Introduction to Multigrid Methods
Department of Applied Mathematics, University of Twente.Professor of Numerical Analysis and Computational Mechanics Department of Applied Mathematics Chair Mathematics of Computational Science Faculty of Electrical Engineering, Mathematics and Computer Science University of Twente Homepage is http://wwwhome.math.utwente.nl/~vegtjjw/
本讲教师：Jaap van der Vegt
Multigrid methods can provide very efficient iterative methods for thesolution of large systems of (non)linear algebraic equations, resulting forinstance from the discretization of partial differential equations. In amultigrid method several coarsened approximations of the algebraic systemand well-designed smoothers are used to accelerate the convergence of theiterative method. This can result in very efficient iterative methods, butif one wants to develop new multigrid algorithms or understand theperformance of existing algorithms, then multilevel analysis isindispensible. In this class an outline of basic multigrid and iterative methods will begiven and mathematical techniques to understand and predict theirperformance will be discussed. No prior knowledge of multigrid or iterativemethods will be required. After this class you should be able to use basic iterative and multigridmethods, analyze and (approximately) predict multigrid performance usingmultilevel analysis and apply these techniques to improve and test multigridalgorithms. The main applications will be from numerical discretizations ofpartial differential equations.