Improved discretization error estimates for first-order system least squares (Q4454895)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 2053279
Language Label Description Also known as
default for all languages
No label defined
    English
    Improved discretization error estimates for first-order system least squares
    scientific article; zbMATH DE number 2053279

      Statements

      Improved discretization error estimates for first-order system least squares (English)
      0 references
      0 references
      0 references
      0 references
      8 March 2004
      0 references
      least squares discretization
      0 references
      finite elements
      0 references
      error bounds
      0 references
      convergence
      0 references
      Dirichlet problem
      0 references
      Poisson equation
      0 references
      continuation methods
      0 references
      Solution of the Dirichlet problem for the Poisson equation in two dimensions using first-order system least squares involves a bilinear functional \[ {\mathcal F}({\mathbf u},p;{\mathbf v},q) := (\nabla p-{\mathbf u},\nabla q-{\mathbf v}) +(\nabla\cdot{\mathbf u},\nabla\cdot{\mathbf v})+(\nabla\times{\mathbf u}, \nabla\times{\mathbf v}). \] For \(\alpha>1\), the bilinear form \(\mathcal F\) is termed \(H^\alpha\)-regular if for every \(g=(g_1,g_2,g_3)\in(L^2)^3\), the solution of the weak equation \[ {\mathcal F}({\mathbf w},r;{\mathbf v},q)=(g;({\mathbf v},q)) \] is in \((H^\alpha)^3\) and satisfies \[ \| ({\mathbf w},r)\| _\alpha\leq C\| g\| _0. \] This paper presents an analysis of \(\mathcal F\) showing it is \(H^2\)-regular on the unit square, on a domain with smooth boundary, and on a polygonal domain with sufficiently small exterior angles at vertices. The proof employs continuation methods. This regularity result is sufficient to show that if the solution to the weak equation satisfies \(p\in H^{2+\epsilon}\) and \({\mathbf u}\in (H^{1+\epsilon})^2\), then approximations \(p_h\) and \({\mathbf u}_h\) converge according to the estimates \(\| p-p_h\| _1=O(h)\), \(\| {\mathbf u}-{\mathbf u}_h\| _1= O(h^\epsilon)\), and \(\| ({\mathbf u},p)-({\mathbf u}_h,p_h)\| _0= O(h^{1+\epsilon})\).
      0 references

      Identifiers

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