Uniqueness and non-uniquess for the mean field control of fisheries

We study a Mean Field Control system arising in the management of fisheries with a special emphasis on non-uniqueness issues.

Namely, we focus on a situation where a group of players coordinate in order to harvest a fishery in the most efficient way possible.

A major challenge in such modelling is the coupling between the dynamics of fish population, which we model through a reaction-diffusion equation, and that of the players, which is seen through the lens of Mean Field Control.

The resulting evolution system consists of four coupled equations.

A central issue, both in the analysis and from the modelling perspective, is the uniqueness of solutions of this system.

By focusing on the ergodic (or static) counterpart of the evolution equation, we show that one should in general expect the emergence of multiple solutions.

Our approach relies on the theory of bifurcation, and the bifurcation parameter we take is the (biologically relevant) total amount of food available to the population.

We also give refined uniqueness criteria that allow to bypass several limitations of previous works on this type of system [39].

This fits within two growing research lines: one on the optimal harvesting of fisheries [40, 39], one on questions of non-uniqueness in Mean Field Games and Mean Field Control [3, 27, 36].