Stationary Mean-Field Games of Singular Control under Knightian Uncertainty

In this work, we study a class of stationary mean-field games of singular stochastic control under model uncertainty. The representative agent adjusts the dynamics of an It\^o diffusion via one-sided singular stochastic control, aiming to maximize a long-term average reward criterion. The mean-field interaction is of scalar type through the stationary distribution of the population. Due to the presence of uncertainty, the problem involves the study of a stochastic zero-sum game, where the decision maker chooses the best singular control policy, while the adversarial player selects the worst probability measure. Using a constructive approach, we prove existence and uniqueness of a stationary mean-field equilibrium. Finally, we provide a stylized numerical benchmark of dirty-capacity reduction under ambiguity and analyze the impact of uncertainty on the mean-field equilibrium.