Numerical Linear Algebra for Dynamical Systems

Prof. Martin Stoll

Major Research Interests

Partial differential equations (PDE) are an ubiquitous mathematical tool for modeling many real-world phenomena. Our group investigates fast and robust solvers for the accurate solution of discretized PDEs. We in particular investigate the use of iterative methods of Krylov-subspace type in combination with tailored preconditioners. For this we require a thourough numerical analysis combined with modern software. Additionally, we are interested in reducing the storage complexity of time-dependent PDEs and optimization problems with PDE constraints. Recently, we have focused our efforts on the solution of so-called phase-field problems, fractional differential equations, and problems from uncertainty quantification.

Current PhD students in the IMPRS program

Kirandeep Kour

Website

PhD project: Tensor methods for machine learning and big data applications

IMPRS-Alumni