Reliability design optimization method based on information reconstruction Kriging model

Abstract

An efficient reliability design optimization method based on information reconstruction Kriging model is developed to further improve the computational efficiency. Inspired by the concept of incremental shifting vector, the conventional nested double-level optimization can be decomposed into updated deterministic optimizations and reliability analysis subproblems, which can simplify the reliability design optimization problem. The history iteration information of the reliability analysis is used to reconstruct the Kriging model and the active Kriging method is employed to address the reliability analysis problems efficiently. The most probable points (MPP) and its gradients of the current iteration process for the reliability constraints are obtained approximately to update the deterministic optimizations. Two numerical examples are investigated to demonstrate the effectiveness and efficiency of the proposed method. It is shown that the proposed method can improve the calculation efficiency while satisfying precision. And the method has the characteristics of high precision and moderate calculation when dealing with nonlinear problems.