Outline

In Episode 11, we interviewed Sébastien Le Digabel about NOMAD a blackbox optimization software. A blackbox is a system which can be viewed in terms of its inputs and outputs, without any knowledge of its internal workings. NOMAD is a software for the optimization of such problems. It implements the Mesh Adaptive Direct Search (MADS) derivative-free optimization algorithm. NOMAD is free and intended to be easy to use. It is designed for solve real-world optimization problems from the industry. It works out of the box, as long as the objective and constraints are provided.

About Sébastien Le Digabel:

Sébastien Le Digabel received the M.Sc.A. degree and the Ph.D. degree in applied mathematics from Polytechnique Montreal, Montreal, Quebec, Canada in 2002 and 2008 respectively, and worked as a postdoctoral fellow at the IBM Watson Research Center and the University of Chicago in 2010 and 2011. He is currently an Associate Professor in the Department of Mathematics and Industrial Engineering at Polytechnique Montreal and a regular member of the GERAD research center. His research interests include the analysis and development of algorithms for derivative-free and blackbox optimization, and the design of related software. All of his work on derivative-free optimization is included in the NOMAD software, a free package for blackbox optimization.