Sébastien Verel (professeur, Université du Littoral Côte d’Opale, LISIC) a animé un Seminar@SystemX en webinaire sur le thème « Optimization for simulation based problems: approaches and examples in mobility problems, and in nuclear energy systemsapproche systémique », le 17 février 2022 de 14h à 15h30.
Résumé :
Digital twins are now common in sciences, and in industries. They have been used to model ecological systems, industrial systems, or to design new objects or services, etc. Digital twins allows to understand, and to test hypotheses on inaccessible systems, or costly real-life systems. As a consequence, the artificial intelligence approaches, and in particular optimization methods can be used on those numerical simulations in order to test, and design more efficient systems which optimize different criteria of interest. From the point of view of optimization methods, such simulation based problems are black-box with no available gradient, or even analytical definition of problems, and moreover, one numerical simulation can be time consuming to compute. To face the challenges, different approaches can be used. This seminar will shows different techniques in this context such as surrogate models, parallel computation, multiobjective optimization with adaptive local search, or evolutionary algorithms. To illustrate the techniques, two examples will be exposed: the nuclear power plant pilotage to introduce intermittent green energies, and in urban mobility to increase the quality of the mobility in a city.
Biographie :
Sébastien Verel is a professor in Computer Science at the Université du Littoral Côte d’Opale, Calais, France, since 2013, and previously at the University of Nice Sophia-Antipolis, France, from 2006 to 2013. He is the director of LISIC lab from 2020. His research interests are in the theory of evolutionary computation, multiobjective optimization, adaptive search, and complex systems. A large part of his research is related to understand optimization algorithms to automatically adapt algorithms to new optimization problems, and applications of evolutionary computation for numerical simulation.