This work presents a model for Markov Decision Processes applied to the problem of keeping two agents in equilibrium with respect to the values they exchange when they interact. Interval mathematics is used to model the qualitative values involved in interactions. The optimal policy is constrained by the adopted model of social interactions. The MDP is assigned to a supervisor, that monitors the agents' actions and makes recommendations to keep them in equilibrium. The agents are autonomous and allowed to not follow the recommendations. Due to the qualitative nature of the exchange values, even when agents follow the recommendations, the decision process is non-trivial. © Springer-Verlag Berlin Heidelberg 2006.
CITATION STYLE
Dimuro, G. P., & Costa, A. C. R. (2006). Interval-based markov decision processes for regulating interactions between two agents in multi-agent systems. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3732 LNCS, pp. 102–111). https://doi.org/10.1007/11558958_12
Mendeley helps you to discover research relevant for your work.