The well-posedness of a multi-population dynamical system with an entropy regularization and its convergence to a suitable mean-field approximation are proved, under a general set of assumptions. Under further assumptions on the evolution of the labels, the case of different time scales between the agents' locations and labels dynamics is considered. The limit system couples a mean-field-type evolution in the space of positions and an instantaneous optimization of the payoff functional in the space of labels.
Keywords: Entropic regularization; Fast reaction limit; Mean-field limit; Population dynamics; Replicator-type dynamics; Superposition principle.
© The Author(s) 2023.