|
|
The research in numerical linear algebra has a long tradition at the
department and is internationally well-established. It comprises the whole
spectrum of problems including perturbation theory and estimation of the
conditioning of the problem, development of algorithms and computable
error bounds and their robust and efficient implementation in
state-of-the- -art software modules.
The main topics include generalized eigenvalue and subspace problems,
canonical forms,
matrix factorizations, matrix equations (Sylvester, Ricatti etc),
matrix functions, least squares problems, and sparse matrix techniques.
The outcome of the research has applications in control systems (e.g.,
computing the controllability subspace or the distance to the closest
uncontrollable system) and several other engineering and scientific areas,
and constitute an important platform for studies in (higher-index)
differential-alge braic systems as well as in nonlinear optimization.
The research also contributes to international numerical libraries, such as
LAPACK (intended for high-performance computing) and the SLICOT library
within the NICONET project.
Some ongoing projects ...
Group members
Former group members
|
|
|
|
