Online Learning of Partitions in Additively Separable Hedonic Games

Online Learning of Partitions in Additively Separable Hedonic Games

Saar Cohen, Noa Agmon

Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence
Main Track. Pages 2722-2730. https://doi.org/10.24963/ijcai.2024/301

Coalition formation involves partitioning agents into disjoint coalitions based on their preferences over other agents. In reality, agents may lack enough information to assess their preferences before interacting with others. This motivates us to initiate the research on coalition formation from the viewpoint of online learning. At each round, a possibly different subset of a given set of agents arrives, that a learner then partitions into coalitions. Only afterwards, the agents' preferences, which possibly change over time, are revealed. The learner's goal is optimizing social cost by minimizing his (static or dynamic) regret. We show that even no-static regret is hard to approximate, and constant approximation in polynomial time is unattainable. Yet, for a fractional relaxation of our problem, we devise an algorithm that simultaneously gives the optimal static and dynamic regret. We then present a rounding scheme with an optimal dynamic regret, which converts our algorithm's output into a solution for our original problem.
Keywords:
Game Theory and Economic Paradigms: GTEP: Cooperative games
Game Theory and Economic Paradigms: GTEP: Computational social choice
Machine Learning: ML: Online learning
Machine Learning: ML: Optimization