依據品質良率指標Yq應用複式抽樣方法於供應商選擇

Abstract

製程良率 (Process Yield) 仍是應用於製造工業中對於量測製程好壞的一個判斷標準，而更為先進的量測公式稱為品質良率指標Yq 又將顧客損失考慮在內。Ng and Tsui在1992年指出品質良率指標針對產品品質特性偏離目標值之變異程度對良率作一個處罰的動作，即把平均的產品損失考慮出來。本研究應用品質良率指標來比較兩家供應商的製程能力。我們使用假設檢定來比較兩個指標，並且改寫成差距 (H0:Yq2-Yq1<=0 ,H1:Yq2-Yq1>0 ) 以及比例 (H0:Yq2/Yq1<=1 ,H1:Yq2/Yq1>1 ) 的形式分別來作比較。由於抽樣分配較為複雜，因此本研究使用複式抽樣的模擬方法來建立差距以及比例值的信賴下界。我們也針對了錯誤機率以及檢定力去比較四種複式抽樣方法間的表現。為了實務上的便利，本研究依據所需的檢定力整理出建議的樣本數量表並且基於我們建議的方法去設計一個流程來衡量兩家供應商的品質。最後，我們以一個製造電漿面板產品的製程為例，套用我們的流程去幫助製造商選擇一個製程品質較好的供應商。

Process yield has been used in the manufacturing industry for measuring process performance. A more advanced measurement formula called the quality index Yq, has been proposed to calculate the quality yield for arbitrary processes by taking customer loss into consideration. Ng and Tsui (1992) indicated that Yq penalizes yield for the variation of the product characteristics form its target, which presents a measure of the average product loss. In this paper, we use the index Yq to compare which supplier has better process performance. The supplier selection decisions would be based on the hypothesis testing comparing the two Yq values, H0:Yq2-Yq1<=0 versus H1:Yq2-Yq1>0 (difference testing), or H0:Yq2/Yq1<=1 versus H0:Yq2/Yq1>1 (ratio testing). Due to the complexity of the sampling probability, we use a bootstrap resampling simulation to construct a lower confidence bound for the difference and the ratio statistics. To compare the performance of those four Bootstrap methods, further simulations of error probability and selection power analysis are conducted. For practical application, we tabulate the different sample size required for designated selecting power. A selection procedure is developed to select a better supplier in quality performance with the method. Finally, we also investigate a real-world case on the PDP (Plasma Display Panel) and apply the selecting procedure to help the manufacturer selecting a better supplier.