Failure Probability Estimation Using Asymptotic Sampling and Its Dependence upon the Selected Sampling Scheme

The article examines the use of Asymptotic Sampling (AS) for the estimation of failure probability. The AS algorithm requires samples of multidimensional Gaussian random vectors, which may be obtained by many alternative means that influence the performance of the AS method. Several reliability problems (test functions) have been selected in order to test AS with various sampling schemes: (i) Monte Carlo designs; (ii) LHS designs optimized using the Periodic Audze-Eglājs (PAE) criterion; (iii) designs prepared using Sobol’ sequences. All results are compared with the exact failure probability value