Computation of nonlinear functionals in particle methods

From MaRDI portal





Let \(f\) be a one-dimensional real density which is Lipschitz continuous and positive on \(B_R= \{x\in \mathbb{R}_1\): \(|x|\leq R\}\), \(R>0\). The author is interested in the estimation of functionals of the form \[ \int \varphi (f(v)) dv \] where \(\varphi: \mathbb{R}^+\to \mathbb{R}\) is continuous and \(\varphi (0) =0\). In section 2 an error estimate for the approximation of these functionals using a coarse graining of the particle space is given. In Section 3, the author assumes moreover that its density can be represented by a finite series of Hermite functions and shows the improvement in speed of convergence compared to the coarse graining method. Finally, in Section 4, numerical examples are given and the entropy is computed for different types of distribution functions occuring in the kinetic theory (Maxwell, Navier-Stokes and Mott-Smith distributions). Moreover, the time development of the entropy for the BKW solution of the homogeneous Boltzmann equation is shown.











This page was built for publication: Computation of nonlinear functionals in particle methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1900693)