Abstract
Burn-in tests help manufacturers detect defective items and remove them before being sold to customers. In a competitive marketplace, cost is a major consideration and not employing a burn-in test may result in higher and needless expenses. With this in mind, we consider degradation-based burn-in tests in which the degradation path follows a Wiener process and weak items are identified when the process crosses a piecewise linear function. We also study linear functions as a special case of such a piecewise linear barrier. Within this setup, we apply a cost model to determine the optimal burn-in test. Finally, we discuss an illustrative example using GaAs laser degradation data and present an optimal burn-in test for it.