Selecting the most reliable design under type-II censoring

In product reliability R & D stage, the decision maker usually faces the problem of selecting the most reliable design among k available designs. To save the life-testing time, it is reasonable to test n items from each design until there are r (r ≤ n) failures from each. Under this kind of censoring (type-II censoring) plan and when the life distribution of the underlying design follows the Weibull distribution, this paper proposes a natural selection rule to achieve the goal. The sample size, the censoring size and the critical value, which are needed by this rule, are computed with the protection of probability of correct selection and the achievement of a specified time-saving factor by the experimenter.