台灣半導體產業因同業競爭越來越激烈，製程及設備人員需做出有效的分析與規劃，徹底掌握生產設備及產品之生產效能需求降低時間成本，進而提升產品產能與公司的競爭力。故本研究針對製程及設備產能之需求加以全盤分析考量及改善，依據實際生產資料比較及製程改善進行設備修正幫助管理階層以綜觀全局的角度降低設備生產週期時間(cycle time)、維修時間、消除產能瓶頸、以提升設備產能利用率降低設備採購成本支出持續進行改善提高公司獲利及企業競爭優勢，並利用時間序列分析的方式，建立統計預測模型，以精進半導體製程機台生產效率。本研究透過2017年10月8日至2018年2月8日，共四個月16台封裝機檯實際產線資料為研究樣本。將Uph(每單位小時產能)、Totally time(完全生產時間)、Postplace(後位)、Wrong work state(工作狀態錯誤)、Down Time(設備停機時間)及MTTA(Max) (平均協助時間)等變數的變化圖像化，解讀各變數於PDCA管理模式下的變化情形，接著建構ARIMA 模型，進行時間序列分析。
研究發現PDCA管理模式控管相當有效，當Postplace時間周期變化時，Wrong work state則呈現平穩的狀態; 在本研究期間中UPH微幅上升，Totally time明顯下降趨勢;Down Time & MTTA(Max)呈現正相關係。根據本研究的結果，得到下列結論，第一，如能控管好製程過程中Postplace時間周期變化，將能夠減少Wrong work state的發生。第二，不斷的執行PDCA模式，可將錯誤持續反饋更正，使得錯誤訊息變少，且機台效能穩定成週期性，利於操作人員控管機台。第三，UPH在整體機台的控制下相對穩定，公司需更加著重於各機台Totally time的控管。公司需有效地降低MTTA(Max)時間，以降低Down Time。最後，公司可聘用專門負責處理統計數據的人才，以協助各機台的數據處理，並將處理好的資料提供給相關技術人員及管理階層。
綜上，後續可針對半導體產業統計分析人才作為主題，加以研究半討體業與資料分析兩者間關係；同時可加入更多管理學要素，探討如何控管以分配半導體管理階層、技術人員及分析人員，以達到半導體業產能效益的極大化。

Due to the increasingly fierce competition in the Taiwan semiconductor industry, process and equipment personnel need to make effective analysis and planning, thoroughly master the production equipment and product production efficiency requirements to reduce time costs, and thereby enhance product production capacity and the company's competitiveness. Therefore, this research aims at the overall analysis, consideration and improvement of the requirements of the production process and equipment capacity. Based on the above-mentioned data comparison and process improvement, equipment modification is carried out to help the management to reduce the equipment production cycle time (cycle time), maintenance time, and eliminate production capacity from an overall perspective. Bottlenecks, improve equipment capacity utilization, reduce equipment procurement costs and continue to improve the company’s profitability and corporate competitive advantage.From October 8, 2017 to February 8, 2018, a total of four months of packaging machine operation data, Uph (per unit hour capacity), Totally time (full production time), Postplace (postplace), Wrong The work state (work state error), Down Time (equipment down time) and MTTA (Max) (mean assist time) and other variables are visualized, and the changes of each variable under the PDCA management mode are interpreted, and then the ARIMA model is constructed. According to the results of this research, the following conclusions are drawn. First, if the Postplace time period change during the manufacturing process can be controlled, the occurrence of Wrong work state will be reduced. Second, the continuous implementation of PDCA mode can continuously feedback and correct errors, so that there are fewer error messages, and the machine performance is stabilized periodically, which is beneficial for the operator to control the machine. Third, UPH is relatively stable under the control of the overall machine, and the company needs to pay more attention to the total time control of each machine. The company needs to effectively reduce the MTTA (Max) time to reduce the Down Time. Finally, the company can hire talents who specialize in processing statistical data to assist in the data processing of each machine, and provide the processed data to relevant technical personnel and management.
Based on the results of this research, the following conclusions are drawn. First of all, if the Postplace time period change in the manufacturing process can be controlled, it will reduce the occurrence of erroneous working conditions. Secondly, the continuous execution of PDCA mode can continuously feedback and correct errors, thereby reducing error messages and stabilizing machine performance on a regular basis, which is very beneficial for the operator to control the machine. Third, UPH is relatively stable under the control of the entire machine, so the company needs to pay more attention to the total time control of each machine. Companies need to effectively reduce the MTTA (maximum) time to reduce downtime. Finally, the company can hire talents who specialize in processing statistical data to assist in the data processing of each machine, and provide the processed data to relevant technical personnel and management personnel.
To sum up, we can focus on the statistical analysis talents of the semiconductor industry as the theme, and study the relationship between the semi-discussive industry and data analysis; at the same time, we can add more management elements to discuss how to control and allocate semiconductor management, technical personnel and Analysts in order to maximize the efficiency of the semiconductor industry.

