Singular limits for models of selection and mutations with heavy-tailed mutation distribution (Q2286495)

This article presents the model of selection-mutation \[ \begin{cases} \partial_tn+(-\Delta)^\alpha n=n R(x,I) \\ n(x,0)=n^0(x), \quad x\in\mathbb{R}^d \end{cases} \] where \[ I(t)=\int_{\mathbb{R}^d}n(t,x)dx \] and hence the reaction term $R$ is nonlocal. The term $(-\Delta)^\alpha$ denotes the fractional Laplacian \[ (-\Delta)^\alpha n(t,x)=\int_{h\in\mathbb{R}^d}\frac{n(t,x)-n(t,x+h)}{|h|^{d+2\alpha}} dh. \] The author undertakes a (biologically relevant) rescaling of the equations with the view of examining a concentration phenomenon with an extended Hamilton-Jacobi approach that was previously developed for mutation kernels possessing fat-tails. These mutation kernels arise after the relevant rescaling. The asymptotic analysis on such evolutionary biological models, and the subsequent concentration phenomena, entails using mutation kernels with small variance (the tails are algebraic but with a large power). In the setting presented here, the ``WKB transformation of the solution does not converge to a viscosity solution of a Hamilton-Jacobi equation, but to a viscosity super-solution of such equation which is minimal in a certain class of super-solutions.'' The concentration of the population density as an evolving Dirac mass is established under a monotony condition on the growth rate. The results here generalize some of those in [\textit{E. Bouin} et al., SIAM J. Math. Anal. 50, No. 3, 3365--3394 (2018; Zbl 1403.45001)].