Isotonic Rules for Selecting Good Truncated Exponential Populations

摘要

The problem of selecting truncated exponential populations better than a control under an ordering prior is studied. Based on some prior information, it is reasonable to set lower bounds for the concerned parameters. Through this consideration, an isotonic selection rule is proposed. This selection rule always satisfies the requirement that the probability of a correct selection is at least equal to some prespecified value p:. A criterion is proposed to evaluate the performance of the selection rules. Simulation results indicate that this rule always performs better than some other earlier existing isotonic selection rules. (Author)