“…The surrogate model is typically created by evaluating the response of the real model at several points, called design points/sites. Types of metamodels that have been employed in the past with regards to structural reliability analysis include polynomials [6,10,17,18,19], Kriging [20,21,22], and Neural Networks (NNs) [23,24]. Although surrogate models are not as accurate as the real model over the entire domain being investigated, they can act as very effective substitutes for the real model over a small part of the domain.…”