Directory of Experts

You are here

Back to search results
Sébastien Le Digabel
Dipl. Ing. (ISIMA, 1999), M.Sc. (Poly, 2002), Ph.D. (Poly, 2008)

Research interests and affiliations

Research interests My research focuses on the design of algorithms for the optimization of complex applications usually encountered in engineering. These problems are typically defined by functions viewed as blackboxes in the sense that no property of the function is available. In this context, usual optimization methods based on derivatives are prohibited, hence the choice of derivative-free optimization algorithms and in particular the use of models and surrogates less expensive to evaluate. My projects focus on three areas:
  1. Development of optimization methods
  2. Application to engineering problems
  3. Development of optimization software

Keywords: blackbox optimization, derivative-free methods, industrial applications, optimization software.


  • Head of the Review Committee of Frauds
  • Member of the sub-committe for undergraduate studies
Expertise type(s) (NSERC subjects)
  • 1601 Operations research and management science
  • 2715 Optimization
  • 2956 Optimization and optimal control theory


Recent publications
Journal article
de Sousa, V.J.R., Anjos, M.F. & Le Digabel, S. (2017). Computational study of valid inequalities for the maximum k-cut problem. Annals of Operations Research, 23 pages. Retrieved from
Journal article
Audet, C., Kokkolaras, M., Le Digabel, S. & Talgorn, B. (2017). Order-based error for managing ensembles of surrogates in mesh adaptive direct search. Journal of Global Optimization, 31 pages. Retrieved from


Sébastien Le Digabel is an Associate Professor of Mathematics at the Ecole Polytechnique in Montreal, and a regular member of the GERAD research center. Before that, he obtained a Ph.D. in applied mathematics from the Ecole Polytechnique in 2008, and worked as a postdoctoral fellow at the IBM Watson Research Center and the University of Chicago in 2010 and 2011.

His research interests include the analysis and development of algorithms for blackbox optimization, and the design of related software. Blackbox optimization occurs when the functions to optimize are given by numerical simulations for which derivative information is not available. In this context, derivative-free optimization may be considered, and in particular the Mesh Adaptive Direct Search (MADS) method of Audet and Dennis, for which Le Digabel's thesis brought some extensions and upgrades. All of his work on MADS is included in the NOMAD software, a free package for blackbox optimization available at

S. Le Digabel's research is funded by AFOSR, the Canadian NSERC foundation, and the Quebec FRQNT fund.

Supervision at Polytechnique


  • Ph.D. (6)

    • Amaioua, Nadir. Amélioration des méthodes quadratiques et de la parallélisation pour l'optimisation des boîtes noires.
    • Bingane, Christian. Optimal Power Flow: Semidefinite Programming Approach.
    • De Souza Dutra, Michael David. Optimisation de la gestion de l'énergie des maisons intelligentes.
    • Dzahini, Kwassi Joseph.
    • Lakhmiri, Dounia.
    • Rodrigues De Sousa, Vilmar Jefte. Solving Large-Scale Maximum K-Cut Problem.


  • Ph.D. Thesis (1)

  • Master's Thesis (4)

    • Lemyre Garneau, M. (2015). Modelling of a Solar Thermal Power Plant for Benchmarking Blackbox Optimization Solvers (Master's Thesis, École Polytechnique de Montréal). Retrieved from
    • Ihaddadene, A. (2014). Algorithme de recherche directe pour l'optimisation robuste de fonctions bruitées (Master's Thesis, École Polytechnique de Montréal). Retrieved from
    • Cartier, D. (2012). Optimisation sous contraintes d'un modèle hydrologique pour une représentation de la physique des processus (Master's Thesis, École Polytechnique de Montréal). Retrieved from
    • Duclos, E. (2012). ACRE: un générateur automatique d'aspect pour tester des logiciels écrits en C++ (Master's Thesis, École Polytechnique de Montréal). Retrieved from