Confidence bounds for reliability of coherent systems based on binomially distributed component data. (English) Zbl 0606.62113
Reliability and quality control, Proc. Int. Conf., Columbia/Mo. 1984, 265-280 (1986).
[For the entire collection see Zbl 0588.00028.]
Empirical Bayes techniques for parameter estimation are applied to binomially distributed component test data. The transformed data are used in the log-likelihood ratio in order to obtain approximate confidence bounds for the system reliability, R. Properties of these bounds are studied for the case of series systems. Results from theoretical and numerical studies show that these methods compare favorably with other approximate methods.
Empirical Bayes techniques for parameter estimation are applied to binomially distributed component test data. The transformed data are used in the log-likelihood ratio in order to obtain approximate confidence bounds for the system reliability, R. Properties of these bounds are studied for the case of series systems. Results from theoretical and numerical studies show that these methods compare favorably with other approximate methods.
MSC:
62N05 | Reliability and life testing |
62F25 | Parametric tolerance and confidence regions |
62F15 | Bayesian inference |
62F10 | Point estimation |