* Dates within the next 7 days are marked by a star.
Asst. Prof. Philine Schiewe (Aalto University)
Multiobjective network optimization
* Today * Tuesday 08 September 2026, 15:15, M1 (M232)
In multiobjective optimization, several - often conflicting - objective functions are minimized simultaneously. As a consequence, there is generally no single total ordering of the feasible solutions and hence no unique notion of optimality. Instead, the goal is to identify efficient solutions, whose images are the non-dominated points forming the so-called Pareto front. Many network design tasks involve conflicting goals, such as cost, travel time, and demand attraction. Here, we examine two such multiobjective network optimization problems in detail. For both problems, we propose tailored scalarizations that retain problem structure, and analyze complexity and size of the Pareto front. We test the developed algorithms in an extensive computational study, on both generated and real-world data.
Konstantinos Bampouras (Aalto University)
Boundary-weighted Fourier inequalities for convex domains
* Wednesday 09 September 2026, 10:15, M3 (M234)
Analysis seminar / Hytönen
Jesse Piispanen
On well-rounded lattices and theta function minimization (MSc thesis presentation)
* Thursday 10 September 2026, 14:15, M3 (M234)
Advisor: Max Forst, Supervisor: Camilla Hollanti
ANTA Seminar / Hollanti et al.
Talal Alrawajfeh (Univ. Helsinki)
TBA (the first meeting of Aalto-Helsinki formalized mathematics seminar)
* Friday 11 September 2026, 13:15, U4062, University of Helsinki main building (Fabianinkatu 33)
Aalto-Helsinki formalized mathematics seminar
Jaime Pardo
Proximal Tensor-Free Approach to Bilinear Inverse Problems
Tuesday 15 September 2026, 15:15, M1 (M232)
Numerical Analysis seminar
Inka Hirvinen (Umicore Battery Materials Finland)
Uncertainty-Aware Data Mapping and Interpretable Modelling for Complex Process Systems (MSc thesis presentation)
Thursday 17 September 2026, 15:15, Zoom
Battery materials play a significant role in enabling the transition to clean energy
and transportation, thus making the development of new battery materials essential.
The commonly employed trial-and-error approach to battery materials development
is costly, which makes computational modelling an attractive option. In this thesis
we utilize historical data from trial runs to train and compare multiple models to
predict the specific surface area (SSA) of precursor cathode active materials (pCAM).
Subsequently, we employ the best-performing model as a surrogate model to optimize
process parameters, which enables the production of battery materials with desired
properties.
Compared to a previously utilized linear regression model, several models achieve
higher predictive performance. The XGBoost regression model achieves both the
lowest prediction error and the highest computational efficiency, which makes it the
most promising method for SSA prediction. Additionally, we identified a Pareto-
efficient front of optimal process parameter combinations using the trained XGBoost
model to approximate the desired properties.
Lorenzo Zacchini (Aalto University)
TBA (midterm review)
Wednesday 23 September 2026, 10:15, M3 (M234)
Analysis seminar / Hytönen
Jani Onninen (Syracuse University)
TBA
Wednesday 23 September 2026, 11:15, M3 (M234)
Analysis seminar
Juho Korkeala
Exotic mean oscillation and compactness of commutators (Master thesis presentation)
Thursday 24 September 2026, 10:15, M3 (M234)
Analysis seminar (extra meeting) / Hytönen
Aleksis Koski
TBA
Wednesday 07 October 2026, 10:15, M3 (M234)
Analysis seminar
Henri Lahdelma
Midterm review talk
Wednesday 21 October 2026, 10:15, M3 (M234)
Analysis seminar
Wontae Kim (Korea Institute for Advanced Study)
TBA
Wednesday 04 November 2026, 10:15, M3 (M234)
Analysis seminar
Show the events of the past year
Page content by: webmaster-math [at] list [dot] aalto [dot] fi