Research Seminars in Mathematics
The research seminars are in the subjects of pure, applied, and computational mathematics and are usually held at the afternoons on Fridays. All are welcome to attend!
-
Research Seminar in Mathematics - Spectra of transposition generated Cayley graphs
Niklas Eriksen, institutionen för naturvetenskap och teknik, Örebro universitet
-
Research Seminar in Mathematics - Research on posing modelling problems: My work with AI
Morten Gulliksson, Institutionen för Naturvetenskap och Teknik, Örebro universitet.
-
Research Seminar in Mathematics - Research on posing modelling problems: What mathematics teacher candidates could gain from problem-posing experience?
Serife Sevinc
-
Research Seminar in Mathematics - Machine Learning for Molecular Dynamics Simulations of Carbon Nanomaterials Growth
Andreas Larsson, Institutionen för teknikvetenskap och matematik, Luleå tekniska universitet.
-
Research Seminar in Mathematics - Adaptive evolutionary trajectories in complexity: repeated transitions between unicellularity and differentiated multicellularity
Hanna Isaksson, Örebro universitet
-
Research Seminar in Mathematics - Unbiased approaches to compressive sensing and low rank matrix estimation
Marcus Carlsson, Matematikcentrum, Lunds universitet
Please contact magnus.ogren@oru.se if you have any questions regarding this seminar series.
2026
Speaker: Yifan Zhang, University of Ostrava
Date: Wednesday 19 August 13.15
Location: T207
Title: Coverings of Complete Graphs by Small Cliques and Applications
Abstract:
When solving partial differential equations by fast Boundary Element Methods, one is naturally led to large dense matrices whose computation must be split into many smaller pieces for parallel execution on multiple cores. This motivates a combinatorial load-balancing problem that can be modelled by covering the edges of a complete graph with small cliques, each representing a computational task.
In this talk, I will discuss recent results on such coverings of complete graphs by cliques of small order. The problem is closely related to classical questions in design theory, but is also driven by applications. If several clique sizes are permitted, then minimising the number of blocks alone does not adequately describe the quality of a covering. We therefore study a refined criterion involving both the block count and the excess, namely the number of edges covered more than once. I will present sharp results for coverings with 3- and 4-cliques, and then turn to the more difficult setting where 5-cliques are also allowed, including some open cases.
More broadly, the talk illustrates how abstract combinatorial structures can arise from concrete computational challenges, and how their study can feed back into the design of efficient parallel algorithms.
Speaker: Sweta Das
Date: 29 May 13.15
Location: T211
Title: Sweta Das 90%-seminarie, Canonical forms and their perturbations.
Abstract:
Canonical forms provide a fundamental framework for studying matrices and matrix pencils through invariants under various equivalence transformations. These forms reveal structural information of the underlying objects, e.g., eigenvalues, their multiplicities, and minimal indices. This information plays an important role in differential-algebraic equations, control theory, generalized eigenvalue problems, etc. However, in practical applications, canonical invariants are often sensitive to perturbations arising from measurement errors and numerical computations. This seminar presents an overview of canonical forms and their perturbation theory.
Speaker: Peter Johansson
Date: Friday 8 May 13.15
Location: T211
Titel: Optical forces
Abstract:
This talk will give a general background to optical forces and torques. I will discuss their physical background and show how they can be calculated in systems consisting of nanoparticles. The methods are based on generalized Mie theory, a scattering-theoretic framework in which the response of each particle is treated in terms of a multipole expansion. These methods will then be applied to describe the coordinated motion of a collection of gold nanoparticle that self-organize and move in a synchronized way.
Speaker: Johan Andersson, Örebro universitet
Date: Thursday 12 March, 15.15
Location: T211
Abstract: We report on our current research project where we prove the following results:
Let $\varphi: [0,\infty) \to [0,\infty)$ be a strictly increasing continuous function with $\varphi(0)=0$.
Then there exists a non-trivial entire function $f$ such that
\begin{gather*}
\int_{\mathbb{C}} \varphi(|f(z)|) dA(z) < \infty \\
\intertext{if and only if}
\int_0^\infty \frac{dx}{1+\varphi(e^{e^x})} = \infty.
\end{gather*}
Furthermore, when this divergence condition holds, the space of entire functions is dense in the Orlicz space $L_\varphi(\mathbb{C})$.
Similarly we prove that for any compact set $K \subset \mathbb{C}$ with non-empty interior, then the set of polynomials is dense in $L_\varphi(K)$ if and only if this integral diverges.
Remarkably, this allows us to approximate non-analytic functions, such as $f(z)=\bar{z}$, by polynomials in this Orlicz space.
In particular, our results resolves a gap left open by Kalton (1980) for the case where $\varphi(x) = (\log^+ \log^+ x)^p$ with $1 < p \le 2$.
Methods of proofs include a Carleman-type differential inequality and the tangential Arakelyan approximation theorem.
Note: This is our first piece of research where we have used a large language model (Gemini Pro 2.5, 3.0, 3.1) as an assistant, helpful for brain storming, editing part of the text and finding relevant references.