Archiv Kolloquium / Former Colloquia
Multirate Integration Schemes for Coupled Systems
Prof. Dr. Michael Günther
(Bergische Universität Wuppertal)
23.10.2018, 16:00 Uhr
C_Fun_Gene, Konferenzraum A+B
Felix-Haussdorff-Str. 8 (Eingang vom Beitz-Platz)
In technical applications one is often confronted with numerically solving coupled systems of time-dependent differential equations of various types. These systems are usually equipped with dynamics acting on different time scales. To exploit this multirate behaviour, multistep integration schemes use different time steps tailored to the dynamics of the subsystems. This talk will give a survey on these schemes, including their development, analysis with respect to consistency and stability and applications in refined network analysis.
Michael Günther was born in Munich in 1967. He studied Mathematics with minor subject Physics (diploma in 1992), received a Ph.D. in Computer Science in 1995, both at Technische Universität München, and obtained his venia legendi (Habilitation) at Universität Karlsruhe (TH) in 2001. After positions as research assistant, assistant professor and professor at proxy at Technische Universität Darmstadt, Universität Karlsruhe (TH) and Universität Ulm, he joined the Bergische Universität Wuppertal in 2003, where he currently holds the chair of applied mathematics and numerical analysis. Since 2016 he is dean of the school of mathematics and natural sciences at the Bergische Universität Wuppertal .
Michael Günther's research interests include numerical analysis and simulation of time-dependent differential equations inclduing differential algebraic equations and coupled problems. The fields of application comprise computational physics (lattice QCD), computational finance (financial deriivatives and option pricing) and computational electronics.
Interessenten sind herzlich willkommen. Prof. Dr. Roland Pulch
Assembling the Network of Life
Prof. Dr. Vincent Moulton
(University of East Anglia)
8.6.2018, 14:15 Uhr
C_Fun_Gene, Konferenzraum A+B+C
Felix-Haussdorff-Str. 8 (Eingang vom Beitz-Platz)
Biologists commonly represent the evolution of organisms using a phylogenetic tree, that is, a leaf-labeled graph-theoretical tree, the tree-of-life providing a well-known example of such a tree. In recent years, however, it has become increasingly recognised that the evolutionary history of certain organisms (e.g. plants, viruses and bacteria) is not always best represented by a tree. This is due to evolutionary processes that take place on the molecular level, such as recombination, lateral gene transfer and recombination. In such cases, phylogenetic networks - more complex leaf-labeled graphs - can provide a useful alternative to trees.
In this talk, we discuss the problem of building up phylogenetic networks from simpler networks, which leads to some interesting mathematical and computational problems.
Vincent Moulton is Professor in Computational Biology at University of East Anglia. After completing his undergraduate studies at University of Warwick in 1987, he moved to the USA, where he completed an MSc at University of Washington in 1991, and his PhD at Duke University in 1994.
He then worked as a researcher at University of Bielefeld, DE, and University of Canterbury/Massey University, NZ. In 1997 he moved to Sweden, where he was employed as Senior Lecturer at Mid Sweden University and, as of 2002, Professor in Bioinformatics at Uppsala University. He joined University of East Anglia in 2004.
Interessenten sind herzlich willkommen. Prof. Dr. Marc Hellmuth
Genus and Unknotting Number of a Knot
Prof. Dr. Stefan Friedl
30.4.2018, 16:00 Uhr
A knot is closed curve in R³. Every knot bounds an orientable surface, and the minimal genus of such a surface is called the genus of the knot. Every knot can be transformed into the trivial knot by a finite number of crossing changes. The minimal number of crossing changes is called the unknotting number. We will introduce these invariants of knots and discuss some properties and open questions.
Stefan Friedl studied Mathematics and Physics in Regensburg and later at Brandeis University in Massachusetts. He obtained a Ph.D in Mathematics from Brandeis University in 2003. He held postdoc positions in Munich, Houston, and Montreal. He became assistant professor at University of Warwick, then professor at University of Cologne, and since he is 2013 professor at University of Regensburg.
Stefan Friedl's research interests are Knot theory, low–dimensional topology, symplectic 4–manifolds, and related algebra.
