雙邊規格下多品質特性製程之群組選擇

Abstract

本研究主要探討群組選擇問題，並藉由雙邊規格下多品質特性的製程能力指標 C_pk^T 來選擇出擁有較佳製程能力的供應商。本文以Bonferroni法來解決問題，將任意兩家供應商的 C ̂_pk^T 相減作為檢定統計量，並以假設檢定比較任意兩家供應商的製程能力及差距。本文亦提供假設檢定所對應的臨界值、比較Bonferroni法與另一多重比較法MCB法之檢定力的差異及在要求之檢定力下所需的樣本數，並驗證此研究的可行性，作為供應商選擇的輔助資訊。最後，我們舉三個例子說明如何使用這兩個多重比較法做供應商選擇。

In this thesis, we consider the supplier selection problem which deals with comparing two-sided processes with multiple characteristics and selecting a group that has a significantly higher capability value based on the process capability index C_pk^T. We present the Bonferroni method to tackle the supplier selection problem. We take the difference C ̂_pk^T values of any two suppliers as the test statistic. For the Bonferroni method, we provide the testing hypothesis for comparing any of two suppliers’ performance and examining the process capability difference between two suppliers. We present the critical values for the testing procedure, comparison the power of the Bonferroni method and the MCB method, required sample sizes for designated selection powers and to illustrate the practicality of this approach. Three application examples on a two-sided process with multiple characteristics are presented to compare with the Bonferroni method and the MCB method.