Advances in Reduced Order Modeling and System Identification

The field of reduced order modeling and system identification is witnessing significant developments, driven by the increasing need for efficient and accurate modeling of complex systems. Researchers are exploring innovative methods to improve the accuracy and interpretability of reduced order models, including the use of machine learning techniques and data-driven approaches. A key direction in this area is the development of non-intrusive reduced order modeling techniques, which can handle high-dimensional systems and provide interpretable results. Another important trend is the application of system identification methods to real-world problems, such as modeling the dynamics of buck converters and predicting the behavior of robot appendages interacting with granular materials. Notable papers in this area include the development of a data-driven approach to enhance the accuracy of non-intrusive Reduced Order Models using Multi-Input Operators Network, and the application of Kolmogorov-Arnold Networks to model and analyze the dynamics of a buck converter system. The use of data-driven moment matching methods for parametric reduced-order models is also a promising direction, as demonstrated by its application to linear and nonlinear parametric systems.

Sources

2N-storage Runge-Kutta methods: c-reflection symmetry and factorization of the Butcher tableau

Machine Learning-based quadratic closures for non-intrusive Reduced Order Models

Data-driven balanced truncation for second-order systems with generalized proportional damping

Interpretable and flexible non-intrusive reduced-order models using reproducing kernel Hilbert spaces

System Identification Using Kolmogorov-Arnold Networks: A Case Study on Buck Converters

Data-Driven Model Reduction by Moment Matching for Linear and Nonlinear Parametric Systems

Data-Driven Prediction of Dynamic Interactions Between Robot Appendage and Granular Material

Built with on top of