Chordality for Model Predictive Control

In this presentation we will show how chordality of graphs can be used to improve performance of optimization solvers for model predictive control. This performance improvement is possible to achieve since the problem is what is called partially separable. We will first cover the standard model predictive control problem. Then we will discuss extensions to stochastic formulations using scenarios, and to control of distributed systems. Finally, we will discuss variable horizon model predictive control will applications to rendezvous maneuvers for a drone and a boat.

Download Slides