Discussion Committee:
Dr. Anwar Saleh - Supervisor
Dr. Saed Mallak - External Examiner
Dr. Samir Mater - Internal Examiner
Authors:
Rania Taleb Mohammad Wannan
Abstract:
Partial differential equations appear in mathematical models that describe natural phenomena. Various methods can be used for solving such equations. In this thesis, an overview of classical iterative methods, as well as, the most recent multigrain methods is given. The classical iterative methods used are; the Jacobi, the Gauss-Seidel, and the SOR methods. Jacobi and Gauss-Seidel methods are efficient in smoothing the error but not in reducing it.
The smoothing property of some classical methods motivated the work done on multigrain methods. Poisson's problem in one and two dimensions has been used as model problem in the study of multigrain methods. The study shows that the rate of convergence of multigrain methods does not depend on the mesh size, a feature that makes multigrain methods good accelerator of classical methods like Gauss-Seidel.