@inproceedings { santos16, abstract = {There are numerous human decisions and social preferences whose features are not easy to grasp mathematically. Fairness is certainly one of the most pressing. In this paper, we study a multiplayer extension of the well-known Ultimatum Game through the lens of a reinforcement learning algorithm. This game allows us to study fair behaviors beyond the traditional pairwise interaction models. Here, a proposal is made to a quorum of Responders, and the overall acceptance depends on reaching a threshold of individual acceptances. We show that, while considerations regarding the sub-game perfect equilibrium of the game remain untouched, learning agents coordinate their behavior into different strategies, depending on factors such as the group acceptance threshold, the group size or disagreement costs. Overall, our simulations show that stringent group criteria trigger fairer proposals and the effect of group size on fairness depends on the same group acceptance criteria. Fairness can be boosted by the imposition of disagreement costs on the Proposer side.}, booktitle = {Proceedings of Adaptive and Learning Agents Workshop (ALA 2016), part of 2016 Int. Conf. Autonomous Agents and Multiagent Systems (AAMAS 2016)}, keywords = {Multi-Agent Societies;Game Theory;Miscellaneous;Reinforcement Learning;}, publisher = {(Best paper award)}, title = {Multiplayer ultimatum game in populations of autonomous agents}, year = {2016}, author = {Fernando P. Santos and Francisco C. Santos and Francisco S. Melo and Ana Paiva and Jorge M. Pacheco} }

hide this content