Accelerated life tests(ALTs) and accelerated degradation tests(ADTs) are frequently used to quickly assess the reliability of components or materials. Many authors have developed ALT plans under various assumptions on the failure time distribution, stress loading method and inspection method. Unlike the case of ALTs, however, little work has been conducted on designing ADTs. Moreover, most previous researches on ADT plans are concerned with the case of a single stress. In this thesis, optimal ADT plans with two stress variables are developed in the case of destructive measurement. The performance characteristic is assumed to follow a lognormal distribution, and the relationship between the elapsed time and the mean of the log-transformed performance characteristic is assumed to follow a simple constant rate relationship with the degradation rate depending on the stress variables according to the Eyring law. The stress levels and the proportion of test units allocated to each stress level are determined such that the asymptotic variance of the maximum likelihood estimator of the q th quantile of the lifetime distribution at the use condition is minimized. Although the maximum number of test points is allowed to be five, certain optimal plans require fewer than five test points and various patterns of optimal plans are identified. They include 2-point test, 3-point test and 5-point test. Computational Results are also tabulated to make them readily available to the experimenters.
