<p>This paper mainly focuses on the superconvergence analysis of the discontinuous Galerkin solutions for the second-kind Volterra integral equations (VIEs) with weakly singular kernel. To improve the numerical accuracy of the discontinuous Galerkin solution and achieve a good superconvergence result, two types of postprocessing techniques based on iteration and interpolation are proposed for the discontinuous Galerkin solution, and the superconvergence properties are investigated in detail for both postprocessing techniques. For the iterated DG postprocessing solution, superconvergence is obtained under the same regularity assumptions as that for convergence, unlike for the iterated collocation method, where one has to improve the regularity assumptions to obtain superconvergence. For the interpolated postprocessing solution, the same highest attainable superconvergence order can be arrived under a higher regularity requirement, however, comparing with the iterated postprocessing technique, the computational cost of the interpolation postprocessing technique is less. Numerical examples are presented to illustrate the performance of the method.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Superconvergence and postprocessing of discontinuous Galerkin solutions for weakly singular Volterra integral equations

  • Wenping Yuan,
  • Hui Liang

摘要

This paper mainly focuses on the superconvergence analysis of the discontinuous Galerkin solutions for the second-kind Volterra integral equations (VIEs) with weakly singular kernel. To improve the numerical accuracy of the discontinuous Galerkin solution and achieve a good superconvergence result, two types of postprocessing techniques based on iteration and interpolation are proposed for the discontinuous Galerkin solution, and the superconvergence properties are investigated in detail for both postprocessing techniques. For the iterated DG postprocessing solution, superconvergence is obtained under the same regularity assumptions as that for convergence, unlike for the iterated collocation method, where one has to improve the regularity assumptions to obtain superconvergence. For the interpolated postprocessing solution, the same highest attainable superconvergence order can be arrived under a higher regularity requirement, however, comparing with the iterated postprocessing technique, the computational cost of the interpolation postprocessing technique is less. Numerical examples are presented to illustrate the performance of the method.