A new method is proposed to tackle the huge computation cost involved in Successive Response Surface Methodology applied to the reliability analysis, in which Space Mapping technique is combined with Response Surface Methodology. While the new approach is performed, the limit state function is only fitted at the first iteration; at other iterations Space Mapping technique is employed to map the original limit state function into the new ones. Experimental design, corresponding model evaluations and response surface fitting of the limit state function are not done repetitively as what we do while SRSM is used, which leads to the great cutting down of computational efforts.
Reduced Sampling for Construction of Quadratic Response Surface Approximations Using Adaptive Experimental Design
