Optimal Design of Multi-stress Accelerated Degradation Tests Based on Bootstrap Method

Abstract

Step-stress accelerated degradation tests (SSADTs) have been applied to obtain products’ reliability information when the available test time is limited. As the products performance is generally affected by multiple stresses, optimal design for multiple SSADT is very necessary to guarantee the assessment accuracy and improve the test efficiency. In the current study, a multi-stress step-stress accelerated degradation model is first proposed, which can reasonably depict the comprehensive influence of accelerated stress levels on degradation regulation Then an optimal design method involving a bootstrap-based objective function is constructed, where multiple decision variables involving sample allocation and test time arrangement for each stress level can be simultaneously optimized based on genetic algorithm (GA) and Monte Carlo simulation. Finally, a simulation example is implemented, and a series of conventional methods are compared to illustrate the superiority of the proposed method. The comparative results and sensitivity analysis results both demonstrate the effectiveness and rationality of the proposed method.