Аннотация:After a brief review of properties of the high-dimensional standard normal space, the orthogonal plane sampling (OPS) method is investigated in the context of the high-dimensional structural component reliability problem. A geometric interpretation of the OPS in high dimensions is developed. Using this interpretation, a modified OPS approach is proposed by fixing the sampling space to a hypersphere and picking highly regular sample points. Comparisons with results obtained with Monte Carlo simulation for several numerical examples illustrate the accuracy and efficiency of the modified OPS relative to the original OPS. It is shown that the modified OPS is at least as accurate as the conventional OPS, but generally much more efficient.
