不同於傳統的壽命試驗方式，一次性設備的壽命資料只有 左設限與右設限。而隨著現在的工業技術越來越發達，產品具 有高壽命與高可靠度的特性，所以我們通常會使用加速壽命試 驗，使元件早期失效以便於我們更快收集到元件的壽命資訊。 本研究係以一次性設備在加速壽命試驗下，做出元件壽命為廣 義指數分配的參數估計，並且討論以下兩個主題： 第一個主題是在壽命試驗中，收集越多關於元件的壽命資 訊，我們越能有效評估元件的可靠度。但如果要獲得更多的壽 命資訊，實驗成本就會增加，而實驗成本常會影響整個實驗的 規模。因此，我們給定成本預算的限制下，採用 D 型最適性的 最佳化準則，找出最佳的實驗配置。 可靠度是產品重要的品質特徵，可靠度越大代表其品質越 好。抽樣檢查即是常見用來判斷一批貨是否可被允收的方法， 如何決定實驗配置與允收臨界值將會影響整體的實驗結果。因 此，第二個主題是我們給定生產者風險與消費者風險下，建立 一個使得成本最小的可靠度抽樣計畫。 Unlike the conventional life testing experiments, the lifetime data col- lected in a one-shot device test are always left or right censoring. With the development of technology, the products with high reliability usually can work for a long time. In order to quickly obtain lifetime data, the accelerated life tests are usually conducted. In this study, we will investigate some problems related to accelerated life test with one-shot device testing. A generalized exponential lifetime distribution at each stress-level combination is considered. There are two main topics: To conduct a one-shot device accelerated life test more efficiently, one has to address the problem of determining optimal setting that produces the best estimation results. One practical problem arising from designing such a test is the budget of experiment. The size of budget always affects the decision of experimental setting and hence, affects the precision of estimation. Therefore, the first topic of this study is to determine the optimal experimental setting under D-optimality criterion with cost constraint. Reliability sampling is an important statistical tool in the area of quality control. It is used to make determination on accepting or rejecting a lot of product. The second topic of this study is to explore the optimal design of sampling plan and critical point to minimize the total cost incurred in conducting a reliability sampling plan under one-shot device accelerated life test with desired producer’s and consumer’s risks.
