Public Lecture in English

This lecture shows how Artificial Neural Networks can be used to significantly simplify the computationally intensive calculation of the Boltzmann collision operator in simulations. The focus is on how such networks can be designed to accurately reproduce important physical properties – in particular, the entropy decay as described by the second law of thermodynamics.

It also explains how training data can be selected effectively. Finally, numerical results demonstrate how well and reliably the presented approaches work.

Speaker:

Prof. Dr. Martin Frank of the Karlsruhe Institute of Technology (KIT)