Global convergence and local superconvergence of first-kind Volterra integral equation approximations (Q2902205)

From MaRDI portal





scientific article; zbMATH DE number 6067217
Language Label Description Also known as
default for all languages
No label defined
    English
    Global convergence and local superconvergence of first-kind Volterra integral equation approximations
    scientific article; zbMATH DE number 6067217

      Statements

      Global convergence and local superconvergence of first-kind Volterra integral equation approximations (English)
      0 references
      0 references
      0 references
      0 references
      17 August 2012
      0 references
      Volterra integral equations of the first kind
      0 references
      collocation
      0 references
      quadrature Galerkin method
      0 references
      discontinuous Galerkin method
      0 references
      global convergence
      0 references
      local superconvergence
      0 references
      numerical examples
      0 references
      discontinuous piecewise polynomial methods
      0 references
      convolution kernel
      0 references
      The results of the paper are concerning the convergence and superconvergence analysis of three discontinuous piecewise polynomial methods for the approximation of the solution of first kind Volterra integral equations with convolution kernel: discontinuous collocation, discontinuous Galerkin (DG), and quadrature discontinuous Galerkin (QDG) methods. In order to simplify the convergence analysis for collocation methods, the authors introduce new polynomial basis functions: the derivatives of the Lagrange polynomials of degree \(m + 1\). This choice leads to a new result about the superconvergence of discontinuous collocation methods.NEWLINENEWLINE Moreover, it is proved that a QDG method with degree of precision equal to \(m\) and \(m + 1\) quadrature points is equivalent to a collocation scheme such that its convergence analysis reduces it to those of the collocation method. The convergence of QDG methods is separately investigated for odd and even \(m\). As a consequence, in the even case the superconvergence of the Gauss-Legendre QDG method is exactly determined. Finally, it is proved that if the degree of precision of a quadrature rule is at least \(2m+2\), then the convergence and superconvergence behaviour of the QDG scheme is the same as of the DG scheme. The optimality of the theoretical convergence rates is confirmed by some numerical tests.
      0 references

      Identifiers