“…Several kinds of surrogate meta-models have been developed, including polynomial response surface (Rajashekhar et al, 1993;Zheng et al, 2000), Kriging meta-modeling (Kaymaz, 2005;Echard et al, 2009;Echard et al, 2010), Gaussian meta-models (Marrel et al, 2009), artificial neural networks (Papadrakakis et al, 1996;Elhewy et al, 2006;Cheng et al, 2008), radial basis functions and support vector machines (Hurtado, 2007;Li et al, 2010). Even if Monte Carlo methods are universal, they are time computing expansive and sometimes impossible to use.…”