Acceptance Sampling Plan of Accelerated Life Testing for Exponential Distribution under Time-Censoring

This paper develops the design of accelerated life testing sampling plans (ALSPs) for exponential distribution under time-censoring. With the consideration of both producer and consumer risks, according to the case of whether AF is known or not, two accelerated LSPs are designed respectively. When AF is known, the life testing is assumed to be conducted at an accelerated condition under time-censoring, the producer and consumer risks are satisfied, the size of testing sample and acceptance number are optimized. Then, sensitivity analyses are conducted to evaluate the influence which the uncertainty in the assumed AF brings on the actual producer and consumer risks. When AF is unknown, two levels of accelerated stress are used, and the sample sizes, sample proportion allocated to each stress level are optimized. The acceptability constant which satisfies the producer and consumer risks is obtained by minimizing the generalized asymptotic variance of the test statistics. Finally, the properties of the two ALSPs are investigated respectively.