No label defined (Q4358716)

This is an article on numerical schemes for solving linear and nonlinear hyperbolic conservation laws in one and two dimensions. It is a survey article in the sense that it shows different ideas and techniques. However it is not meant to be an exhaustive review of the literature. First scalar, linear advection problems in 1 and 2 space dimensions are considered. The framework, put forward by the author, for a large group of schemes in one space dimension is to reconstruct a piecewise parabolic solution from a piecewise linear one, transport this by evolution and project the result back to the space of piecewise linear functions. For a survey of these methods one is referred to \textit{K. W. Morton} [Numerical solution of convection-diffusion problems (1996; Zbl 0861.65070)]. The finite volume approach is also mentioned. It is argued that the schemes with an odd order of accuracy on uniform mesh are to be preferred. For convection in two space dimensions a semi-Lagrangian method and the Lagrangian-Galerkin method as applied to the incompressible Navier-Stokes equation is explained. For the scalar conservation law in two dimension the finite volume schemes are discussed. Some basic facts on the Euler equations of gas dynamics in one and two space dimensions for unsteady and steady flows are given. In a more detailed section on evolutionary hyperbolic problems first characteristic based formulas for the wave equation in two space dimensions are derived. One can write the solution at the tip of the characteristic cone as integrals of the solution at the base and along the surface of the cone. To obtain a numerical scheme the surface integral is approximated. These formulas had been exploited by \textit{D. S. Butler} [Proc. Roy. Soc. Lond., Ser. A 255, 232-252 (1960; Zbl 0099.41501)] and more recently by \textit{S. Ostkamp} [Multidimensional characteristic Galerkin schemes and evolution operators for hyperbolic systems (1995; Zbl 0831.76067)]. It is indicated how this approach extends to general systems and in particular to the Euler equation and how it relates to the method of \textit{M. Fey} [Ein echt mehrdimensionales Verfahren zur Lösung der Eulergleichungen, Dissertation, ETH Zürich (1993)]. A whole section is devoted to steady problems and the relation of the schemes to unsteady problems. In particular fluctuation splitting methods in one and two space dimensions and the cell vertex finite volume methods are described. A short section is devoted to fluctuation splitting for elliptic/hyperbolic problems where the decomposition is done accordingly. For the entire collection see [Zbl 0869.00046].