A nonparametric approach for selecting the most reliable population

Abstract

In this paper, we present a nonparametric method for selecting the most reliable population in reliability life testing. We propose a new statistic, the transitional order statistic, and give a natural selection rule based on this statistic. Under the proposed selection rule, the probability of correct selection is calculated by defining a nonlinear indifference zone under the least favorable configuration. Finally, we present an example to illustrate the selection procedure and validate the results.