Three professors from the SimTech community at the University of Stuttgart are currently contributing to the international symposium “Mathematical and Applied Aspects of Complexity Reduction Methods” at the Collège de France in Paris: Benjamin Stamm, Jörg Fehr, and Bernard Haasdonk.
Held from 22 to 24 June 2026, the symposium brings together leading researchers working in approximation theory, numerical analysis, reduced-order modelling, scientific computing, machine learning, and industrial simulation. Its focus is on methods that make complex simulations more efficient while retaining mathematically controlled accuracy and reliable error estimates.
Benjamin Stamm presented “The Reduced Basis Method for Ground-State Eigenvalue Computations.” His talk addressed efficient reduced-basis approaches for parametrised eigenvalue problems, motivated in particular by many-body quantum systems. The methods aim to enable the analysis of ground-state phase diagrams while addressing challenges such as degenerate eigenstates and reliable error certification.
Jörg Fehr spoke on “From Latent Space Representations to Practical Surrogate Models for Structural Dynamical Systems.” He discussed how model order reduction, system identification, and machine learning can be combined to develop efficient and physically meaningful surrogate models for engineering applications, including crash simulations, disc-brake models, and coupled structural and fluid-dynamic systems.
Bernard Haasdonk will present “Greedy Kernel-Based Surrogates for Approximating Parametric PDEs.” His contribution introduces kernel-based approximation methods for parametrised partial differential equations that can construct surrogate models without requiring precomputed solution snapshots. The approach is particularly relevant for challenging settings involving moving sources, non-affine geometries, or high-dimensional domains.
The strong Stuttgart presence at this renowned venue highlights SimTech’s international visibility and expertise in complexity reduction, model order reduction, and data-integrated simulation.