A0646
Title: Optimal planning for accelerated life tests with progressive stress loadings
Authors: Xiaojian Xu - Brock University (Canada) [presenting]
Abstract: Accelerated life testing is an effective method for reducing testing time by subjecting products to elevated stress levels to obtain sufficient failure data within a shorter period. Optimal design strategies are developed for ALT experiments incorporating progressive stress loading schemes in conjunction with progressive Type-II censoring. To enhance the efficiency of life tests, an approach is proposed for optimizing progressive stress loading rates under various optimality criteria, including c*-optimality, D-optimality, and A-optimality. When the lifetime of a test unit follows an exponential distribution, explicit expressions are derived for the optimal stress loading rates under both D- and A-optimality, applicable to single-group and balanced two-group loading scenarios. For the single-group case under c*-optimality, the optimal loading rate is expressed as a function of a specified failure proportion. It is further extended cases where lifetimes follow a Weibull distribution, supported by numerical results and illustrative examples. Design guidance is provided for scenarios of either increasing or decreasing failure rates, with an emphasis on practical considerations. Notably, the proposed c*-, D-, and A-optimal designs demonstrate high efficiency across all cases examined, especially when failure proportions are strategically chosen during the planning stage.
