Multiple coarse grid acceleration for multiscale multigrid computation

Ruxin Dai, Jun Zhang, Yin Wang

Research output: Contribution to journalArticlepeer-review

11 Scopus citations

Abstract

The multiscale multigrid method uses an iterative refinement procedure with the Richardson extrapolation technique to obtain a higher order solution. The computational cost for the iterative refinement procedure may be significant for some ill conditioned coefficient matrices. In this paper, we proposed an alternative strategy using multiple coarse grids to eliminate the iterative refinement procedure, and thus accelerate the multiscale multigrid computation. Numerical investigations show that our multiple coarse grid computing strategy is more efficient and scalable than the iterative refinement procedure. The multiscale multigrid method with multiple coarse grid strategy is used to solve two dimensional (2D) Poisson equation and convection diffusion equation, but the idea can be used to solve other partial differential equations.

Original languageEnglish
Pages (from-to)75-85
Number of pages11
JournalJournal of Computational and Applied Mathematics
Volume269
DOIs
StatePublished - Oct 15 2014

Bibliographical note

Funding Information:
The third author’s research work was supported in part by NSF under grant CNS-1157162 and NVIDIA CUDA Teaching Center .

Keywords

  • Higher order solution
  • Multiple coarse grids
  • Multiscale multigrid method
  • Richardson extrapolation

ASJC Scopus subject areas

  • Computational Mathematics
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Multiple coarse grid acceleration for multiscale multigrid computation'. Together they form a unique fingerprint.

Cite this