“…To overcome the computational problems, an alternative strategy of constructing approximate models (also known as surrogate models or response surface models) to replace the expensive FEM simulation in the optimization process has been extensively used in recent years. A number of methods have been successfully applied to build the global/local approximation models for postbuckling anal-ysis of composite structures, for example, the polynomial regression function [8], Articial Neural Networks [9,10], Radial Basis Functions [10] and Kriging method [9]. The minimum weight design of composite structural components (stiened panels and shells) considering their postbuckling response was carried out using these tools or approximate models.…”