Bayesian reliability of the Weibull failure model with bivariate characterization of the parameters (Q794378)

The author develops Bayes estimators of the Weibull reliability function when the scale and shape parameters in the Weibull failure model have a bivariate uniform or bivariate exponential distribution. These Bayes estimators do not have computationally convenient forms and must be numerically evaluated. Computational experience, using Monte Carlo simulation, in evaluating the MLE and Bayes estimates for the Weibull failure model is reported. A related conclusion of importance to the practitioner is that these Bayes estimates of the reliability function slightly overestimate the specific value of the reliability.