Hybrid Modeling, Simulation & Optimization (HybridMSO)

The Hybrid Modeling, Simulation, and Optimization (HybridMSO) group
researches the exciting convergence of knowledge-based and data-driven
modeling approaches. Their work is centered on accelerating process
development and improving online optimization, particularly through the
creation of digital twins for conventional and modular process plants using
hybrid modeling. Importantly, their approach significantly reduces the data
requirements for model development compared to purely data-driven
methods. They specialize in capturing complex and hard-to-model phenomena
through data-driven techniques and contribute to the development of design
of experiments methodologies for process and model enhancement.

Pyramid with methods, lectures and projects of the HybridMSO group

Research focus

How can knowledge-based and data-driven modeling approaches be
optimally combined?
How should optimal experiments for creating hybrid knowledge-based and
data-driven models be designed?
How can hybrid models be efficiently identified?
How can knowledge in the form of behavioral models be efficiently shared
to account for the distribution of knowledge in the process industry?
How can prior knowledge in the form of information models be
automatically incorporated into the development of knowledge-based and
data-driven behavioral models?