單邊規格下多品質特性製程之群組選擇

Abstract

決策者在挑選供應商時，常以良率為挑選條件之一，除此之外，也會面臨需要從多個供應商中選出符合需求及有較佳製程能力的供應商，而製程良率指標為廣泛使用且能有效衡量高品質製程良率的工具之一，因此，本篇研究主要探討供應商群組選擇問題，並藉由單邊規格下多品質特性的製程能力指標CpuT來選擇出具有較佳製程能力的供應商。本篇論文提供一個挑選流程，使得可以從k (k > 2)個供應商中選出一群有較佳製程能力的供應商，其中，利用Bonferroni校正解決多重比較問題。本篇論文亦提供假設檢定所對應的臨界值及在要求之檢定力下所需的樣本數，作為供應商群組選擇的輔助資訊。此外，我們針對供應商群組選擇問題應用Bonferroni校正與其他多重比較方法作檢定力分析及比較。最後，我們應用三個例子來說明如何在多個供應商中選出一群符合需求及有較佳製程能力的供應商，驗證此研究的可行性。

The decision-maker usually faces the problem of selecting the best manufacturing material supplier – the one with the largest capability index – from several available suppliers in the initial stage of production setting. In this thesis, we consider the group selection problem for one-sided processes with multiple independent characteristics. More specifically, the aim of this thesis is to select all the possibly best suppliers into a group among k (k > 2) suppliers based on CpuT. An effective selection procedure is developed for practical applications. We apply the Bonferroni method to deal with multiple comparisons problem. For practitioners’ convenience, critical values for testing procedure and required sample sizes for a designated selection power are also investigated and tabulated. In addition, we compare the power between two multiple comparisons techniques, Bonferroni method and multiple comparisons with the best (MCB) method. The application examples for one-sided processes with multiple characteristics are presented to illustrate the practicality of this approach.