Kriging-Based Reliability-Based Design Optimization Using Single Loop Approach

Reliability-Based Design Optimization (RBDO) is a powerful tool in engineering structural design, it tries to find a balance between cost and reliability for structural designs under uncertainty. Several RBDO formulations are developed to solve the RBDO problem, such as double loop methods, single loop methods, and decoupled methods. Despite, these new formulations of RBDO, they are unable to deal for engineering complex problems, due to the computational cost. The Kriging surrogate has been widely used to replace the time-consuming mechanical constraints. In this paper, a single loop RBDO approach (SLA) is coupled with the Kriging surrogate, the most probable points (MPP) of each loop are used as new sample points to update the Kriging model. The Kriging-SLA is running iteratively until it reaches the converge criteria. Compared with other sampling methods, this method can be started with very few training points and converges to the right minimum very efficiently. 2 benchmark examples are used to demonstrate this method.