Adaptation of MPC via RL: fundamental principles

Sebastien Gros

NTNU Trondheim

Monday, August 02, 2021, 9:00

virtual, see link below

Reinforcement Learning and MPC have natural connections that we will explore in this lecture. We will then discuss a recent, fundamental result in learning-based MPC, which formally explains how the tuning of MPC should take place in order to achieve optimal closed-loop performance with inaccurate MPC models. Finally, we will discuss some implementation aspects of RL-based MPC.

This talk is part of the summer school on model predictive control and reinforcement learning and openly available.

Join Zoom Meeting

Meeting ID: 654 3915 8240
Passcode: 0uu1tq0ek