Courses persons responsible
Robert Piche

Implementations

	Period 1	Period 2	Period 3	Period 4	Period 5	Summer
Lecture	4 h/week	4 h/week	-	-	-	-
Exercise work	-	24 h/per	-	-	-	-

(Timetable for academic year 2007-2008)

Objectives
Learn the mathematical basis and practical issues related to modern numerical methods for the direct and iterative solution of linear equations, least squares problems and eigenvalue problems, and develop your proficiency in matrix mathematics along the way. The main themes are matrix factorizations and decompositions, perturbation theory and conditioning,floating point arithmetic and roundoff effects, structure-exploiting algorithms, and the engineering of numerical software.

Content

Content	Core content	Complementary knowledge	Specialist knowledge
1.	LU decomposition: algorithm, stability, implementation in multilevel architectures
2.	normal equations, QR and SVD decomposition, stability	data compression
3.	computing eigenvalues, perturbation analysis	the PageRank algorithm
4.	iterative methods: Jacobi, conjugate gradient, FFT, multigrid	Poisson PDE, medical imaging inverse problem

Requirements for completing the course
Term project and participation in lectures.

Evaluation criteria for the course

Study material

Type	Name	Auhor	ISBN	URL	Edition, availability...	Exam material	Language
Book	Applied Numerical Linear Algebra	J. W. Demmel	0-89871-389-7		SIAM 1997	No	English

Prerequisites

Code	Course	Credits	M/R
MAT-31090	MAT-31090 Matrix Algebra 1	5	Mandatory

Prequisite relations (Sign up to TUT Intranet required)

Remarks

Taught every second year. It will be taught in 2007-8. Lectures Wednesdays 2-4pm and Fridays noon-2pm in Tc133, first lecture 29.8.2007.

Scaling

Methods of instruction	Hours
Lectures	96
Assignments	60
Total sum	156

Correspondence of content
MAT-51620 Matrix Algebra 2

Course homepage

Last modified	20.08.2007
Modified by	Robert Piche