Accelerated life testing (ALT) is a commonly used method in high reliability and long life evaluation, and it is adopted widely in life-determination and life-extension of systems and productions. Based on the differences among productions in life and characteristic, we consider simple stress model and multi-stress levels model with Type-â…¡censored data. The exact confidence intervals of the MLEs for parameters are obtained in simple stress model, and third-order confidence intervals inference are obtained in multi-stress levels for the exponentially distributions. Meanwhile, make comparisons of coverage probabilities between normal-approximation intervals and third-order intervals through a Monte Carlo simulation study. Finally, we propose a test statistic G to test validity of the exponentially distributions assumption. And Using Monte Carlo Method, we study the power of the proposed test procedure.
|