Research

Follow me on Google Scholar to be informed of my new publications.

Paper and proceedings

• S. Colla and J. M. Hendrickx, “Exploiting Agent Symmetries for Performance Analysis of Distributed Optimization Methods”, submitted to Open Journal of Mathematical Optimization, 2024.

  [PDF, arXiv ]

• S. Colla and J. M. Hendrickx, “On the Optimal Communication Weights in Distributed Optimization Algorithms”, MTNS 2024.

  [PDF, arXiv ]

• A. Rubbens, N. Bousselmi, S. Colla and J. M. Hendrickx, “Interpolation Constraints for Computing Worst-Case Bounds in Performance Estimation Problems”, tutorial paper for 62th IEEE Conference on Decision and Control (CDC), 2023.

  [PDF, arXiv, DOI ]

• S. Colla, J. M. Hendrickx, “Automatic Performance Estimation for Decentralized Optimization”, IEEE Transactions on Automatic Control, accepted, 2023.

  [PDF, arXiv, DOI ]

• S. Colla, J. M. Hendrickx, “Automated Performance Estimation for Decentralized Optimization via Network Size Independent Problems”, in 61th IEEE Conference on Decision and Control (CDC), 2022.

  [PDF, arXiv, DOI, slides ]

• S. Colla, J. M. Hendrickx, “Automated Worst-Case Performance Analysis of Decentralized Gradient Descent”, in 60th IEEE Conference on Decision and Control (CDC), 2021.

  [PDF, arXiv, DOI, slides, video]

PhD Thesis

• S. Colla, “Computer-Aided Analysis of Decentralized Optimization Methods”, PhD thesis, UCLouvain 2024 [DOI ]

Talks

• 43th Benelux Meeting on Systems and Control, March 2024 in Blankenberge (Belgium).

  What are the Optimal Communication Weights for Decentralized Optimization?

  [abstract, slides]

• 20th EUROpt Workshop, August 2023 in Budapest (Hungary).

  Exploiting Agent Symmetries for Automatic Performance Analysis of Distributed Optimization Methods. [slides]

• A Size Independent Formulation for Automatic Performance Evaluation of Decentralized Optimization Algorithms

  • 42th Benelux Meeting on Systems and Control, March 2023 in Elspeet (Netherlands). [abstract]
  • PEP-talks, February 2023 in Louvain-la-Neuve (Belgium). [slides, video]

• Automatic Performance Estimation for Decentralized Optimization

  • ECE Miami University (Graduate Seminar), October 2022 at Oxford (OH, USA). [slides]
  • ORC MIT, September 2022 at Boston (MA, USA). [slides]
  • SeLMA seminar, May 2022 at KULeuven (Belgium). [slides]

• 41th Benelux Meeting on Systems and Control, July 2022 in Brussels (Belgium).

  Performance Estimation Problem for Decentralized Optimization Methods

  [abstract, slides]

• 40th Benelux Meeting on Systems and Control, June 2021 in Rotterdam (Netherlands).

  Automated Worst-Case Performance Analysis of Decentralized Optimization Methods

  [abstract, slides]

• European Control Conference 2021 (ECC21), June 2021.

  Decentralized Estimation in Open Multi-Agent Systems

  [extended abstract, slides]

Poster

• 7th International Conference on Continuous Optimization (ICCOPT), July 2022 in Bethlehem (USA).

  Automatic Performance Estimation for Decentralized Optimization

  [PDF]

Toolbox

The PESTO toolbox allows to easily formulate and solve Performance Estimation Problems (PEP) in order to automatically find the worst-case performance of optimization methods. I have contributed to this toolbox by implementing my recent results on PEP for decentralized optimization. This allows anyone to easily analyze worst-case performance of many decentralized algorithms.

Master Thesis

• 19. Colla, “Data fusion in Open Multi-Agent Systems for decentralized estimation”, École polytechnique de Louvain, Université catholique de Louvain, 2020. (Advisor: Julien M. Hendrickx) [PDF].