機台產能配置輔助系統與實證

NHUIR > College of Science and Technology > Department of Information Management > Disserations and Theses > Item 987654321/20537

Please use this identifier to cite or link to this item: http://nhuir.nhu.edu.tw/handle/987654321/20537

Title:	機台產能配置輔助系統與實證
Other Titles:	An empirical study on the tool capacity allocation system
Authors:	練鴻展 Lian, Hong-jhan
Contributors:	資訊管理學系碩士班李翔詣 Hsiang-yi Lee
Keywords:	整數規劃;最佳化;產能配置 capacity allocation;optimization;integer programming
Date:	2008
Issue Date:	2015-05-10 15:50:19 (UTC+8)
Abstract:	本研究目的是為晶圓代工廠建立一個產能配置決策支援系統。在現今競爭激烈的商業環境裡，高度的生產效率是保持競爭優勢的關鍵因素。而充份利用工廠產能的產能規劃，才能得到最佳化的產出，提高生產效率。然而，由於不同機台產能上限與不同產品的產能需求不一，要得到產能最佳化配置並不容易。必須建立出正確的產能規劃模型並配合數學規劃軟體的計算。在本研究的個案中，最主要問題是工廠機台在跨製程使用時，產能會因為製程變動而下降。當機台使用於單一製程時，產能使用上限為1.0，當機台跨製程使用時，使用上限則為0.9。本研究利用整數規劃具有可依據變數值選擇限制式的特性，將機台使用狀況以二元變數來表示。當二元變數為1時表示機台有使用，當二元變數為0時表示機台沒有被使用。再使用另一個二元變數來表示機台是否跨製程使用，以決定限制性的上限為何。本系統建構是以數學規劃軟體Lingo為核心，以Excel檔案為資料來源。結果顯示，個案工廠目前採用的手動方式配置產能需時超過10分鐘，且僅能確認產能配置是否可行，無法知道最佳產量為何。而本系統可以在1-2分鐘內得到結果，且機台產能配置可以確保產量為最佳。顯示本系統的確可以對於IE人員的產能規劃有具體幫助。　　The purpose of this study is to establish a capacity allocation decision support system for a wafer foundry factory. In today’s fierce competing world, one of the key factors for maintaining competitive advantage is to reach production efficiency. And this is based on the fully utilization of tool capacity. However, this is not an easy job since tool capacity varies as tool type varies and different products usually consume different tool capacity. A correct math programming model must be set up to gain the optimal allocation and computer software may be required. The challenge this study faced is that tool capacity changes as it is used in a single process or multiple process. That is, the capacity limitation is not fixed, it is 1.0 when a tool is assigned for one single process only and it is 0.9 when the tool is assigned for multiple processes. The study then adapts an integer programming model with two types of binary integer variables. The first one is to indicate that the tool is used in certain process or not. It becomes 1 when the tool is allocated or 0 when the tool is not allocated. Another binary integer is then used to indicate whether the tool is used in single process or cross-process used. The result is then used to determine the right capacity constrain. The system is developed via an Excel file with a Lingo kernel. The results indicate that the system could get the result within 1-2 min, while the manual way used in that factory took more than 10 min. The system also makes sure the output is optimized with the capacity allocation. This indicates that this system could be helpful for Industrial Engineer when making capacity planning.
Appears in Collections:	[Department of Information Management] Disserations and Theses

Files in This Item:

File	Description	Size	Format
096NHU05396017-001.pdf		1123Kb	Adobe PDF	425	View/Open
index.html		0Kb	HTML	249	View/Open

Loading...