A stochastic optimization problem is considered: n items are put on life test with random observation limits. The test is stopped for the item if a failure occurs or the observation limit is reached. The probability of observing a failure has to be maximized by choosing a suitable distribution of observation limits while the total time of test is given. It will be shown that for life distributions with increasing failure rate the observation limits should be random variables. For life distributions with decreasing failure rate equal observation limits for all items will be optimal.
展开▼
机译：考虑了一个随机优化问题：对 n 个项目进行具有随机观测极限的寿命测试。如果发生故障或达到观察限制，则停止物料的测试。在给定测试总时间的同时，必须通过选择合适的观察限分布来最大化观察到故障的概率。结果表明，对于故障率增加的寿命分布，观测极限应为随机变量。对于故障率递减的寿命分布，所有项目的观测限相等将是最佳的。
展开▼
