DN2252/DN2251 Numerical Algebra, 10.5/9 ECTS

NEWS: Welcome to the course! I hope that you will find this course interesting and useful. /Michael

The next round will be given in autumn 2008

Problems of linear algebra are fundamental to all algorithms in Scientific Computing. These problems are so basic since many (if not all) other numerical problems are reduced to one or a sequence of linear algebra problems. As a consequence, well-developed software is available (e.g., LAPACK, matlab,...). You have had an introduction to linear algebra algorithms in the basic course of numerical analysis.

So, why do it once again?

The aim of the present course is two-fold:

• Giving an understanding of the functioning of modern software. What can be expected from the codes?
• Being able to select algorithms and codes for your applications. What are good algorithms, and what are bad ones? Why?

Our analysis includes problems like solving linear systems of equations, linear least-squares problems, and eigenvalue problems. We will consider the difficulties when solving small scale, medium scale, and large scale problems (the latter including problems with millions of unknowns such that high-performance computers are necessary).

For further informations please visit the following pages.

The present course consists of two parts

• Part 1: Scientific Computation Computational Algebra. This part is also an obligotory component of the course DN2251 tilnum3.
• Part 2: Methods for large matrices

Other rounds can be found here.