MTH 610: Geometric Multilevel Methods
Fall 2014
Instructor
Jay Gopalakrishnan
Venue
SEC 156
Times
Tue, Thu: 17:15 - 18:30
Office Hours
Tue 13:15 - 14:15 (in NH 309)
or by appointment.
Course Objective
The course aims to introduce mathematically oriented graduate
students to multigrid solvers in a geometric multilevel
setting. (Multigrid solvers are one of the most efficient
solvers available today for the numerical solution of certain
partial differential equations.)
Prerequisites
Consent of the instructor.
Technical Outline
We will begin with a brief discussion of the conjugate
gradient method as an iterative solver and how preconditioners
help us solve problems that are otherwise intractable.
We then proceed to auxiliary space preconditioners, the
fictitious space lemma of Nepomnyaschikh, and the XZ identity.
Next, we introduce the additive multilevel BPX preconditioner,
the multiplicative multilevel V-cycle algorithm, and a range
of convergence theorems like the Braess-Hackbusch
theorem.
Diary of results discussed daily
Click
here
for class notes/diary.
References
There is no textbook. However, these references will be useful:
- Bramble, `Multigrid Methods,' 1993.
- Hackbusch, `Multi-Grid Methods and Applications,' 1985.
- Briggs, Henson, and McCormick, `A Multigrid Tutorial,
2nd Edition,' SIAM publications, 2000.
Course materials will be available at the course D2L page.
Evaluation
Grades will be assigned based on class projects.
Jay Gopalakrishnan