An Energy-minimizing Interpolation for Robust Multigrid Methods

From MaRDI portal





The authors propose an energy minimization formulation with global constraint framework to construct robust interpolation which can handle elliptic partial differential equations (PDEs) with rough coefficients on structured or unstructured grids with one unifying approach. The method does not assume any underlying coarsening stategy, and hence can be used in combination with semicoarsening or algebraic multigrid coarsening. The authors discuss the connections between the present method and the smoothed aggregation method of \textit{P. Vaněk, J. Mandel}, and \textit{M. Brezina} [Computing 56, No. 3, 179-196 (1996; Zbl 0851.65087)]. The effectiveness of the method is demonstrated numerically mainly in two dimensions by applying it to a discontinuous problem, an oscillatory coefficient problem, and an anisotropic problem. The results show that the present method leads to a multigrid convergence rate comparable to algebraic multigrid. Empirically, the convergence rate is independent of the coefficients of the underlying PDE, in addition to being independent of the mesh size.




Cited in
(42)








This page was built for publication: An Energy-minimizing Interpolation for Robust Multigrid Methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4509791)