A probability feasible region enhanced important boundary sampling method for reliability-based design optimization

Reliability-based design optimization (RBDO) is powerful for probabilistic constraint problems. Metamodeling is usually used in RBDO to reduce the computational cost. Kriging model-based RBDO is very suitable to solve engineering problems with implicit constraint functions. However, the efficiency and accuracy of the kriging model constrain its use in RBDO. In this research, the importance boundary sampling (IBS) method is enhanced by the probability feasible region (PFR) method to fit kriging model with high accuracy. The proposed probability feasible region enhanced importance boundary sampling (PFRE-IBS) method selects sample points for inactive constraint functions only in its important region, thus reducing the number of sample points to improve the efficiency of sampling method. In order to verify the efficiency and accuracy of the proposed PFRE-IBS method, three RBDO problems are used in this paper. The comparison results with other sampling methods show that the proposed PFRE-IBS method is very efficient and accurate.