工商時報(2019)，半導體不畏情勢動盪 台積電引領產值續創新高。2019.06.30。https://ctee.com.tw/topic/2019tw0050/semi
工商時報(2020)，6/23 MES雲端化 線上研討會。2019.06.20。https://ctee.com.tw/industrynews/activity/289141.html
工商時報(2020)，台積、聯發科母雞帶小雞 13檔悍股旺到明年。2020.11.21。https://ctee.com.tw/news/stock/373618.html
中時新聞網(2020)，資訊服務實威助家寬實業導入DELMIAWORK 打造工業4.0工廠。2020.10.28。https://www.chinatimes.com/realtimenews/20201028002108-260410?chdtv
今周刊(2017)，如何善用PDCA 避免犯同樣的錯？。2017.04.20。https://www.businesstoday.com.tw/article/category/80408/post/201704200046/%E5%A6%82%E4%BD%95%E5%96%84%E7%94%A8PDCA%20%E9%81%BF%E5%85%8D%E7%8A%AF%E5%90%8C%E6%A8%A3%E7%9A%84%E9%8C%AF%EF%BC%9F
今周刊(2017)，連企業老闆都搶著學！讓你收入翻倍的PDCA管理術。2017.03.02。https://www.businesstoday.com.tw/article-content-80392-162236
日本設備維護協會(2008)，「TPM 實踐法:全員參與的設備保養實施大全」，先鋒企業管理發展中心TPM 硏究小組譯，和昌，桃園。
江建勳(2009)，《積體電路設計廠商的策略分析》。國立臺灣大學企業管理所論文。
何年琛(2003)。半導體封裝業MES導入成功因素之探討。中央大學管理學院高階主管企管研究所論文。
沈家志、林冠宇、張家豪、劉志宏（2019）。《電漿蝕刻技術於晶圓切割產業之應用》。先進製造技術專輯。2019.9。
林書毅(2012)，「設備綜合效率改善之個案研究-以L公司為例」，國立高雄應用科技大學高階經營管理研究所碩士在職專班碩士論文。
邱俊斌（2006），「半導體設備綜合效力之探討─以半導體測試廠為例」，國立成功大學工業與資訊管理學系碩士班碩士論文。
科技報橘(2019)，為什麼積體電路產業中使用的「矽」，純度要達到 99．999999999％？。2019.10.01。https://buzzorange.com/techorange/2019/10/01/silicon-chip-industry/
科技新報(2020)，台積電利多頻傳，成熟製程、先進封裝都豐收。2020.11.19。https://technews.tw/2020/11/19/taiwan-semiconductor-manufacturing-co-ltd-has-a-lot-of-good-news-mature-process-and-advanced-packaging-have-a-good-harvest/
唐唯譯(2017)，「半導體封裝製程生命週期評估研究-以台灣某公司之球閘陣列產品為例」，國立臺北科技大學環境工程與管理研究所碩士論文。
高福成(2005)，《台灣企業實施TPM的適合性研究》。中原大學企業管理研究所論文。
高福成(2009)，《TPM全面生產維護推進實務》。機械工業出版社。
高福成(2005)，「台灣企業實施TPM 的適合性研究」，中原大學企業管理研究所碩士論文。
張淇樺(2020)，「應用 OEE 與限制理論提升封裝製程 效率-以 S 公司為例」，逢甲大學智能製造與工程管理碩士在職學位學程。
許竹佑(2019)，「應用六個標準差品管策略改進半導體封裝製程」，國立中央大學工業管理研究所在職專班。
陳宣汝(2015)，「運用PDCA循環提升財會作業效能之研究：以A公司清潔用品製造產業為例」，國立臺北大學國際財務金融碩士在職專班碩士論文。
陳紀航(2017)，「 應用智慧型代理人於製程參數優化及實證研究-以精密製造和IC封裝製程為例」，國立清華大學工業工程與工程管理學系所碩士論文。
陳虹銘(2018)，「應用PDCA管理循環於捲裝載帶製程生產良率提昇之研究—以L 公司為例」，國立高雄大學碩士論文。
曾愛文（2003），「應用結構化行動理論探討ERP MES導入效益」，國立中央大學資訊管理研究所碩士論文。
黃文琪(2013)，「半導體封裝廠設備綜合效率指標應用」，逢甲大學工業工程與系統管理所碩士論文。
黃茹鈺(2009)，「MES系統運用於收集生產工時之改善」，逢甲大學經營管理碩士在職專班碩士論文。
微光雜誌(2020)，PDCAF 循環式品質管理，數位行銷該要學會的成長駭客。2020.05.03。https://medium.com/vito-hsu/%E6%95%B8%E4%BD%8D%E8%A1%8C%E9%8A%B7%E8%A9%B2%E8%A6%81%E5%AD%B8%E6%9C%83%E7%9A%84-pdca-%E5%BE%AA%E7%92%B0%E8%A1%8C%E9%8A%B7%E7%AE%A1%E7%90%86-%E6%88%B4%E6%98%8E%E7%92%B0-144479431ff7
經理人(2018)，每天改善1%，1年強大37倍！一張圖，讀懂不斷改善的PDCA循環法則。2018.02.22。https://www.managertoday.com.tw/articles/view/55730
經濟日報(2020)，無懼疫情 資策會估台灣今年IC封測產值微幅成長。2020.06.04。https://money.udn.com/money/story/5612/4612012
經濟日報(2020)，經濟部次長林全能參訪東台 關注智慧製造產線發展。2020.11.20。https://money.udn.com/money/story/5612/5031262
趙豊昌(2009)，「整合良率預估及缺陷樣式辨識之晶圓缺陷診斷系統」，國立交通大學工業工程與管理所碩士論文。
數位時代(2015)，預知維修保養 開創工業4.0智慧服務先河。2015.08.25。https://www.digitimes.com.tw/iot/article.asp?cat=158&id=0000439578_yxn4juhj8eybw176jx5ix
數位時代(2020)，一顆新5G晶片的誕生會經過哪些步驟？白話文解析IC產業鏈全貌。2020.03.02。https://www.bnext.com.tw/article/56653/5g-ic
聯合新聞網(2020)，大量科技跨足半導體 攻檢測設備。2020.10.24。https://udn.com/news/story/7241/4959149
聯合新聞網(2020)，先進封裝技術大盤點。2020.08.14。https://udn.com/news/story/6871/4779184
