Inference on Constant Stress Accelerated Life Tests Under Exponentiated Exponential Distribution

Accelerated life tests have become increasingly important because of highercustomer expectations for better reliability, more complicated products withmore components, rapidly changing technologies advances, and the clear needfor rapid product development. Hence, accelerated life tests have been widelyused in manufacturing industries, particularly to obtain timely informationon the reliability. Maximum likelihood estimation is the starting point whenit comes to estimating the parameters of the model. In this paper, besides themethod of maximum likelihood, nine other frequentist estimation methodsare proposed to obtain the estimates of the exponentiated exponential distribution parameters under constant stress accelerated life testing. We considertwo parametric bootstrap confedence intervals based on different methods ofestimation. Furthermore, we use the different estimates to predict the shapeparameter and the reliability function of the distribution under the usualconditions. The performance of the ten proposed estimation methods isevaluated via an extensive simulation study. As an empirical illustration,the proposed estimation methods are applied to an accelerated life test dataset.