Besov regularity and new error estimates for finite volume approximations of the \(p\)-Laplacian (Q2501248)

From MaRDI portal





scientific article; zbMATH DE number 5051396
Language Label Description Also known as
default for all languages
No label defined
    English
    Besov regularity and new error estimates for finite volume approximations of the \(p\)-Laplacian
    scientific article; zbMATH DE number 5051396

      Statements

      Besov regularity and new error estimates for finite volume approximations of the \(p\)-Laplacian (English)
      0 references
      0 references
      0 references
      0 references
      4 September 2006
      0 references
      The authors investigate the finite volume numerical approximation on rectangular meshes of solutions to the \(p\)-Laplacian with homogeneous Dirichlet boundary conditions \((1 < p < +\infty)\): \[ \begin{aligned} -\text{div} (| \nabla\bar u| ^{p-2}\nabla\bar u) = f & \quad \text{on } \Omega,\\ \bar u = 0 & \quad \text{on } \delta\Omega. \end{aligned} \] An optimal error bound is proved in the discrete \(W^{1,p}(\Omega)\) norm for the finite volume approximation of the \(p\)-Laplacian \((p > 2)\) with general \(L^{p'}(\Omega)\) data on uniform Cartesian grids. In this case, where the exact solution does not belong to \(W^{2,p}(\Omega)\), classical finite volume techniques can not be used to estimate the consistency error of numerical fluxes. As a consequence of this fact, the authors' proof is based on taking advantages of the variational structure of the continuous problem and of its discrete approximation. The error estimate follows if one shows a discrete Besov estimate independent of the mesh size for the approximate solution but also for the projection of the exact solution on the space of piecewise constant functions. Moreover, numerical results for some analytical solutions with critical regularity give some evidences that the convergence rate obtained is optimal.
      0 references
      0 references
      finite volume schemes
      0 references
      a priori error estimates
      0 references
      consistency
      0 references
      numerical results
      0 references
      convergence
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references