Mathieu Besançon ☕️

Mathieu Besançon

Researcher in mathematical optimization

Zuse Institute Berlin

I am a researcher in computational optimization at the Zuse Institute Berlin, in the AI in Society, Science, and Technology department. I am associated with the MODAL-SynLab DFG project and a member of the MATH+ Berlin Mathematics Research Center.

My research interests span the theory, methods, and algorithms in mathematical optimization, either on various generic or specific classes of problems. Slightly more specifically, I have been interested in structured constrained optimization in various settings (see below for a list of topics). Those include solution methods, computational models, and software in MI(N)LP and convex optimization and in particular around the SCIP framework and Frank-Wolfe related approaches.

I graduated with a double PhD (cotutelle) between Polytechnique Montréal, at the GERAD lab and Centrale Lille, at INRIA & the Cristal lab, in mathematical optimization. My thesis focused on bilevel optimization, an extension coined near-optimality robustness, and pricing for demand response in smart grids. It was co-supervised by Luce Brotcorne (INRIA) & Miguel F. Anjos (University of Edinburgh).

I am involved in several open-source projects around optimization and scientific computing in the Julia programming language and around JuMP but like looking around on new development. Before starting the PhD, I worked in various industries, from a hardware startup to steel manufacturing. I did my joint Bachelor-Master in Process Engineering at the UTC in France with a semester at the TUBS in Germany and Polytechnique Montreal.

On a personal note, I read both fiction (mostly history, detective, thrillers and fantasy) and non-fiction books, on economic policy, education, and entrepreneurship, a more detailed list can be found on goodread. I also enjoy games in various formats (tabletop, video, board, card) and cooking.


Postdoctoral Researcher
Zuse Institute Berlin
January 2021 – Present Berlin, Germany
Research in optimization methods and computation.
Doctoral Researcher
Polytechnique Montréal, Inria Lille
September 2017 – December 2020 Montréal, Canada & Lille, France
Double PhD program in mathematical optimization for pricing of Demand Response programs in a smart grid context.
Research Engineer, Data Scientist
Equisense SAS
July 2016 – August 2017 Lille, France
Research and development for a startup building connected devices and associated products for horse-riders.


Filter publications here.

(2023). Cutting Plane Selection with Analytic Centers and Multiregression. CPAIOR 2023.


(2023). Improved local models and new Bell inequalities via Frank-Wolfe algorithms. arXiv.


(2022). Convex mixed-integer optimization with Frank-Wolfe algorithms.


(2022). Flexible Differentiable Optimization via Model Transformations. In minor revision.


(2022). FrankWolfe.jl: A High-Performance and Flexible Toolbox for Frank--Wolfe Algorithms and Conditional Gradients. INFORMS Journal on Computing.



The most reliable way to reach me is per email.