Arash Bahari Kordabad
|
Arash Bahari Kordabad |
News
|
Research Interests
My research interests primarily lie at the intersection of Markov Decision Processes (MDPs), Economic Model Predictive Control (EMPC), and Reinforcement Learning (RL), with applications in energy systems and autonomous vehicles. Another key focus of my work is providing probabilistic guarantees for stochastic systems under temporal logic specifications. Additionally, my research encompasses distributionally robust optimization, second-order RL algorithms, and data-driven MPC.
Education & Research Experience
|
Max Planck Institute for Software Systems, Kaiserslautern, Germany Postdoctoral Researcher May 2023 - present |
|
Topic: "Multi-agent awareness and control with temporal logic specifications" Supervisor: Prof. Sadegh Soudjani Project: SymAware |
|
|
Norwegian University of Science and Technology (NTNU), Trondheim,
Norway Ph.D., Department of Engineering Cybernetics Feb 2020 - April 2023 |
|
Thesis Topic: "Theoretical properties of learning-based MPC", link Supervisor: Prof. Sébastien Gros Co-Supervisor: Prof. Anastasios Lekkas Committee: Prof. Ole Morten Aamo, Prof. Lars Grüne, and Prof. Rolf Findeisen |
|
|
Aalborg University, Aalborg, Denmark Visiting Ph.D. Researcher, Department of Electronic Systems Nov 2021 - Aug 2022 |
|
Research topic: Safe Reinforcement Learning Host: Prof. Rafal Wisniewski |
|
|
Sharif University of Technology, Tehran, Iran M.Sc., Department of Mechanical Engineering Sep 2017 - Sep 2019 |
|
Thesis Topic: "Control of bifurcation and chatter suppression in peripheral milling process" Supervisor: Prof. Hamed Moradi GPA: 19.41/20 |
|
|
University of Tabriz, Tabriz, Iran B.Sc., Department of Mechanical Engineering Sep 2013 - Sep 2017 |
|
Thesis Topic: "On the muscle models as viscoelastic material and comparison of force-length models for active skeletal muscle" Supervisor: Prof. Kamal Jahani GPA: 18.1/20 |
|
Selected Publications (Top 5)
- A. Bahari Kordabad, M. Zanon, S. Gros, "Equivalence of optimality criteria for Markov decision process and model predictive control", IEEE Transactions on Automatic Control, 2024.
- A. Bahari Kordabad, M. Charitidou, D. V. Dimarogonas, and S. Soudjani, "Control barrier functions for stochastic systems under signal temporal logic tasks", 22nd European Control Conference (ECC), 2024.
- A. Bahari Kordabad, D. Reinhardt, A. S. Anand, and S. Gros, "Reinforcement Learning for MPC fundamentals and current challenges", 22nd IFAC World Congress, 2023.
- A. Bahari Kordabad, S. Gros, "Q-Learning of the storage function in economic nonlinear model predictive control", Engineering Applications of Artificial Intelligence, 2022.
- A. Bahari Kordabad, H. Nejatbakhsh Esfahani, W. Cai, and S. Gros, "Quasi-Newton iteration in deterministic policy gradient", American Control Conference (ACC), 2022.
Selected Talks
- "Reinforcement Learning for MPC: Fundamentals and Current Challenges" in an invited session entitled Recent Advances in Automated Learning and Calibration of MPC Policies, IFAC, +75 audience, Japan. (photo) July 2023
- "Introduction to optimization with temporal logic", NTNU, Trondheim, Norway. Mar 2023
- "Intersection of Reinforcement Learning and MPC", Eindhoven University of Technology, Netherlands, Host: Prof. Dinesh Krishnamoorthy. Feb 2023
- "MPC-based Reinforcement learning", École polytechnique fédérale de Lausanne (EPFL), Lausanne, Switzerland, Host: Prof. Alireza Karimi. Jan 2023
- "Optimality Equivalence of MDP and MPC", KTH, Stockholm, Sweden, Host: Prof. Bo Wahlberg. Jan 2023
- "Intelligent control for time-delay systems", KTH, Stockholm, Sweden, Host: Prof. Håkan Hjalmarsson. Jun 2019