Advances in Control and Modeling of Complex Systems

The field of control and modeling of complex systems is rapidly evolving, with a focus on developing innovative methods to address the challenges posed by nonlinear dynamics, parameter variations, and high-dimensional data. Recent research has explored the use of neural networks, reduced-order modeling, and operator learning to improve the accuracy and efficiency of control and modeling techniques. Notably, the integration of hyper neural networks into predictive control frameworks has shown promise in capturing system nonlinearities and parameter variations. Additionally, the development of adaptive reduced basis trust region methods and neural co-state regulators has enabled more efficient and accurate solutions to inverse problems and optimal control problems. Noteworthy papers include: The Neural Parameter-varying Data-enabled Predictive Control framework, which adaptively captures system nonlinearities and parameter variations. The Reduced-Order Neural Operator Modeling framework, which bridges concepts from reduced-order modeling and operator learning to achieve improved discretization convergence and robustness.

Sources

Neural Parameter-varying Data-enabled Predictive Control of Cold Atmospheric Pressure Plasma Jets

Controllable Patching for Compute-Adaptive Surrogate Modeling of Partial Differential Equations

Some Super-approximation Rates of ReLU Neural Networks for Korobov Functions

Adaptive Reduced Basis Trust Region Methods for Parabolic Inverse Problems

Neural Co-state Regulator: A Data-Driven Paradigm for Real-time Optimal Control with Input Constraints

Deep Bilinear Koopman Model for Real-Time Vehicle Control in Frenet Frame

RONOM: Reduced-Order Neural Operator Modeling

Vertical Vibration Reduction of Maglev Vehicles using Nonlinear MPC

Built with on top of