DATP-based sequential optimization and reliability assessment for RBDO

Sequential optimization and reliability assessment (SORA) has been widely used in reliability-based design optimization (RBDO), but it is hindered by the unaffordable computational burden of high dimensional cases. In this work, we propose a new strategy to improve the efficiency of the SORA while performing high dimensional RBDO. The dimension-adaptive tensor-product (DATP) algorithm and hierarchical interpolation scheme are introduced to replace the complicated black-box performance function with an approximate model, where the DATP is to overcome the so-called curse of dimension and the hierarchical interpolation scheme is to save the cost of function evaluations. Moreover, the hybrid mean value (HMV) method is adopted for the reliability assessment in performance measure approach (PMA). A two dimensional example and a ten dimensional example are included to verify the computational capability of the proposed method.