Performs automated transport separation of transport-dominated systems with neural networks

Frequency domain reduced order model (ROM) for second-order systems with frequency dependent matrices

individual convolutional autoencoders (iCAEs) for low-dimensional parametrization

CARDIOKOOP - Control-aware Koopman deep learning framework for real-time hemodynamic forecasting and cardiovascular digital twin applications.

Frequency domain reduced order model (ROM) for second-order systems with frequency dependent matrices

Implementation of balanced truncation method for semi-discretized heat equation, boundary control, Neumann observation. In 2018.

Dimensionality reduction using CUR decomposition

Codes for a research project about a robust high order model-order reduction methods for large scale ODEs systems.

Model reduction for port-Hamiltonian and passive systems in Julia

Optimal control for a class of linear transport dominated systems via the shifted proper orthogonal decomposition

Repository that implements the neural network shifted POD

Python implementations of structure-preserving solvers for constrained mechanical systems.

Solving a non-linear elliptic problem via POD, PINN and POD-nn approaches

Vanishing Viscosity solution predicting Graph Neural Networks and Domain Decomposable Reduced Order Models based on the Discontinuous Galerking method applied to Friedrichs' systems

Uncertainty Quantification for the cgDNA framework

This repository implements a new model order reduction approach for moving fronts.

For full documentation, see:

Code for Master thesis "Model order reduction using group convolutional autoencoders", University of Münster. Supervision by Prof. Dr. Mario Ohlberger and Dr. Stephan Rave.