Analysis of numerical methods

The first lecture took place Tuesday the 30th of August. A complete schedule can be found here. For further information about the course, see the syllabus or the tentative declaration of contents for the lectures.

The exam and solutions

can be downloaded from here.
Exam 20111019, proposed solutions.

Teachers

Gunilla Kreiss, lectures.
Emil Kieri, assignments and seminars.

Examination

Written exam and mandatory assignments.

The exam will take place Wednesday October 19th. Allowed tools: Mathematics Handbook by Råde and Westergren.

There will be four mandatory assignments, which should be solved in groups of between one and three, preferably two, students. Discussion about the assignments between the groups is encouraged. Each assignment should be handed in before each corresponding seminar. It is not necessary that you have completely solved all problems before the seminar, partial solutions will be accepted. It is enough that you have worked on the problems to such a degree that you can discuss solution techniques and difficulties at the seminar. However, if you do not hand in the assignment before the seminar, or do not attend the seminar, a complete and correct solution is required.

Deadlines for the assignments are,
Assignment 1: September 16th,
Assignment 2: September 30th,
Assignment 3: October 7th,
Assignment 4: October 14th.

Completed assignments should be left in the assignment inbox number 57 on floor 2, building 2, or given directly to Emil at the seminar, or submitted electronically through Studentportalen. The absolute deadline for this year's assignments is October 28th. Assignments submitted after this date will not be considered, but are referred to the course instance next year.

The assignments can be found at Studentportalen.

Litterature

The recommended course textbook is

Bertil Gustafsson, High Order Difference Methods for Time Dependent PDE. Springer, Berlin, 2008.

For the earlier 7.5 hp version of the course,

John Strikwerda, Finite Difference Schemes and Partial Differential Equations. SIAM, Philadelphia, 2007.

was used. It still has a decent coverage of most of the course contents.

Matlab

The implementation part of the assignments should be written in Matlab. The university provides a Matlab license for students to use on their private computers, see Matlab for students for more information.

Old exams

Old exams from a previous version of the course can be found here. The most important differences are that normal mode (GKS) analysis and multigrid methods have left, and spectral methods and the fast Fourier transform have entered the course.