Modeling Model Predictive Control: A Category Theoretic Framework for Multistage Control Problems

arXiv2305.03820MaRDI QIDQ6509975FDOQ6509975

Authors: Tyler Hanks, Baike She, Matthew Hale, Evan Patterson, Matthew Klawonn, James P. Fairbanks 

Abstract: Model predictive control (MPC) is an optimal control technique which involves solving a sequence of constrained optimization problems across a given time horizon. In this paper we show that the structure of problems in this sequence is elegantly captured via a novel category-theoretic viewpoint. Specifically, we develop a category - called Para(Conv) - whose objects are Euclidean spaces and whose morphisms represent constrained convex optimization problems. We go on to show that convex MPC optimization problems result from compositions of simpler subproblems in Para(Conv), and we also show that Para(Conv) can be used to model both initial and terminal state constraints, which often arise in practice. We present a novel Julia library that leverages our theoretical results to automate the implementation of correct-by-construction MPC problems in software. Specifically we provide code snippets showing the ease with which MPC problems can be implemented and solved using our library.

This page was built for publication: Modeling Model Predictive Control: A Category Theoretic Framework for Multistage Control Problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6509975)