Research of Decision-Making in the Multi-Agent System Based on Interactive Influence Diagrams

Abstract:

Interactive influence diagrams(I-IDs) offer a transparent and representation for the decision-making in multiagent settings. In I-IDs, for the sake of predicting the behavior of other agent accurately, the modeling agent starts from an initial set of possible models for another agent and then maintains belief about which of those models applies. This initial set of models in the model node is typically a fully specification of possible agent types. Although such a rich space gives the modeling agent high accuracy in its beliefs, it will also incur high cost in maintaining those beliefs. In this paper, we demonstrate that we can choose a minimal, but sufficient, space of mental models by combining models that action or utility equivalence. We illustrate our framework using the two-tiger game and provide empirical results by showing the minimal mental model spaces that it generates.

• Periodical: Key Engineering Materials (Volumes 467-469)
• Pages: 1947-1952
• DOI: https://doi.org/10.4028/www.scientific.net/KEM.467-469.1947
• Online since: February 2011
• Authors: Bo Li, Jian Luo, Jin Fa Zhuang
• Keywords: Action Equivalence, Interactive Influence Diagrams, Minimal Mental Model Sets, Multi-Agent Decision-Making, Utility Equivalence

References

[1] G. Kaminka, D. V. Pynadath and M. Tambe: Monitoring teams by overhearing: A multi-agent planrecognition approach. Journal of Artificial Intelligence Research 17: 83–135(2002).

DOI: 10.1613/jair.970

[2] D. Suryadi and P. Gmytrasiewicz: Learning models of other agents using influence diagrams. In Proceedings of the Seventh International Conference on User Modeling (UM), pages 223–232, (1999).

DOI: 10.1007/978-3-7091-2490-1_22

[3] D. Koller and B. Milch: Multi-agent influence diagrams for representing and solving games. In Proceedings of International Joint Conference on Artificial Intelligence (IJCAI), pages 1027–1034, ( 2001).

[4] Y. Gal and A. Pfeffer: Networks of influence diagrams: A formalism for reasoning about agents' decision-making processes. Journal of Artificial Intelligence Research (JAIR), 33: 109–147, (2008).

DOI: 10.1613/jair.2503

[5] P. Doshi, Y. Zeng, and Q. Chen: Graphical models for interactive pomds: Reresentations and solutions. Journal of Autonomous Agents and Multi-Agents Systems(JAAMAS), 18(3): 376-416, (2009).

DOI: 10.1007/s10458-008-9064-7

[6] M. Kumaran, P. Doshi and Y. F. Zeng: Approximate Solutions of Interactive Dynamic Influence Diagrams Using epsilon-Behavioral Equivalence. In International Symposium on Artificial Intelligence and Mathematics, Ft. Lauderdale, Florida. (2010).

[7] J. A. Tatman and R. D. Shachter: Dynamic programming and influence diagrams. IEEE Transaction on Systems, Man, and Cybernetics, 20(2): 365–379, (1990).

DOI: 10.1109/21.52548

[8] Y. F. Zeng and P. Doshi: Model Identification in Interactive Influence Diagrams Using Mutual Information, Web Intelligence and Agent Systems (WIAS), Volume 8, Number 3, 313-327. (2010).

DOI: 10.3233/wia-2010-0194