and applications. In: Gendreau, M., Potvin, J.Y. (eds.) Handbook of Metaheuristics. International Series in Operations Research & Management Science, vol. 146, pp. 363–397. Springer, Boston (2010). doi:10.1007/978-1-4419-1665-5_12
Schoenauer, M., Teytaud, F., Teytaud, O.: A rigorous runtime analysis for quasi-random restarts and decreasing stepsize. In: Hao, J.-K., Legrand, P., Collet, P., Monmarché, N., Lutton, E., Schoenauer, M. (eds.) EA 2011. LNCS, vol. 7401, pp. 37–48. Springer, Heidelberg (2012). doi:10.1007/978-3-642-35533-2_4
Biere, A.: Adaptive restart strategies for conflict driven SAT solvers. In: Kleine Büning, H., Zhao, X. (eds.) SAT 2008. LNCS, vol. 4996, pp. 28–33. Springer, Heidelberg (2008). doi:10.1007/978-3-540-79719-7_4
Huang, J.: The effect of restarts on the efficiency of clause learning. In: International Joint Conference on Artificial Intelligence (IJCAI), pp. 2318–2323 (2007)
Lalla-Ruiz, E., Voß, S.: Improving solver performance through redundancy. Syst. Sci. Syst. Eng. 25(3), 303–325 (2016)
Fischetti, M., Monaci, M.: Exploiting erraticism in search. Oper. Res. 62(1), 114–122 (2014)
Friedrich, T., Kötzing, T., Wagner, M.: A generic bet-and-run strategy for speeding up stochastic local search. In: Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence, pp. 801–807 (2017)
Lissovoi, A., Sudholt, D., Wagner, M., Zarges, C.: Theoretical results on bet-and-run as an initialisation strategy. In: Genetic and Evolutionary Computation Conference (GECCO) (2017, accepted for publication)
Stützle, T., López-Ibáñez, M.: Automatic (offline) configuration of algorithms. In: Genetic and Evolutionary Computation Conference (GECCO), pp. 795–818 (2016)
Bezerra, L.C.T., López-Ibáñez, M., Stützle, T.: Automatic component-wise design of multiobjective evolutionary algorithms. IEEE Trans. Evol. Comput. 20(3), 403–417 (2016)
Ansótegui, C., Pon, J., Tierney, K., Sellmann., M.: Reactive dialectic search portfolios for MaxSAT. In: AAAI Conference on Artificial Intelligence (2017, accepted for publication)
Kadioglu, S., Sellmann, M.: Dialectic search. In: Gent, I.P. (ed.) CP 2009. LNCS, vol. 5732, pp. 486–500. Springer, Heidelberg (2009). doi:10.1007/978-3-642-04244-7_39
Ansótegui, C., Sellmann, M., Tierney, K.: A gender-based genetic algorithm for the automatic configuration of algorithms. In: Gent, I.P. (ed.) CP 2009. LNCS, vol. 5732, pp. 142–157. Springer, Heidelberg (2009). doi:10.1007/978-3-642-04244-7_14
Hartigan, J.A.: Bounding the maximum of dependent random variables. Electron. J. Stat. 8(2), 3126–3140 (2014)
Polacek, M., Doerner, K.F., Hartl, R.F., Kiechle, G., Reimann, M.: Scheduling periodic customer visits for a traveling salesperson. Eur. J. Oper. Res. 179, 823–837 (2007)
Applegate, D.L., Bixby, R.E., Chvatal, V., Cook, W.J.: The Traveling Salesman Problem: A Computational Study. Princeton University Press, Princeton (2011)
Applegate, D.L., Cook, W.J., Rohe, A.: Chained Lin-Kernighan for large traveling salesman problems. INFORMS J. Comput. 15(1), 82–92 (2003)
Cook, W.: The Traveling Salesperson Problem: Downloads (Website) (2003). http://www.math.uwaterloo.ca/tsp/concorde/downloads/downloads.htm. Accessed 21 Dec 2016
Reinelt, G.: TSPLIB - A traveling salesman problem library. ORSA J. Comput. 3(4), 376–384 (1991). Instances: http://comopt.ifi.uni-heidelberg.de/software/TSPLIB95/tsp/. Accessed 21 Dec 2016
Gomes, F.C., Meneses, C.N., Pardalos, P.M., Viana, G.V.R.: Experimental analysis of approximation algorithms for the vertex cover and set covering problems. Comput. Oper. Res. 33(12), 3520–3534 (2006)
Abu-Khzam, F.N., Langston, M.A., Shanbhag, P., Symons, C.T.: Scalable parallel algorithms for FPT problems. Algorithmica 45(3), 269–284 (2006)
Cai, S.: Balance between complexity and quality: local search for minimum vertex cover in massive graphs. In: International Joint Conference on Artificial Intelligence (IJCAI), pp. 747–753 (2015). Code: http://lcs.ios.ac.cn/caisw/MVC.html. Accessed 21 Dec 2016
Cai, S., Su, K., Luo, C., Sattar, A.: NuMVC: an efficient local search algorithm for minimum vertex cover. J. Artif. Intell. Res. 46(1), 687–716 (2013)
Ansótegui, C., Malitsky, Y., Samulowitz, H., Sellmann, M., Tierney, K.: Model-based genetic algorithms for algorithm configuration. In: International Joint Conference on Artificial Intelligence (IJCAI), pp. 733–739 (2015)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
Kadioglu, S., Sellmann, M., Wagner, M. (2017). Learning a Reactive Restart Strategy to Improve Stochastic Search. In: Battiti, R., Kvasov, D., Sergeyev, Y. (eds) Learning and Intelligent Optimization. LION 2017. Lecture Notes in Computer Science(), vol 10556. Springer, Cham. https://doi.org/10.1007/978-3-319-69404-7_8
Download citation
DOI: https://doi.org/10.1007/978-3-319-69404-7_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-69403-0
Online ISBN: 978-3-319-69404-7
eBook Packages: Computer ScienceComputer Science (R0)