Reproduce “Certified Neural Approximation of Robust Nonlinear Model Predictive Control” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hendrik Alsmeier, Maik Pfefferkorn, Rolf Findeisen
The authors replace robust MPC with a fixed neural controller, designing the MPC to tolerate bounded approximation error. Under coverage and regularity conditions, the method retains feasibility and stability guarantees. In a four-state reactor simulation, mean evaluation time fell from 20.73 ms to 0.59 ms, about 35×, with no observed violations; evidence is limited to simulation and stated assumptions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.