Abstract
Rationality and convergence are two important criterions for multi-agent learning. A novel method called Pareto-Q learning is prompted for cooperative general-sum games, with the Pareto Optimum allowing rationality and social conventions benefiting the convergence. Experiments with the grid game suggest the efficiency of Pareto-Q. Compared with the single-agent Q-learning and Nash agent Q-learning, Pareto-Q learning performs best.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Bowling, M., Veloso, M.: Existence of Multiagent Equilibria with Limited Agents. Journal of Artificial Intelligence Research 22, 353–384 (2004)
Hu, J., Wellman, M.P.: Nash Q-Learning for General-sum Stochastic Games. Journal of Machine Learning Research 4, 1039–1069 (2003)
Littman, M.L., Szepesvari, C.: A Generalized Reinforcement Learning Model: Convergence and Applications. In: Proceedings of the 13th International Conference on Machine Learning, Bari, Italy, pp. 310–318 (1996)
Deb, K.: Multi-Objective Evolutionary Algorithms: Introducing Bias among Pareto-Optimal Solutions. KanGAL report 99002, Indian Institute of Technology, Kanpur, India (1999)
Kok, J.R., Spaan, M.T.J., Vlassis, N.: An Approach to Noncommunicative Multiagent Coordination in Continuous Domains. In: Proceedings of the Twelfth Belgian-Dutch Conference on Machine Learning, Utrecht, Netherlands, pp. 46–52 (2002)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2005 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Song, M., Gu, G., Zhang, G. (2005). Pareto-Q Learning Algorithm for Cooperative Agents in General-Sum Games. In: Pěchouček, M., Petta, P., Varga, L.Z. (eds) Multi-Agent Systems and Applications IV. CEEMAS 2005. Lecture Notes in Computer Science(), vol 3690. Springer, Berlin, Heidelberg. https://doi.org/10.1007/11559221_64
Download citation
DOI: https://doi.org/10.1007/11559221_64
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-29046-9
Online ISBN: 978-3-540-31731-9
eBook Packages: Computer ScienceComputer Science (R0)