Data-driven optimal control via linear programming: boundedness guarantees

no code implementations • 30 Oct 2023 • Lucia Falconi, Andrea Martinelli, John Lygeros

The linear programming (LP) approach is, together with value iteration and policy iteration, one of the three fundamental methods to solve optimal control problems in a dynamic programming setting.

Reconfigurable Plug-and-play Distributed Model Predictive Control for Reference Tracking

no code implementations • 19 Jul 2022 • Ahmed Aboudonia, Andrea Martinelli, Nicolas Hoischen, John Lygeros

In the redesign phase, passivity-based control is used to ensure that asymptotic stability of the network is preserved.

Model Predictive Control

Data-Driven Optimal Control of Affine Systems: A Linear Programming Perspective

no code implementations • 22 Mar 2022 • Andrea Martinelli, Matilde Gargiani, Marina Draskovic, John Lygeros

In this letter, we discuss the problem of optimal control for affine systems in the context of data-driven linear programming.

LEMMA

PAGE-PG: A Simple and Loopless Variance-Reduced Policy Gradient Method with Probabilistic Gradient Estimation

no code implementations • 1 Feb 2022 • Matilde Gargiani, Andrea Zanelli, Andrea Martinelli, Tyler Summers, John Lygeros

A numerical evaluation confirms the competitive performance of our method on classical control tasks.

Policy Gradient Methods

Passivity-based Decentralized Control for Discrete-time Large-scale Systems

no code implementations • 15 Jul 2021 • Ahmed Aboudonia, Andrea Martinelli, John Lygeros

The controller is synthesized by locally solving a semidefinite program offline for each subsystem in a decentralized fashion.