Bayesian estimation of the difference between two mean times to failure

Abstract

This paper is concerned with the problem of allocating a fixed number of trials between the failure times of two independent components with unknown mean times to failure , in order to estimate with squared error loss. Introducing independent gamma priors on θ and ω, a fully sequential procedure is derived and shown to be nearly optimal. Monte Carlo results indicate that the fully sequential procedure performs much better than the best fixed allocation procedure

Keywords:

• Bayes Risk
• Efficiency
• Sequential
• Uniform Integrability