OPTIMUM STEP STRESS ACCELERATED DEGRADATION TEST USING GOAL PROGRAMMING

Step stress accelerated degradation test (SSADT) is an important technical strategy for estimating the long term performance of high reliability products. Much of the previous works designed optimal SSADT by maximizing estimation precision under the constraint that experimental cost does not exceed a predetermined budget without any restrictions on the number of measurements. However, a compromise between precision and cost is still a problem to solve. Besides, fixing measurement frequencies under different stress levels of accelerated life test was another drawback in previous studies that was dealt with in this study. This paper, therefore, presents a goal programming model that minimizing weighted sum of asymptotic variance of mean time to failure and also the experimental cost. Some design variables as sample size, measurement frequencies and number of measurements under different stress levels are optimized under the constraint of increasing number of measurements with higher stress levels. For illustration, a comparative study is conducted between proposed model and previous studies. The sensitivity of the SSADT plan is studied, showing that the proposed optimal test plan is robust to small departure from the values of the parameters.