Weibull Geometric Process Model for the Analysis of Accelerated Life Testing with Complete Data

In this paper the Weibull geometric process model is utilized for the analysis of accelerated life testing under constant stress. By assuming that the lifetimes under increasing stress levels form a geometric process, the maximum likelihood estimates of the parameters and their confidence intervals (CIs) using both asymptotic and parametric bootstrap method are derived. The performance of the estimators is evaluated by a simulation study with different pre-fixed parameters. This paper also compares the geometric process model with the traditional log-linear model. A simulation study is also performed to compare the performances of the geometric model and the log-linear model.