A New Convergence Proof for Finite Volume Schemes Using the Kinetic Formulation of Conservation Laws (Q4943640)

Many physical phenomena, as e.g. compressible inviscid fluid flow, can be described by (systems of) nonlinear hyperbolic conservation laws. Unfortunately, analysis and numerics is rather difficult even for multidimensional scalar conservation laws, which are considered by the authors. The method of choice for the numerical approximation is the finite volume method. Their convergence, which has been studied by many authors, relies on the famous results by \textit{S. N. Kružkov} [Math. USSR Sb. 10, 217-243 (1970; Zbl 0215.16203)] who has proved existence and uniqueness of entropy solutions via the vanishing viscosity method. In the present paper, the authors give an alternative convergence proof based upon the kinematic formulation, which has been introduced by \textit{P.-L. Lions}, \textit{B. Perthame} and \textit{E. Tadmor} [J. Am. Math. Soc. 7, 169-191 (1994; Zbl 0820.35049)], a discrete entropy inequality, and the velocity averaging technique. The kinematic formulation enables to associate a solution to a linear transport equation (with a nonlinear constraint) with the weak entropy solution to the original conservation law. By means of a discrete entropy inequality, the entropy dissipation over each cell can be estimated. For proving the strong convergence of a subsequence of approximate solutions in \(L^1_{\text{loc}}\), the velocity averaging technique is needed. The weak\(^*\) convergence in \(L^\infty\) is also employed. The finite volume schemes considered use piecewise constant approximations on an unstructured mesh consisting of convex polygons. The numerical flux functions belong to the class of Osher's E-fluxes (being convex combinations of the well-known Godunov and Lax-Friedrichs fluxes). Besides the new proof, the paper also gives a nice overview of the different convergence approaches known from the literature.