Model predictive control (MPC) for linear systems with quadratic costs and linear constraints is shown to admit an exact representation as an implicit neural network. A method to "unravel" the implicit neural network of MPC into an explicit one is also introduced. As well as building links between model-based and data-driven control, these results emphasize the capability of implicit neural networks for representing solutions of optimisation problems, as such problems are themselves implicitly defined functions.
@article{arxiv.2404.12030,
title = {Mapping back and forth between model predictive control and neural networks},
author = {Ross Drummond and Pablo R Baldivieso-Monasterios and Giorgio Valmorbida},
journal= {arXiv preprint arXiv:2404.12030},
year = {2024}
}