Efficient Global Optimization (EGO) is an optimization strategy based on approximating functions, namely Gaussian process models. We show the application of this technique to a model calibration problem referred to a geomechanical application. By means of the approximating function an objective relevance ranking among the problem parameters can be produced, offering valuable and reusable information on the physical problem.
Olga Fink, Chao Hu, Sayan Ghosh
Dominique Bonvin, Alejandro Gabriel Marchetti, Sean Costello