Fast Numerical Algorithms for Large-Scale Problems

The course DN2230, Fast Numerical Algorithms for Large-Scale Problems, starts Monday October 25 in room E51. It is a D level course which gives 7.5 credits, and it focuses primarily on numerical linear algebra; how to efficiently solve large linear systems and eigenvalue problems. We will present Krylov space iteration techniques, the QR algorithm, and Multigrid. We will also look at fast multi-pole methods. These were introduced to speed up computations for the n-body problem, i.e. the problem of computing the evolution of a system of particles that interact via Coulomb forces, e.g. stars or electrically charged atoms. The method can also be applied to integral equations related to boundary value problems for PDE.