Adaptive-loop method for non-deterministic design optimization

Abstract

In the last two decades, significant attention has been paid to developing nondeterministic design optimization methodologies: reliability-based design optimization (RBDO), possibility-based design optimization (PBDO), and evidence-based design optimization (EBDO). As a result, numerous methods for non-deterministic design optimization have been proposed. They are categorized as: parallel-loop method, serial-loop method, and single-loop method. It has been reported that each method has its own strong and weak points. The current paper attempts to understand better the various methods, and proposes to develop an integrated framework for non-deterministic design optimization. In short, the integrated design framework adaptively involves different loop methods (deterministic design optimization, parallel-loop method, single-loop method) to maximize numerical efficiency and stability without losing numerical accuracy in the non-deterministic design optimization process. While deterministic design optimization and the single-loop method will improve numerical efficiency at the beginning and end of non-deterministic design optimization, the parallel-loop method maintains good numerical stability in the middle. Thus, the proposed method is referred to as the adaptive-loop method. It will be shown that the adaptive-loop method is applicable for nondeterministic design optimization, such as RBDO, PBDO. Examples are used to demonstrate the effectiveness of the adaptive-loop method in terms of numerical efficiency and stability.