Asymptotic of grazing collisions and particle approximation for the Kac equation without Cutoff (Q694987)

This work is concerned with the Kac equation, which is a one-dimensional version of the spatially homogeneous Boltzmann equation for the non-equilibrium dynamics of particles in a gas. This is a nonlinear integro-differential equation for the density of particles that move with a certain velocity at a given instant of time. The rate at which collisions among particles take place is proportional to the cross-section function. The integral of this function is not assumed to be finite, and in particular this means that the case in which an infinite number of collisions for each particle happens during any time interval is allowed. One of the goals of this article is finding the asymptotic form of the Kac equation in the limit of grazing collisions, consisting in more and more collisions taking place, and at the same time they generate smaller and smaller deviations. This asymptotic form is a Fokker-Planck equation, and, in fact, the authors prove uniform convergence in time in terms of the second Wasserstein distance between the solutions to either equation; also, an explicit rate of convergence is given. Another goal is substituting small collisions by a diffusion term, in the spirit of the previous asymptotic calculation, in order to get an approximation of the Kac equation. The authors prove a convergence rate of this approximation to the actual equation that improves the one obtained in the case of an approximation that simply truncates small collisions. This result should be of special interest for the numerical resolution of the Kac equation, since an infinite number of collisions cannot be directly accomplished this way. Building on these previous developments, the authors are able to find a system of stochastic particles that both diffuse and undergo collisions, and that is able to approximate the solution to the Kac equation under study. Estimates on the rate of convergence are provided in this case as well.