Skip to Main Content
We consider a framework with two conflicting objectives in the planning of ALT; i.e., meeting a desired level of statistical precision for an estimate of interest, and meeting a cost target for conducting the test. This can easily be generalized to multiple objectives with, say, precision targets for various estimates, stress limits constraint, sample size constraint, etc. In the absence of actual cost, the cost of conducting the test is related to the total test time at different stress levels, and the total sample size needed for the test. The framework is applied to planning of constant stress ALT, in which the logarithm of the lifetime follows the extreme value distribution. The proposed framework consistently leads to more cost-effective test plans without compromising the precision of the estimate, and is equally robust to misspecification of initial estimates, compared with other test plans presented in the literature.