Multigrid Multigrid

Multigrid - Definition and Overview

The multigrid methods are a group of algorithms for solving differential equations using a hierarchy of discretization. It has the advantage over other methods that it scales linearly with the number of discrete nodes used.

The key idea is the following: Assuming you have a differential equation which can be solved approximately (with a given accuracy) on a grid <math>i<math> with a given grid point density <math>N_i<math>. Assuming furthermore that a solution on any grid <math>N_i<math> may be obtained with a given effort <math>W_i = \rho K N_i<math> from a solution on a coarser grid <math>i+1<math> with grid point density <math>N_{i+1} = \rho N_i<math> (i.e. <math>K<math> is not dependent on <math>i<math>).

Using the Geometric_series, we then find for the effort involved in finding the solution on the finest grid <math>N_1<math>

<math>W_1 = W_2 + \rho K N_1<math>
<math>W_1 = W_3 + \rho^2 K N_1 + \rho K N_1<math>
<math>W_1 / (K N_1) + 1 = 1 + \sum_p \rho^p <math>
<math>W_1 / (K N_1) + 1 = 1 / (1 - \rho)<math>
<math>W_1 = (K N_1) (1 / (1 - \rho) - 1)<math>

i.e. a solution may be obtained in <math>O(N)<math> time.

See also

References and External links

Example Usage of Multigrid

Sciencia: Improving algebraic Multigrid interpolation operators for linear elasticity problems http://kele.es/or6
Copyright 2009 WordIQ.com - Privacy Policy  :: Terms of Use  :: Contact Us  :: About Us
This article is licensed under the GNU Free Documentation License. It uses material from the this Wikipedia article.