雙邊規格下多條生產線製程之群組選擇

Abstract

製程能力指標被用來解決極低不良率問題，且常被用來衡量製造產品符合規格的能力。對於製造業而言，經常面臨供應商選擇問題，因此本研究主要探討供應商的群組選擇問題並藉由雙邊規格下多條生產線的製程能力指標CpkM來評選出一群有較佳製程能力的供應商。本文提供一套有系統性的方法來解決多條生產線之供應商群組選擇問題，並搭配Bonferroni方法調整多重比較時之誤差膨脹問題。另外，本文也提供了進行供應商比較時所對應之臨界值及在要求之檢定力下之所需要的樣本數。最後，我們比較Bonferroni與 Multiple Comparison with the Best (MCB)兩種方法之檢定力大小，並應用不同例子說明兩種方法之決策結果不盡相同。當實際只有一個較好供應商時，可運用MCB方法，而若有兩個以上較好供應商時，則建議使用Bonferroni方法。

Process Capability indices (PCIs) have been proposed to deal with the problem for extremely low fraction of defectives and to provide numerical measures in process capability. For manufacturing industries, they often face the group supplier selection problem. Hence, in this thesis, we consider the group selection problem for two-sided processes with multiple manufacturing lines and select a group which contains suppliers having a higher capability value based on the process capability index CpkM. We present a systematic approach to tackle the group selection problem and to control the overall error inflation problem due to multiple tests by the Bonferroni adjustment. In addition, we provide the critical values for testing procedure and sample sizes for designated selection powers. Finally, we compare the powers of the Bonferroni method and the MCB method and illustrate the decisions of the two methods are not quite similar by giving numerical examples. We prefer to use MCB method when there is only one best supplier and use Bonferroni method when there are more than or equal to two best suppliers.