Inference for Accelerated Life Testing Experiments with Generalized Inverted Exponential Model Under Hybrid Censoring

{Simulation is used to demonstrate a significant improvement in the quality of highly reliable and sophisticated products using accelerated life testing. We have discussed the estimation of unknown parameters using classical and Bayesian estimation methods for the generalized inverted exponential model, employing simple step-stress accelerated life testing with a hybrid censoring scheme.} This paper develops classical and Bayesian inferential techniques for unknown parameters of the generalized inverted exponential distribution by using a simple step-stress accelerated life test with type-I hybrid censoring scheme. The cumulative exposure model is used to connect the lifetime distributions before and after the stress change. Maximum likelihood estimates and their corresponding asymptotic confidence intervals are obtained using the observed likelihood and Fisher information matrix. For Bayesian inference, informative and non-informative independent gamma priors are considered under the squared error loss function. Since the conditional posterior distributions do not have standard forms, posterior samples are generated by employing the Metropolis–Hastings algorithm to obtain Bayes estimates and their associated highest posterior density credible intervals. For illustrative purposes, a simulation study is performed under different experimental settings to compare the proposed methods of estimation with the help of bias, mean squared error, average length of intervals, and their coverage percentages. The results show that estimation accuracy generally improves with increasing sample size and decreasing censoring proportion, while Bayesian estimation, particularly with informative priors, provides competitive performance. In the end, two real data sets are examined to show the suitability of the distribution and the practical application of the suggested methodology.