An adaptive decoupling approach for reliability-based design optimization

Reliability-based design optimization (RBDO) is a powerful tool for design optimization when considering the probabilistic characteristics of the design variables. However, its use in practical applications is hindered by the huge computational cost during the evaluation of structure reliability. A number of algorithms have been proposed, such as reliability index approach (RIA), performance measure approach (PMA), decoupled approach and single-loop approach. In this paper, the adaptive decoupling approach (ADA) is developed. It adopts a new update angle strategy and a novel feasibility-checking method to improve the efficiency of probabilistic optimization. The update angle is utilized to reduce the number of performance function evaluations. And the probabilistic constraints are divided into feasible, active and violated categories, and only violated and active probabilistic constraints are evaluated precisely. The computation capability of ADA is demonstrated using five examples which are widely used to test RBDO methods. The comparison results show that ADA has the similar accuracy as existing methods, and it is also very efficient.