scientific article; zbMATH DE number 3974220

From MaRDI portal
Publication:3740168





[For the entire collection see Zbl 0591.00024.] The optimality of Petrov-Galerkin approximations for the linear diffusion-convection equation in two dimensions is analysed. The main problem with the Petrov-Galerkin method is to select the test space for a given trial space. Since the Riesz representation operators in two dimensions are not available in explicit form, the decomposition of the space into the direct sum of curl-free and divergence-free subspaces is used. This leads to an error estimation with optimal constant for the upwind Petrov-Galerkin approximation. A numerical example for the problem in a rectangular domain is given.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3740168)