A third Strang lemma and an Aubin-Nitsche trick for schemes in fully discrete formulation (Q1616104): Difference between revisions

The authors present an interesting abstract analysis framework, in the spirit of Strang's second lemma, for approximations of linear partial differential equations (PDEs) in weak form. Contrary to Strang's lemma, the approximations can be written in fully discrete form, with test and trial spaces that are not spaces of functions -- and thus not manipulable together with the continuous test and trial spaces. The framework identifies a general consistency error that bounds, under an inf-sup condition, the discrete norm of the difference between the approximation solution and an interpolant of the exact solution. Some improved estimates in a weaker norm, using the Aubin-Nitsche trick, are established. This framework is applied to an anisotropic heterogeneous diffusion model and two classical families of schemes for this model conforming and non-conforming virtual element and finite volume methods. For each of these methods, the authors derive energy error estimates. Optimal \(L^2\)-error estimates were also proved for virtual element method. A clear notion of consistency for finite volume methods, which leads to a generic error estimate involving the fluxes and valid for a wide range of finite volume schemes is obtained. An important application is the first error estimate for multi-point flux approximation \(L\) and \(G\) methods. The results seem to be entirely new and they are very interesting. The article includes several numerical methods and deserves to be read.