Dozenten: | Volker Mehrmann, Matthias Bollhöfer, Christian Mehl |
Koordination: | Christian Mehl |
LV-Termine: | Do 10-12 in MA 313 |
Inhalt: | Vorträge von Diplomanden, Doktoranden, Postdocs und manchmal auch Gästen zu aktuellen Forschungsthemen |
Vollständige vorläufige Terminplanung: | ||||
Datum | Uhrzeit | Raum | Vortragende(r) | Titel |
---|---|---|---|---|
Do 30.10.2003 | 10:15 | MA 313 |
Vorbesprechung | |
Do 6.11.2003 | 10:15 | MA 313 |
Daniel Kressner | Structured Condition Numbers for Invariant Subspaces (Abstract) |
Do 13.11.2003 | 10:15 | MA 313 |
Olga Holtz | M-matrices satisfy Newton's inequalities (Abstract) |
Do 20.11.2003 | 10:15 | MA 313 |
Michael Karow | On Real and Complex Operator Norms of a Real Matrix |
Do 4.12.2003 | 10:15 | MA 313 |
Simone Bächle | Index reduction by minimal extension for differential-algebraic equations in circuit simulation |
Do 11.12.2003 | 12:00 | MA 464 |
Prochko Prochkov | Numerical Simulation of Train Traffic in Large Networks via Time-optimal Control |
Do 18.12.2003 | 10:00 | MA 313 |
Falk Ebert | Controlled simulator coupling - Part II |
Do 18.12.2003 | 11:00 | MA 313 |
Tim Wichmann (Fraunhofer ITWM, Kaiserslautern) | Symbolic Simplification of Nonlinear DAE-Systems in Analog Circuit Design (Abstract) |
Do 15.1.2004 | 10:15 | MA 313 |
Christian Otto | The Jacobi-Davidson algorithm |
Do 22.1.2004 | 10:00 | MA 313 |
Michael Schmidt | Model reduction via Karhunen-Loeve-expansion/Proper Orthogonal Decomposition |
Do 22.1.2004 | 11:00 | MA 313 |
Sonja Schlauch | Analysis, Modeling and Numerical Simulation of Drop Size Distributions in Stirred Liquid-liquid Dispersions |
Do 29.1.2004 | 10:00 | MA 313 |
Ulrike Baur | Model reduction of linear time-varying systems |
Do 29.1.2004 | 11:00 | MA 313 |
Andreas Hilliges | Simulation of railway tracks under periodic excitation |
Do 12.2.2004 | 10:00 | MA 313 |
Chunchao Shi | Characterizations of the distance to singularity for matrix pencils |
Do 12.2.2004 | 11:00 | MA 313 |
Elena Virnik | Algebraic Multigrid (AMG) preconditioner applied to solving (I-A^T)x=0 with GMRES where A is transition matrix of a Markovian process |
Do 19.2.2004 | 10:00 | MA 313 |
Stefan Seidel | Simulation and Optimisation of Rail Traffic in Realistic Networks |
Do 19.2.2004 | 11:00 | MA 313 |
Alex Streller | Finite Element Discretization of the Navier-Stokes Equation for the Backward Facing Step |
Interessenten sind herzlich eingeladen!
Rückblick:
Weitere Vorträge siehe auch:
Abstracts zu den Vorträgen:
Abstract:
The purpose of this talk is to present some known and new results
on the perturbation theory for invariant subspaces of matrices.
The first part is a review of well-known perturbation bounds, and
it is shown how the implicit function theorem can be used to derive
such bounds. The condition number of an invariant subspace is
introduced and connected to the singular values of a certain
Sylvester operator. If the perturbation is known to preserve
certain matrix structures then this condition number could severly
overestimate the actual sensitivity of an invariant subspace.
This question is discussed in the second part of this talk, in
particular for
Abstract:
Newton's inequalities c_n^2 ≤ c_{n-1}c_{n+1} are shown
to hold for the normalized coefficients c_n of the characteristic
polynomial of any M- or inverse M-matrix. They are derived by
establishing first an auxiliary set of inequalities also valid for
both of these classes. Finally, a connection between Newton's inequlities
and the inverse eigenvalue problem for nonnegative matrices is discussed.
Abstract:
In general, the behaviour of an analog circuit can mathematically be
described by a nonlinear differential-algebraic equation system (DAE). In
the field of symbolic circuit analysis this equation system is set up
symbolically using computer algebra methods. Then it is further analyzed
in order to get insights into the circuit's behavior. Unfortunately, the
complexity of the symbolic equation system limits this approach to very
simple circuits. To extend the application field to practical examples
from industry, symbolic simplification techniques have been developed which
allow to generate a simplified equation system of much smaller complexity
which is still accurate within a user-given error bound. For linear
circuits this approach has been successfully used in industry for several
years now. Recently, research started to extend these techniques to
nonlinear circuits. In this talk we will present an overview of the
symbolic simplification techniques for nonlinear analog circuits, i.e. for
nonlinear DAE-systems. We describe the general approach as well as some
details of the algorithms, especially the so-called ranking method, a
technique which both improves the result and speeds up the computation. At
the end we will show that the idea of symbolic simplification is general
enough to be extended to additional application fields like the analysis of
pipelines.
Impressum | Christian Mehl 18.2.2004 |