This class will introduce iterative methods in numerical linear algebra
including standard methods (Jacobi, Gauss-Seidel, SOR),
projection methods (steepest descent), and Krylov space methods
(Arnoldi, Lanczos, CG, GMRES, QMR). Other topics include preconditioning
and parallel implementations, and eigenvalue problems. If time permits we will
also study iterative techniques for solving nonlinear problems.
We shall use the following texts:
``Iterative Methods for Solving Linear Systems ''
by Anne Greenbaum,
SIAM, Philadelphia, 1997.
``Iterative Methods for Sparse Linear Systems''
by
Youssef Saad, WPS, 1996. This book is made publicly available by the author:
Additional materials will be given in class. Some of them may be found below.
If you have questions, come and see me in 226C Fisher
every Tuesday after class. If you cannot make it,
please send me e-mail to arrange for a mutual aggreeable time.
For detailed information please consult the syllabus
(
PDF).
There will be several homework projects, on which the
final grade will be based.
Oct 10. Implement LU decomposition
(
example); Orthogonalization by modified Gram-Schmidt,
QR decomposition(
example); Residual norm steepest descent; Full orthogonalization method.
A set of large, sparse matrices from MatrixMarket
(
information and the tarred, gzipped
code).