Andrea Pesare

Pubblicazioni

Titolo Pubblicato in Anno
Online identification and control of PDEs via reinforcement learning methods ADVANCES IN COMPUTATIONAL MATHEMATICS 2024
A New Algorithm for the LQR Problem with Partially Unknown Dynamics Large-Scale Scientific Computing 2022
Convergence results for an averaged LQR problem with applications to reinforcement learning MATHEMATICS OF CONTROL SIGNALS AND SYSTEMS 2021

ERC

  • PE1_18
  • PE1_20

Interessi di ricerca

I am working on a mathematical framework for some probabilistic model-based Reinforcement Learning methods. Our model is based on optimal control problems with uncertain dynamics, where the cost is averaged over a family of possible dynamics. Other research interests include numerical methods for optimal control problems, numerical methods for Hamilton-Jacobi equations, 3D reconstruction and Photometric Stereo.

Keywords

Optimal control
Reinforcement Learning
numerical methods
uncertain systems
3D reconstruction

© Università degli Studi di Roma "La Sapienza" - Piazzale Aldo Moro 5, 00185 Roma