Title: A probabilistic approach to Dirac concentration in nonlocal models of adaptation with several resources

Abstract: This work is devoted to the study of scaling limits in small mutations and large time of the solutions u^$\epsilon$ of two deterministic models of phenotypic adaptation, where the parameter $\epsilon$ > 0 scales the size of mutations. The first model is the so-called Lotka-Volterra parabolic PDE in R d with an arbitrary number of resources and the second one is an adaptation of the first model to a finite phenotype space. The solutions of such systems typically concentrate as Dirac masses in the limit $\epsilon$ $\rightarrow$ 0. Our main results are, in both cases, the representation of the limits of $\epsilon$ log u^$\epsilon$ as solutions of variational problems and regularity results for these limits. The method mainly relies on Feynman-Kac type representations of u $\epsilon$ and Varadhan's Lemma. Our probabilistic approach applies to multi-resources situations not covered by standard analytical methods and makes the link between variational limit problems and Hamilton-Jacobi equations with irregular Hamiltonians that arise naturally from analytical methods. The finite case presents substantial difficulties since the rate function of the associated large deviation principle has non-compact level sets. In that case, we are also able to obtain uniqueness of the solution of the variational problem and of the associated differential problem which can be interpreted as a Hamilton-Jacobi equation in finite state space.