Abstract
Max-sum is a version of Belief propagation, used for solving DCOPs. On tree-structured problems, Max-sum converges to the optimal solution in linear time. Unfortunately when the constraint graph representing the problem includes multiple cycles (as in many standard DCOP benchmarks), Max-sum does not converge and explores low quality solutions. Recent attempts to address this limitation proposed versions of Max-sum that guarantee convergence, by changing the constraint graph structure. Damping is a method that is often used for increasing the chances that Belief propagation will converge, however, it was not mentioned in studies that proposed Max-sum for solving DCOPs.
In this paper we investigate the effect of damping on Max-sum. We prove that, while it slows down the propagation of information among agents, on tree-structured graphs, Max-sum with damping is guaranteed to converge to the optimal solution in weakly polynomial time. Our empirical results demonstrate a drastic improvement in the performance of Max-sum, when using damping. However, in contrast to the common assumption, that it performs best when converging, we demonstrate that non converging versions perform efficient exploration, and produce high quality results, when implemented within an anytime framework. On most benchmarks, the best results were achieved using a high damping factor (A preliminary version of this paper was accepted as a two page extended abstract to a coming up conference.)
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
We say that a variable is involved in a constraint if it is one of the variables the constraint refers to.
- 2.
- 3.
t-tests established that the Damped Max-Sum anytime solutions of all values of the parameter \(\lambda \) were better on average than the anytime solutions reported for standard Max-Sum, with statistical significance of p = 0.01, and better on average than DSA’s solutions for \(\lambda \) values of 0.7 and 0.9 (except for the 0.9 version on graph coloring problems), with the same significance level.
References
Dueck, D.: Affinity propagation: clustering data by passing messages. Ph.D. thesis, University of Toronto (2009)
Farinelli, A., Rogers, A., Petcu, A., Jennings, N.R.: Decentralized coordination of low-power embedded devices using the max-sum algorithm. In: AAMAS, pp. 639–646 (2008)
Kiekintveld, C., Yin, Z., Kumar, A., Tambe, M.: Asynchronous algorithms for approximate distributed constraint optimization with quality bounds. In: AAMAS, pp. 133–140 (2010)
Kschischang, F.R., Frey, B.J., Loeliger, H.A.: Factor graphs and the sum-product algorithm. IEEE Trans. Inf. Theor. 47(2), 181–208 (2001)
Lazic, N., Frey, B., Aarabi, P.: Solving the uncapacitated facility location problem using message passing algorithms. In: International Conference on Artificial Intelligence and Statistics, pp. 429–436 (2010)
Verman, A.B.M., Stutz, P.: Solving distributed constraint optimization problems using ranks. In: AAAI Workshop Statistical Relational Artificial Intelligence (2014)
Modi, P.J., Shen, W., Tambe, M., Yokoo, M.: Adopt: asynchronous distributed constraints optimizationwith quality guarantees. Artif. Intell. 161(1–2), 149–180 (2005)
Murphy, K.P., Weiss, Y., Jordan, M.I.: Loopy belief propagation for approximate inference: an empirical study. In: Proceedings of the Fifteenth Conference on Uncertainty in Artificial Intelligence, UAI 1999, Stockholm, Sweden, 30 July–1 August 1999, pp. 467–475 (1999)
Okamoto, S., Zivan, R., Nahon, A.: Distributed breakout: beyond satisfaction. In: Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, IJCAI 2016, New York, NY, USA, 9–15 July 2016, pp. 447–453 (2016)
Som, A.C.P.: Damped belief propagation based near-optimal equalization of severely delay-spread UWB MIMO-ISI channels. In: 2010 IEEE International Conference on Communications (ICC), pp. 1–5 (2010)
Petcu, A., Faltings, B.: A scalable method for multiagent constraint optimization. In: IJCAI, pp. 266–271 (2005)
Pretti, M.: A message-passing algorithm with damping. J. Stat. Mech. Theor. Exp. 11, P11008 (2005)
Rogers, A., Farinelli, A., Stranders, R., Jennings, N.R.: Bounded approximate decentralized coordination via the max-sum algorithm. Artif. Intell. 175(2), 730–759 (2011)
Som, P., Chockalingam, A.: Damped belief propagation based near-optimal equalization of severely delay-spread UWB MIMO-ISI channels. In: 2010 IEEE International Conference on Communications (ICC), pp. 1–5. IEEE (2010)
Sontag, D., Meltzer, T., Globerson, A., Jaakkola, T., Weiss, Y.: Tightening LP relaxations for map using message passing. In: UAI, pp. 503–510 (2008)
Stranders, R., Farinelli, A., Rogers, A., Jennings, N.R.: Decentralised coordination of mobile sensors using the max-sum algorithm. In: Proceedings of the 21st International Joint Conference on Artificial Intelligence, IJCAI 2009, Pasadena, California, USA, 11–17 July 2009, pp. 299–304 (2009)
Tarlow, D., Givoni, I., Zemel, R., Frey, B.: Graph cuts is a max-product algorithm. In: Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence (2011)
Weiss, Y.: Correctness of local probability propagation in graphical models with loops. Neural Comput. 12(1), 1–41 (2000)
Yanover, C., Meltzer, T., Weiss, Y.: Linear programming relaxations and belief propagation - an empirical study. J. Mach. Learn. Res. 7, 1887–1907 (2006)
Zhang, W., Xing, Z., Wang, G., Wittenburg, L.: Distributed stochastic search and distributed breakout: properties, comparishon and applications to constraints optimization problems in sensor networks. Artif. Intell. 161(1–2), 55–88 (2005)
Zivan, R., Okamoto, S., Peled, H.: Explorative anytime local search for distributed constraint optimization. Artif. Intell. 212, 1–26 (2014)
Zivan, R., Peled, H.: Max/min-sum distributed constraint optimization through value propagation on an alternating DAG. In: AAMAS, pp. 265–272 (2012)
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
Cohen, L., Zivan, R. (2017). Max-sum Revisited: The Real Power of Damping. In: Sukthankar, G., Rodriguez-Aguilar, J. (eds) Autonomous Agents and Multiagent Systems. AAMAS 2017. Lecture Notes in Computer Science(), vol 10643. Springer, Cham. https://doi.org/10.1007/978-3-319-71679-4_8
Download citation
DOI: https://doi.org/10.1007/978-3-319-71679-4_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-71678-7
Online ISBN: 978-3-319-71679-4
eBook Packages: Computer ScienceComputer Science (R0)