台灣進口能源依存度高，因此能源價格及其政策制定經常受國際情勢影響，又工業部門在能源佔比最高，因此台灣在制定能源政策時較著重在工業部門，也因為如此製造業對於如何改善能源使用是一個重要的課題；在化工廠中，使用能源的種類及設備數量眾多，且經常區分不同的轄區工場，因此在面對操作變動時，經常只考慮製程工場操作穩定而忽略消耗外購或自產公用物料(如蒸汽、電力…等)產生製程所需能源的公用系統(utilities system)，因此要在符合各工場需求找到最低成本的操作模式非常不易。因此本研究以案例廠為例，考量工廠操作現況以及其操作限制，依據情境分析定義最佳化問題以及設定敏感度分析項目，以混合整數線性規劃建構模型，並進行現況比較確認模型可靠性，最後透過敏感度分析觀察其能源操作最佳化結果的變化，目標找出能源操作成本最小化的能源操作最佳方案。
本研究實證結果，依據近兩年案例廠生產狀況選定(1) 操作建議Case Ⅰ—產能HL(高產能+低產能)、(2) 操作建議Case Ⅲ—產能HS(高產能+停產)進行現況比較，建議操作策略與現況操作的能源成本相比，成本分別下降2.3%及3.3%。透過調整製程操作達到節省能源，除了不須新增改善費用的支出，對工廠就能有及時的能源改善效益，並且亦可將其作為操作人員操作調整以及工廠正常操作能源使用花費的參考。

Taiwan has a problem of energy shortage, so we are very demand on imported energy. Therefore, the energy price and energy policy are always affected by the international situation. And industry sector uses more energy than any other sector. Thus, the development of energy policy usually focuses on industry. So, how to use energy efficiently is a challenge for industry sector.
There are many equipment in petrochemical plants driven by different types of energy. When any condition occurs in the factory that will affect the production, the operator's experience is usually used to adjust the system parameters to restore the stable production. But when the operator adjusts the system parameters, he only ensures that the process system is stable and ignores the operation status of the utility system. Therefore, it is difficult to find the lowest cost operation mode that meets the limiting conditions for operation. This study takes a factory as an example, analyzes the actual operation situation of the factory and then define the research scope and sensitivity analysis parameters. Develop and solves the linear programming model of energy operation then compare with current status to determine reliability of model. Sensitivity analysis of price, the requirement of steam, and equipment efficient are finally conducted. Base on this result, this study will propose an optimization of the steam and electricity operation. Compare the result with actual operating energy cost, it is decreased 2~3%. We don't need the extra investment costs, adjusting the mode of process operation to achieve energy saving. And it can provide the operating method and normal cost of operation for the operator.

一、中文文獻
1. 陳家榮，作業研究課程上課講義, Accessed by March, 2019.
2. Frederick S Hillier/Gerald J. Lieberma，譯者 馮景如，作業研究(I)，滄海書局，1996。
3. 陳建銘。「模擬最佳化應用於半導體後段封裝廠混合流線型排程問題之研究」。博士論文，國立清華大學工業工程與工程管理學系，2015。
4. 姜育辰。「運用系統模擬在連續式生產系統」。碩士論文，東海大學工業工程與經營資訊學系，2013。
5. 周百建。「資料科學與強化學習於石化原物料之價格預測與採購決策」。碩士論文，國立成功大學製造資訊與系統研究所，2018。
6. 謝秉璁。「小型風力發電場之風機配置最佳化研究」。碩士論文，國立臺灣大學機械工程學研究所，2016。
7. 陳偉欽。「利用數學規劃與萬用啟發式演算法最佳化角度限制產品之刀模設計」。碩士論文，國立成功大學工程管理碩士在職專班，2017。
8. 林牧潔。「利用效率排程演算法分析汽電共生廠最佳運轉」。碩士論文，大同大學電機工程學系(所)，2014。
9. 劉峻睿。「多週期汽電共生系統重整之最適化」。碩士論文，東海大學化學工程與材料工程學系，2017。
10. 桃春松。「一家國內化工廠的供應鏈網路最佳化設計」。碩士論文，南台科技大學企業管理系，2010。
11. 化學工程學會會刊，製程工廠能源整合的推動策略與效益，第65卷第一期，2018。
12. 107年能源統計年報, Accessed by June, 2020.https://www.moeaboe.gov.tw/ECW/populace/content/SubMenu.aspx?menu_id=6977
13. PO-LAB / Python-Gurobi-Pulp, Accessed by June, 2020.https://github.com/PO-LAB/Python-Gurobi-Pulp
14. Python Package Index / PuLP 2.1, Accessed by June, 2020.https://pypi.org/project/PuLP/
二、英文文獻
1. Xiaojian Huang, Pei Lu, Xianglong Luo*, Jianyong Chen, Zhi Yang, Yingzong Liang, Chao Wang, Ying Chen. Synthesis and simultaneous MINLP optimization of heat exchanger network, steam Rankine cycle, and organic Rankine cycle. Energy. Volume 195. 1-17. 2020.
2. Xiaojian Huang, Pei Lu, Xianglong Luo*, Jianyong Chen, Zhi Yang, Yingzong Liang, Chao Wang, Ying Chen. A generic algorithm of sustainability (GAS) function for industrial complex steam turbine and utility system optimization. Energy. Volume 164. 881-897. 2018.
3. Xiaojian Huang, Pei Lu, Xianglong Luo*, Jianyong Chen, Zhi Yang, Yingzong Liang, Chao Wang, Ying Chen. A generic algorithm of sustainability (GAS) function for industrial complex steam turbine and utility system optimization. Energy. Volume 164. 881-897. 2018.
4. Sandra R. Micheletto, Maria C.A. Carvalho, Jos´e M. Pinto∗. Operational optimization of the utility system of an oil refinery. Computers and Chemical Engineering. Volume32. 170–185. 2008.
5. Chia-Yen Lee, Andrew L. Johnson*. Proactive data envelopment analysis: Effective production and capacity expansion in stochastic environments. European Journal of Operational Research. Volume232. 537–548. 2014.
6. Chia-Yen Lee, Chen-Fu Chien. Stochastic programming for vendor portfolio selection and order allocation under delivery uncertainty. OR Spectrum. Volume36. 761–797. 2014.