個案公司爲世界前五大封測廠，對產品出貨品質及檢驗資料報告之要求非常嚴謹。為確保產品的品質，公司要求在全自動生產的流程下蒐集各項製程資料，包括產品檢驗報告、統計製程值、生產參數以及生產機台異常狀況報告等，要提供給客戶以滿足需求。 由於公司已導入電腦整合製造 (Computer Integrated Manufacturing, CIM) 系統及高度自動化生產機台，並透過電腦控制、網路與管理技術，已可將產品由工廠投入到成品產出的所有生產過程與控制整合在一起，進而掌控生產品質。而生產機台及系統的穩定度和工廠產出品質及產能更有直接關係，針對如何降低生產機台的異常頻率更是努力的目標。
對於直接產線上的維護者和使用者而言，對生產機台的異常狀態的解決，要在最快時間內判斷異常原因，找到問題，並解決問題，並在最短的時間內儘快讓生產機台回復正常運作，這是不容易的事情。利用案例式推理的手法來解決，但發現此方法是依過去的經驗規則做一鏈結，被用來解決許多相似的問題。反之問題資料不完整且隨時間動態改變的情況下，很容易造成問題的失真而無法解決。針對新的問題發生也會有處理時效不佳的狀況發生，進而降低生產效能。
本研究主要目的是設計一套機制，先蒐集生產機台主要異常訊息並加以分析，透過高度自動化系統的管理機制下，在異常發生時可以停止機台的運作生產，避免異常產品的持續產出。並且依據異常訊息資料的分析彙整，可以提供維護者和使用者正確快速的資訊，並很快找到真正異常原因並解決。除了縮短生產機台異常回復之時間，亦可以降低生產機台異常頻率，進而提升產品品質及生產效率，並得到客戶之信任。同時為了長期管理的考量，建立圖表介面管理監測系統，針對生產機台異常狀況和處理的判斷邏輯，進行歸納和追蹤，根據這些可靠的資料訊息，提供一項決策模式供使用者和管理者使用。

The case company is one of the top five packaging and testing factories in the world. The industry has strict manufacturing requirements. To ensure the manufacturing quality, the company must have strict internal policies for controlling the manufacturing processes to meet customer demands. The case company has installed both highly automated equipments and computer integrated manufacturing (CIM) systems for high quality production. Because production quality is highly correlated with equipment stability, it is critical for the company to achieve high levels of equipment stability and utilizations for avoiding product deficits.
However, it is difficult to identify the root cause of problems within limited time and to act properly to resolve the problems. In most cases, the problems may be resolved by implementing expert system (ES) or case-base reasoning (CBR) systems. However, there still are problems lacking proper resolving procedures because the low frequencies problems and insufficient experiences. For resolving these types of problems, new methods are needed.
The main purpose of this study is to develop reasonable mechanisms and systems to resolve these problems in the production line. By timely analyzing relevant raw data to predict the occurrence of problems in advance, the equipment can be stopped to resolve the problems in time before incurring serious manufacturing interruptions and product deficits. Beside, reducing the frequency of abnormal events will result in good product quality and production efficiency, obtaining the trust of customers. The company also has created graphical user interface (GUI) to monitor the actual performance of all equipments with more convenience in real time. All pre-defined event can be traced and monitored by the system. With the system, it has become easy to identify production and equipment abnormal situations and then have correct instructions and actions taken. This also reduces the troubleshooting time. According to implementation results, the methods developed in this case should provide a new model for decision-making and actions for the manufacturing control of the industry.

【英文文獻】
1.Grove, (1852), “The Phenomenon first Described with Reported Vaporization and Film Formation of Metal Films by Sputtering.”
2.Konopka, J., (1996), “Improving Output in Semiconductor Manufacturing Environment,” Ph.D. Dissertation. Arizona Graduate College.
3.Koch, R., (1998), “The 80/20 Principle: The Secret of Achieving More with Less, Doubleday.”
4.Leachman, R., (1997), “Closed-Loop Measurement of Equipment Efficiency and Equipment Capacity.” IEEE Transaction on Semiconductor Manufacturing.
5.McGraw, Karen L.; Briggs, Karan Harbison, (1989), “Knowledge Acquisition: Principles and Guidelines.”
6.Pareto, V., (1870), “The Fundamental Principles of Equilibrium in Solid Bodies.”. Polytechnic University of Turin.
7.Riesbeck, C.K., (1989), “Case-based Reasoning,” New Jersey: Lawrence Erlbaum Associates.
8.Schmidt, G.,C, (1998), “Case-Based Reasoning for Production Scheduling,” International Journal of Production Economics, Vol.6, pp.14-21.
9.SEMI E13, (1999), Standard for SEMI Equipment Communication Standard Message Service, Global Information & Control Committee.
10.SEMI E30-0600, (1999), Generic Model for Communications and Control of Manufacturing Equipment, Global Information & Control Committee.
11.SEMATECH, (1995), Semiconductor Manufacturing Productivity Overall Equipment Effectiveness Guidebook, Revision 1.0.
12.Watson, I., (1999), “Case-Based Reasoning is a Methodology not a Technology,” Knowledge-Based Systems,” Vol. 12, pp.303-308.
13.Walter A. Shewhart , (1939), “Statistical Method from the Viewpoint of Quality Control.”
 
【中文文獻】
1.QC手法研究小組，(2007)，「品質管制」，台北華泰。
2.李佳蓉，(1995)，「統計製程管制在積體電路製造業之應用研究」，國立交通大學科技管理碩士論文。
3.李婉如，(2001)，「晶圓代工廠之機台當機影響即時估算法」，國立交通大學工業工程與管理研究所碩士論文。
4.林劭函，(2005)，「晶圓廠機台妥善率組合決策之研究」，國立交通大學工業工程與管理研究所碩士論文。
5.黃子圳、王冠智，(2010)，「產業分析報告 TFT-LCD 面板產業分析以奇美為例」，國立虎尾科技大學碩士論文。
6.蔡彰文，(1999)，「製程能力分析與改善」，機械工業雜誌，第194期。
7.張志平、紀勝財譯，(2004)，「品質管制」，東華書局。
8.劉駿凱，(2011)，「應用案例式推理於半導體設備故障診斷輔助系統」，國立高雄應用科用大學電子工程研究所碩士論文。

【網路文獻】
1.CTS，(2011)，http://news.cts.com.tw/udn/money/201108/201108070795252.html. (Downloaded: November 19, 2013)。
2.Miracom Inc, 1998, Manufacturing Operations Management Systems Solution Provider. http://miracomglobal.com. (Downloaded: April 1, 2014)。
3.Wikipedia，http://zh.wikipedia.org/wiki/%E4%BF%A1%E5%BA%A6 (Downloaded: April 1, 2014)。
4.工研院IEK，(2013) http://www.moneydj.com/kmdj/news/NewsViewer.aspx?a=52776b60-fe65-4a08-8f95-2e521467821a#ixzz2jqprZtqF. (Downloaded: February 16, 2014)。
5.MoneyDJ 財經知識庫， http://www.moneydj.com/kmdj/wiki/wikiviewer.aspx?KeyID=d87b59da-fd31-48ea-807a-8885115cf669#ixzz2hHlMsePA. (Downloaded: December 16, 2014)。
6.經理人 Manager Today, http://www.managertoday.com.tw/?p=649. (Downloaded: March 16, 2014)。