Degradation tests are widely used to assess the reliability of highly reliable products which are not likely to fail under traditional life tests or accelerated life tests (ALT). However, for some highly reliable products, the degradation may be very slow, and thus it seems impossible to have a precise assessment within a reasonable test time. In such cases, an alternative technique is to use higher stresses to extrapolate the product’s reliability at the normal use stress. This is called an accelerated degradation test (ADT). In this article, motivated by a LEDs data, we discuss the optimal allocation problem under accelerated degradation experiment when a Wiener process is used to describe the product’s degradation path. We derive the Fisher information and the approximate variance of the estimated mean-time-to-failure (MTTF) under normal use. Three optimality criteria are defined and the optimal allocation of test units are determined.Finally, the LEDs data is illustrated to demonstrate the efficiency of the optimal allocation of test units.
Methods and Applications of Statistics in Engineering, Quality Control, and the Physical Sciences, pp.330-343