Kristina Wicke: Die Qual der Wahl beim Plätzchenbacken - phylogenetische Entscheidungshilfen
Peter Kristel: Fair cake-cutting
13.12.2017, 15:30 Uhr
SR 1, Franz-Mehring-Str. 47
Diagnosis and initialization methods for differential algebraic equations
Prof. Dr. Diana Estévez Schwarz
(Beuth Hochschule für Technik Berlin)
14.11.2017, 16:00 Uhr
Differential algebraic equations (DAEs) arise in many applications, such as the simulation of constraint multibody systems, electrical networks, the discretization of PDEs, etc. In general, the quality of the numerical solutions of DAEs depends on assumptions that are often not checked rigorously. Consequently, simulations may fail or, even worse, may deliver arbitrary results.
In this talk, we discuss some specific difficulties of DAEs and illustrate them with several examples. Particular emphasis will be placed on the determination of the so-called index of the DAE, the diagnosis of singular points, and the computation of consistent initial values using a projector-based approach. Moreover, we present how automatic differentiation (AD) can be used in this context.
Diana Estévez Schwarz was born in 1972, studied mathematics at the Technische Universität Berlin and received her PhD in 2000 from the Humboldt Universität zu Berlin. Afterwards, she worked as development engineer for circuit simulation software at Infineon Technologies in Munich before she obtained a professorship at the Technische Fachhochschule Berlin in 2006, that was renamed Beuth Hochschule für Technik Berlin in 2009. Her research is focused on differential algebraic equations, in particular, for the computation of the index as well as consistent initial values, the characterization of structural properties and the diagnosis of singularities.
Curvature on graphs
Prof. Dr. Matthias Keller
12.07.2017, 14:15 Uhr
The geometry of a graph is tightly linked the spectral theory of graph Laplacian. In this talk we discuss a notion of curvature on planar graphs which goes back to ideas of Descartes. This notion of curvature is then used to deduce spectral bounds from curvature bounds.
Matthias Keller was born in 1980. He studied Mathematics in Chemnitz.
He received his PhD in 2010 from the University of Jena. In 2015 he finished his habilitation also in Jena. He had postdoc and visiting professorship positions in Jerusalem and Haifa before he obtained full professorship in Potsdam. Since 2017 he is a principal investigator in the DFG priority programme "Geometry at Infinity".
The main focus of Matthias Kellers research is on spectral theory of graphs and its relation to spectral geometry of Riemannian manifolds.
Reachability in Generalizations of Topological Spaces - Applications to Evolution and Chemistry
Prof. Dr. Peter F. Stadler
16.05.2017, 12:15 Uhr
Vortragssaal der Universitätsbibliothek E.21
Reachability is a notion that can, in relatively simple cases, be modeled by non-idempotent closure functions. This picture is sufficient e.g. in the context of biological evolution and evolutionary search. The study of chemical universes, comprising a not necessarily finite set of chemical compound organized by chemical reaction that convert them into each other, however, indicates that generalized closures do not convey all the necessary details: after all, chemical universes are directed hypergraphs. An alternative inroad is to start from the complementary notions of separation and connectedness and to generalize these to a non-symmetric setting. In this presentation I will focus on basic properties of the resulting abstract reachability spaces and on open questions.
Peter F. Stadler, born in 1965, received his Ph.D. in Chemistry from the University of Vienna in 1990 following studies in chemistry, mathematics, physics and astronomy. After a PostDoc at the Max Planck Institute for Biophysical Chemistry in Goettingen, he returned to Vienna and qualified as a professor in 1994 in Theoretical Chemistry. Since 1994 he is External Professor at the Santa Fe Institute, a research center focused on Complex Systems. In 2002 he moved to the University of Leipzig as Full Professor of Bioinformatics. Since 2009 he is External Scientific Member of the Max Planck Society and head of the "Discrete Biomathematics" group at the MPI for Mathematics in the Sciences. He is corresponding member abroad of the Austrian Academy of Sciences since 2010.
The general theme of his research is the search for a consistent understanding of biological processes (with an emphasis on (molecular) evolution) at the genotypic, phenotypic, and dynamical level.