Identification of expensive-to-simulate parametric models using Kriging and Stepwise Uncertainty Reduction

Julien Villemonteix (1) , Emmanuel Vazquez (1) , Eric Walter (2)

1 SUPELEC-Campus Gif
2 L2S - Laboratoire des signaux et systèmes

This paper deals with parameter identification for expensive-to-simulate models, and presents a new strategy to address the resulting optimization problem in a context where the budget for simulations is severely limited. Based on Kriging, this approach computes an approximation of the probability distribution of the optimal parameter vector, and selects the next simulation to be conducted so as to optimally reduce the entropy of this distribution. A continuous-time state-space model is used to illustrate the method.

Domaines: Automatique / Robotique

Julien Villemonteix, Emmanuel Vazquez, Eric Walter. Identification of expensive-to-simulate parametric models using Kriging and Stepwise Uncertainty Reduction. Conference on Decision and Control, Dec 2007, New Orleans, United States. pp.5505-5510, ⟨10.1109/CDC.2007.4434190⟩. ⟨hal-00252148⟩