4 papers
Rollout Then Optimize: A One-Step Newton Refinement of Learned Policies for Nonlinear Model Predictive Control
Andrea Ghezzi, Rudolf Reiter, Katrin Baumgärtner +2
We propose a computationally efficient rollout-then-optimize method to improve a learned control policy at deployment time. A learned policy provides a nominal trajectory, which is…
A Sequential Benders-based Mixed-Integer Quadratic Programming Algorithm and Its Implementation in the CAMINO Toolbox
Andrea Ghezzi, Wim Van Roy, Sebastian Sager +1
Sequential quadratic programming and sequential convex programming efficiently solve nonlinear programs (NLPs) by linearizing inner nonlinearities while preserving the outer convex…
A Comparative Study of MINLP and MPVC Formulations for Solving Complex Nonlinear Decision-Making Problems in Aerospace Applications
Andrea Ghezzi, Armin NurkanoviÄ, Avishai Weiss +2
High-level decision-making for dynamical systems often involves performance and safety specifications that are activated or deactivated depending on conditions related to the syste…
AC4MPC: Actor-Critic Reinforcement Learning for Nonlinear Model Predictive Control
Rudolf Reiter, Andrea Ghezzi, Katrin Baumgärtner +3
\Ac{MPC} and \ac{RL} are two powerful control strategies with, arguably, complementary advantages. In this work, we show how actor-critic \ac{RL} techniques can be leveraged to imp…