Data-driven Distributionally Robust MPC
 |  |
|---|
Data-driven distributionally robust MPC (DRMPC) is an optimal control scheme, which - instead of considering only one distribution as in SMPC - determines control actions with respect to the worst-case distribution from a set of distributions (the set is known as ambiguity set). The research mainly focuses on:
- Exact reformulations of underlying distributionally robust optimisation problems for the purpose of computational efficiency.
- Proofs of stability and convergence.
- Propagation of ambiguity sets in dynamical systems.
Nonlinear Wasserstein Distributionally Robust Optimal Control
Zhengang Zhong
Data-Driven Distributionally Robust Control
Zhengang is a PhD candidate at Imperial College London. His PhD research is about data-driven distributionally robust model predictive control.Prior to his PhD research, he holds a Diplom-Ingenieur in Mechatronics from Technische Universität Dresden. His research interests include data-driven optimal control, scientific machine learning, and distributionally robust optimisation.