A Bayes approach to simultaneous evaluation of similar assemblies

Simultaneous qualification testing of similar electronic-assembly designs that have the same reliability objectives lends itself to a Bayes approach. The author presents a method for obtaining the number of units to be put on test (test units) and the number of survivors required for this problem. The Bayes approach can reduce the total number of test units. The similarity of any two assemblies can be qualified using projected failure rates of components common to each design and components that differ between them. Under binomial sampling, survival probabilities follow the conjugate beta distribution. Two systems of linear equations are developed. Their solutions are vectors that describe, for each design: (1) the number of test units, and (2) the minimum number of survivors to demonstrate the reliability objective.<
>