Concentration in Lotka-Volterra parabolic equations: an asymptotic-preserving scheme (Q6165229)

In this paper, an asymptotic-preserving scheme for parabolic Lotka-Volterra equations \[ \left\{ \begin{array}{l l} \partial_t n_\epsilon(t,x) - \epsilon \Delta_x n_\epsilon(t,x) = \frac{n_\epsilon(t,x)}{\epsilon} R(x,I_\epsilon(t)), & x\in \mathbb{R}^d, \,t\ge 0, \\ I_\epsilon(t)=\int_{\mathbb{R}^d} \psi(x) n_\epsilon(t,x) dx, & t\ge 0, \end{array} \right. \] is developed, which model the evolution of a population density. The proposed scheme is converging for fixed \(\epsilon > 0\) and has favorable stability properties in the asymptotics. The limiting scheme for \(\epsilon =0\) is converging towards the unique viscosity solution of the constrained Hamilton-Jacobi equation, which describes the asymptotic regime. The key ingredients for the construction of the asymptotic-preserving scheme are the monotonicity and the implicit treatment of the constraint. Using these properties, the convergence of the scheme is proved, and so is its asymptotic behavior in the vanishing \(\epsilon\) limit. The convergence of the limit scheme is based on compactness arguments, and on the monotonicity of the scheme. The theoretical analysis of the schemes is illustrated with numerical simulations