Rigorous Quantitative Analysis of Multigrid
نویسنده
چکیده
Exact numerical convergence factors for any multigrid cycle can be predicted by local mode (Fourier) analysis. For general linear elliptic PDE systems with piecewise smooth coeecients in general domains discretized by uniform grids, it is proved that, in the limit of small meshsizes, these predicted factors are indeed obtained, provided the cycle is supplemented with a proper processing at and near the boundaries. That processing, it is proved, costs negligible extra computer work. Apart from mode analysis, a Coarse Grid Approximation (CGA) condition is introduced which is both necessary and suucient for the multigrid algorithm to work properly. The present part studies the L 2 convergence in one cycle, for equations with constant coeecients. In the sequel 1 P2], extensions is discussed to many cycles (asymptotic convergence), to more levels with arbitrary cycle types (V , W, etc.), and to FMG algorithms. Various error norms and their relations to the orders of the inter-grid transfer operators are analyzed. Global mode analysis, required to supplement the local analysis in various border cases, is developed and partial relaxation sweeps are systematically introduced into both analysis and practice. 1 Most of the present paper, together with P2], appeared as RL] (containing an erroneous Appendix, later removed).
منابع مشابه
Mass Smoothers in Geometric Multigrid for Isogeometric Analysis
We investigate geometric multigrid methods for solving the large, sparse linear systems which arise in isogeometric discretizations of elliptic partial differential equations. In particular, we study a smoother which incorporates the inverse of the mass matrix as an iteration matrix, and which we call mass-Richardson smoother. We perform a rigorous analysis in a model setting and perform some n...
متن کاملMultiscale Scienti c Computation : Six Year Research
The Gauss Center research on multiscale computational methods is reported, emphasizing main ideas and interrelations between various elds, and listing the relevant bibliography. The reported areas include: top-eeciency multigrid methods in uid dynamics; atmospheric ows and data assimilation; feedback optimal control; PDE solvers on unbounded domains; wave/ray methods for highly indeenite equati...
متن کاملPatch Smoothers for Saddle Point Problems with Applications to PDE-Constrained Optimization Problems
We consider a multigrid method for solving the discretized optimality system of a PDE-constrained optimization problem. In particular, we discuss the construction of an additive Schwarz-type smoother for a class of elliptic optimal control problems. A rigorous multigrid convergence analysis yields level-independent convergence rates. Numerical experiments indicate that the convergence rates are...
متن کاملA Multigrid Semismooth Newton Method for Contact Problems in Linear Elasticity
A multigrid semismooth Newton method for elastic contact problems is developed and analyzed. We show that after a suitable regularization of the contact problem superlinear local convergence is obtained for a class of semismooth Newton methods. In addition, an estimate for the order of the error introduced by the regularization is derived. The main part of the paper is devoted to the analysis o...
متن کاملBlock - Structured Multigrid for the Navier - StokesEquations : Experiences
This paper summarizes investigations concerning the algorithmic scalability of multi-grid methods for partial diierential equations on MIMD distributed memory systems. It is shown that even multigrid methods which are distinguished by h-independent convergence rates are not scalable in a rigorous sense. We develop their parallel asymptotic computational complexity for diierent types of multigri...
متن کامل