The team is working on a methodology of optimization by numerical simulations.
Both the phenomena studied and the modelling tools are the source of many difficulties in the implementation of the optimization. This problem is known to the designers who are faced with delays and costs associated to experimental researches of an acceptable solution. The solution, based on the works of Ficher, Box and Taguchi, is now well established. This is the method of Design of Experiments (Design of Experiments or English) which can be translated in computing science very well under the name of method of Digital Design of Experiments.
The method of Design of Experiments is particularly well suited to optimization problems with a average number of parameters (dozens), an high computational cost for an experience and results subjected to random effects.