Equilibrium selection in general-sum games
Determine principled criteria and algorithms for selecting appropriate policies in general-sum partially observable stochastic games, targeting desirable equilibria among potentially many sub-optimal Nash equilibria.
References
It is not clear how to identify appropriate policies for a general-sum game.
— Multi-agent cooperation through learning-aware policy gradients
(Meulemans et al., 24 Oct 2024) in Section 2.1 (General-sum games and their challenges), Equilibrium selection paragraph