ON CONFIDENCE LIMITS FOR THE RELIABILITY OF SYSTEMS,
Abstract
An application of the asymptotic Chi-square distribution of the log-likelihood ratio is made to obtain approximate confidence intervals (or interval estimates) for the reliability of any structure of the class of structures which can be represented by a monotone Boolean function of Bernoulli variates. This is for the case when the data available consists of a number of Bernoulli trials for the performance or non-performance of the components. Some special methods are obtained for a subclass of structures which contain those which fail if, and only if, a certain prescribed number of their components fail. This generalizes the results of A. Madansky (AD-609 429) for series, parallel and series-parallel systems. Computational procedures are given and some numerical studies are reported which assess how accurate the confidence intervals are as a function of the sample size. Some comparisons are made with exact results in the cases where they are known. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1965
- Accession Number
- AD0627305
Entities
People
- Janet Myhre
- San C. Saunders
Organizations
- Boeing