Statistical reliability analysis based on the two-sided power distribution

Within the scope of the statistical reliability analysis studies, the probability of a component or systems consisting of more than one component to be operational is modeled using survival random variables. The two-sided power (TSP) distribution is useful in modeling life-time data that have values on a limited [a,b] range and have a threshold. In this study, by considering the basic reliability indices of the TSP distribution, change over the time of a component which is modelled by the TSP distribution is examined, firstly. Furthermore, reliability and hazard rate estimates are calculated under different loss functions using maximum likelihood (ML) and Bayesian methods. Then, stress-strength reliability, which has importance in statistical reliability analysis studies, is examined under assumptions of simple and multi-component systems and estimated by ML and Bayesian estimation methods. The bootstrap confidence intervals for all ML estimates and the highest posterior density confidence intervals for Bayesian estimates by using Markov chain Monte Carlo (MCMC) algorithm are obtained. All obtained calculation methods are compared with simulations and exemplified with real data applications