Selecting the most reliable design under type-II censored accelerated testing

Accelerated life tests (ALTs) of products and materials are used to obtain reliability information within a reasonable time frame. However, there are no suitable selection rules for selecting the best product or material as a result of the ALT. Under the type-II censoring plan, a reasonable selection rule for an ALT using the Arrhenius model is proposed. The advantages of this selection rule are compared with a nonALT selection rule by using as criteria the average ratio of time-saving in life testing and sample size. The tradeoff between an increased sample size and a shortened life testing time for the ALT selection procedure is feasible. >