Reliability-based design optimization (RBDO) has attracted considerable attention in the past decades to optimize engineered systems and satisfy reliability requirements in design. However, RBDO under high-dimensional uncertainty is hindered by its huge computational burden. In this paper, we tackle this problem by adopting a recently developed high-dimensional reliability analysis (HDRA) method. The HDRA method optimally combines the strengths of univariate dimension reduction (UDR) and Kriging-based reliability analysis, which achieves satisfactory accuracy and efficiency for high-dimensional reliability analysis problems with strong variate interactions and correlations. The computational efficiency of high-dimensional RBDO is improved by pursuing two new strategies: (i) newly selected samples are updated for all the constraints during the sequential sampling process in HDRA; and (ii) a two-stage surrogate modeling strategy is adopted to first locate a highly probable region of the optimum design and then locally refine the accuracy of the surrogates in this region. Results of two mathematical examples show that the proposed HDRA-based RBDO (RBDO-HDRA) method produces higher accuracy and comparable efficiency than the IJDR-based RBDO (RBDO-UDR) method and the ordinary Kriging-based RBDO (RBDO-Kriging) method. The better performance can be attributed to the capability of RBDO-HDRA to handle both high dimension and strong interactions among variables.
展开▼