應力──強度模式之品質績效評估系統

Stress-Strength Model Quality Performance Assessment System

製造和醫療的商品績效比較，大多數商品的品質特性被假設來自常態分佈，不過，也有不少學者認為商品的品質特性可能來自非常態分配，如偉伯分配。此外，在商品績效的檢測實驗中，實驗者可能由於某些原因無法收集完整的觀測數據，諸如時間限制，人為因素和其他條件，因此，設限資料在實際情況中經常發生。本研究提出一個分析非常態分配、型二右設限且大樣本的兩商品優劣之新方法，並以程式語言C#去發展一個商業智慧中的品質績效評估系統。管理者利用此系統去確定兩商品的優劣。此系統能顯著減少生產成本，也能增強製程能力以提升企業的競爭優勢。

The comparison of product performance in manufacturing and medical. Most product quality characteristics are assumed from normal distribution. However, there are many scholars believe that the quality characteristics of the products may be come from non-normal distributions, such as Weibull distribution. In addition, comparing the performance of products in the quality testing experiment, the experimenter may be unable to collect complete observational data for some reason, such as time constraints, the human factors and other conditions. Therefore, the censored data often may arise in the actual situation. In this study, a new approach of analyzing non-normal, type II right censored sample and large sample quality data is proposed in the superiority-inferiority of two products. Moreover, program language C# is utilized to develop a quality performance assessment system in Business Intelligence. The management can use the system to determine the superiority-inferiority of two products. The system can reduce the product cost significantly, and also can enhance product process capability, to promote competitive advantage of enterprise.