Brit Steiner
Pfaffenwaldring 57
70569 Stuttgart




Directly to

zur Startseite

KerMor - Kernel Methods for Model Order Reduction

Principal investigators
Bernard Haasdonk
Daniel Wirtz
Begin 2/1/10
End 10/31/10

Our project is concerned with tackling the difficulties when handling large scale, parametrized nonlinear dynamical systems that occur naturally in biochemical settings as cell apoptosis simulation, for example.

Due to more sophisticated simulation models in natural and engineering sciences advanced reduction methods are moving more and more into the focus of many researchers and system modelers.

Accessing both the mathematical and computer science background of Daniel Wirtz the project roughly be divided into two parts, which are closely interlocked:

Development of a comprehensive software framework using current scientific paradigms that allows to simulate and reduce a broad range of dynamical systems as well as its integration/embedding into existing model reduction frameworks like RBMatlab.

The other part aims to derive (a-posteriori) error bounds/estimates that directly influence the algorithm design of KerMor and will ensure the quality of the solutions obtained. For validation and test of the work several cooperations throughout the SimTech networks are being established to give access to current models.

The software framework KerMor is being developed during the projects lifetime for implementations of our algorithms and work.


Due to the graduation in the SimTech graduate school several connections to other SimTech projects and researchers are established.

As one application field, we are establishing a cooperation with S. Waldherr regarding biochemical simulations. Another link is created to M. Daub to obtain a nonlinear parametrized model for cell death simulation.

We initiated an interdisciplinary SimTech Seminar on “Model Order Reduction” in which general exchange and bilateral discussions were realized. Some hereby established cooperations include work with the ITM (J. Fehr, P. Eberhard), the IST (S. Waldherr, M. Löhning, F. Allgöwer) and IADM (M. Daub, G. Schneider)