Parameter estimation of Birnbaum-Saunders distribution with genetic algorithm under right censored data

The Birnbaum-Saunders (BS) distribution is a common reliability distribution used in scientific studies. There have been studies in the literature on parameter estimates for this distribution. In addition, in many studies, it is recommended to use genetic algorithm (GA) optimization methods for parameter estimation in modelling. This thesis focuses on the analysis and estimation of model parameters for a two-parameter Birnbaum-Saunders distribution for right-censored reliability data. For the estimation of Birnbaum-Saunders distribution parameters, we propose the genetic algorithm (GA) method as an alternative to the maximum likelihood estimation (ML) method. Psi31 data are often used as an example to show the limitations of prediction methods when using censored data. In addition, the performance of the ML and GA methods were studied by Monte Carlo simulation with different sample sizes and censorship rates.