The Research Unit of Scientific Computing and Modelling consists of the following 6 workgroups:

The scientific aim of this research unit is to model and to simulate problems in the natural sciences, in engineering disciplines, and in industry. In many applications, this results in the task to develop numerical solvers for coupled systems of deterministic and stochastic partial differential equations and algebraic equations. To ensure efficiency, the solvers are adapted to modern computer architectures, while keeping them flexible enough to handle new applications and hardware.

In applications, model parameters are often unknown and must be determined reliably by comparison with measurements. To this end, modern methods from uncertainy quantification are put on theoretic foundations and applied to calculate probability distributions of the unknown parameters, which has important advantages compared to more commonly used methods.

Furthermore, machine-learning algorithms are developed and used to extract knowledge from data and simulations and to solve optimization problems in selected applications including: