多重分配檢測品質績效決策系統

製程能力分析被發展用以品質績效的評估。大多數製程能力指標都假設產品的品質特性是常態分配。然而很多產品的壽命模式是非常態分配，包括指數分配、柏拉圖分配、韋伯分配、冪次分配族等等。並且在產品壽命測試實驗中，設限資料是經常發生的。本研究提出一個分析非常態品質資料的新方法。在產品的壽命模式是冪次分配族而且在型二右設限樣本的假設之下，以發展一個新的假設檢定程序。而此新的假設檢定程序是商業智慧（BI）中的一個多重分配檢測品質績效決策系統，並且利用電腦程式語言C#去開發此多重分配檢測品質績效決策系統。此系統能同時評估指數分配、韋伯分配和柏拉圖分配的產品壽命績效。

Process capability analysis has been developed for assessing quality performance. Most capability indices assume that the quality characteristic has a normal distribution. Nevertheless, the lifetime model of many products generally may possess a non-normal distribution which including exponential, Pareto, Weibull distribution or Power distribution family and so forth. Moreover, censored samples may arise in life testing experiments. In this study, a new approach of analyzing non-normal quality data is proposed. This study developed a new hypothesis testing procedure under the Power distribution family with the right type II censored sample. The new hypothesis testing procedure is a Multiple Distribution Testing Quality Performance Decision System in Business Intelligence(BI). Moreover, program language C# is utilized to develop the Multiple Distribution Testing Quality Performance Decision System. The system can assess the lifetime performance of product with exponential distribution, Weibull distribution and Pareto distribution simultaneously.