蘇淵達(2012)，「架構導向車用電子製造品質管理模式之研究」，國立中山大學資訊管理所碩士論文。

Beenstock, M., E. Goldin ＆ Nabot, D. (1999), the Demand for Electricity in Israel, Energy Economics, 21, 168-183.
Box, G. E. P., & Jenkins, G. M. (1976). Time series analysis: Forecasting and control. San Francisco: Holden Day.
Darbellay, G.A. & Slama, M. (2000),Forecasting the Short-term Demand for Electricity：Do Neural Networks Stand a Better Chance？International Journal of Forecasting, 16, 71-83.
Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive times series with a unit root. Journal of the American Statistical Association, 74, 427-431.
Enders, W. (2004), Applied Econometric Time Series. Hoboken, NJ：Wiley.
Fatai, K., Oxley, L. & Scrimgeoar, F.G. (2003), Modeling and Forecasting the Demand for Electricity in New Zealand：A Comparison of Alternative Approach, Energy Journal, 24, 75-102.
Ghosh, S. a & Das, A. (2002), Short-run Electricity Demand Forecasts in Maharashtra, Applied Economics, 34, 1055-1059.
Gujarati, D.N. (2003), Basic Econometrics. New York：McGraw-Hill.
Hartmann, G. (1992) Antarktische benthische Ostracoden 8. Auswertung der Reise der "Meteor" (Ant. 11/4) in die Gewässer um Elephant Island und der Antarktischen Halbinsel. Helgoländer Meeresuntersuchungen, 46, 405-424.
Hippert, H.S., Bunn, D.W. & Sovza , R.C. (2005), Large Neural Networks for Electricity Load Forecasting：Are They Overfitted？ International Journal of Forecasting, 21, 425-434.
Hondroyiannis, G. (2004), Estimating Residential Demand for Electricity in Greece, Energy Economics, 26, 319-334.
Houthakker, H.S. (1951), Some Calculations of Electricity Consumption in Great Britain, Journal of the Royal Statistical Society(A), No.114,Part Ⅲ,351-371.
Konopka, J.M., (1995), Capacity Utilization Bottleneck Efficiency System-CUBES , IEEE Transactions on Components, Packaging and Manufacturing Technology-Part A, Vol.18, No.3, pp.484-491.
Larsen, B.M. & Nesbakken, R. (2004), Household Electricity End-use Consumption：Results from Econometric and Engineering Models, Energy Economics, 26, 179-200.
Leachman (1995) Benchmarking manufacturing performance in semiconductor industry. Production and Operations Management Vol.38 No.3 (Summer)：pp201-216.
Ljung, C. M., & Box, G. E. P. (1978). On a measure of lack of fit in time series models. Biometrika, 65, 297-303.
McMenamin, J.S. & Monforte, F.A. (1998), Short Term Energy Forecasting with Neural Networks, Energy Journal, 19, 43-61.
Nakajima, S. (1989), TPM Development Program, Productivity Press, Cambridge, MA.
Semiconductor Manufacturing Productivity Overall Equipment Effectiveness Guidebook， Revision 1.0，SEMATECH，April 13，1995。
Soares, L.J. & Souza, L.R. (2006), Forecasting Electricity Demand Using Generalized Long Memory, International Journal of Forecasting, 22, 17-28.
Tserkezos, E.D. (1992), Forecasting Residential Electricity Consumption in Greece Using Monthly and Quarterly Data, Energy Economics, 14, 226-232.
Zachariadis, T. & Pashourtidou, N. (2007), An Empirical Analysis of Electricity Consumption in Cyprus, Energy Economics, 29, 183-198.