整合能源消耗與製程輔具資源綁定之流線型生產排程研究

Flow-Shop Scheduling with Consideration of Energy Consumption and Auxiliary Resource Binding

劉奕和(Yi-He Liu)

指導教授 ： 黃奎隆

國立臺灣大學/工學院/工業工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202001763

摘要

隨著全球環保意識抬頭，綠色生產（Green Production）的議題逐漸受到重視，因此結合能源消耗與生產排程之研究勢必成為未來之趨勢。而實務上許多製程皆須搭配特定輔具資源協助生產，如鑄造產業製作外沙箱之輔具鐵斗、半導體產業微影製程之輔具光罩等，若未將輔具資源等因素考慮至排程作業中，常會導致生產排程績效不佳。目前大多數研究僅針對個別議題進行探討，並未同時考量能源消耗與輔具資源搭配之整合問題，若能同時有效整合此兩種議題，將可使生產排程規劃更有效率，並達到降低能耗成本之目的。
本研究探討不同工件需搭配不同輔具資源綁定生產，且各生產輔具皆有其相對應之能源消耗量，在滿足各生產時間點總能源消耗量之限制下，如何以最佳輔具資源配對，解決流線型工廠之生產排程問題如工件等待特定輔具資源釋放導致機台被迫閒置，或工件選擇能源消耗量高之輔具生產導致生產成本提高等議題。故本研究提出以基因演算法為基結合混整數線性規劃模型之兩階段啟發式演算法（GABTSO），並以最小化總完工時間與加權總能源消耗成本為目標求得品質優良的排程最適解。而在基因演算法的部分，本研究亦提出一演算法參數最佳化之機器學習模型，僅需告知期望排程之生產情況如工件數、輔具數等資訊，該模型即會自動輸出最佳之基因演算法參數設置以獲得最佳之求解效果。
本研究數值分析將四個探討因子包含(1)工件數、(2)輔具數、(3)能耗峰值限制與(4)能源成本等，分別設定不同之水準數形成共24種情境，探討不同因子水準對於GABTSO求解之影響，並藉數值分析結果提供相關之策略建議。最後並以國內大型鑄造廠實際資料為範例進行實務驗證，將最適生產排程規劃提供給公司管理階層，以作為決策時之考量依據，藉以提升公司之競爭力。

關鍵字

綠色生產排程 ； 輔具資源綁定 ； 基因演算法 ； 混整數線性規劃 ； 機器學習

並列摘要

Some manufacturing processes can only be performed when machines and the corresponding auxiliary resources are simultaneously available. Particularly, with the characteristics where the auxiliary resources and jobs are many-to-many correspondence. In addition, when job is assigned with different types of auxiliary, the energy consumption may vary accordingly. To promote the green production and to reduce cost, joining assignment of auxiliary resources with reduction of energy consumption is investigated.
We propose a two-stage heuristic algorithm which adopts the machine learning parameter optimization based genetic algorithm combined with a linear programming model to solve the flow-shop scheduling problem with requirement of auxiliary resources and energy consumption. Once an auxiliary tool is assigned to a job, that tool can’t be released until the job completes the operations those need the auxiliary tool. Jobs and auxiliary tools are many-to-many correspondence, and each auxiliary tool has different energy consumption. Every time point of the scheduling period has different energy consumption peak and energy cost. The objective of the problem is to determine a production scheduling that minimizes the weighted makespan and the total energy consumption cost under all of above constraints.
This study conducted numerical analysis to analyze how the four discussion factors include 1. Job, 2. Auxiliary, 3. Energy Peak, 4. Energy Cost to affect our heuristic algorithm and give some management insight based on numerical result. Lastly, a real scheduling problem will be used to verify the effect of this study.

並列關鍵字

Green Scheduling ； Auxiliary Resource Binding ； Genetic Algorithm ； Mixed Integer Linear Programming ； Machine Learning

參考文獻

1. Peter, G., Rebeka, L. (2007). Review of sustainability terms and their definitions. Journal of Cleaner Production, 15(18), 1875-1885.

2. International Energy Outlook 2019 (#IEO2019). Washington, DC: U.S. Energy Information Administration.

3. Baines, T., Steve, B., Ornella, B., Peter, B. (2012). Examining green production and its role within the competitive strategy of manufacturers. Journal of Industrial Engineering and Management, 5(1), 53-87.

4. Zheng, X. L., Wang, L. (2018). A Collaborative Multiobjective Fruit Fly Optimization Algorithm for the Resource Constrained Unrelated Parallel Machine Green Scheduling Problem. IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS: SYSTEMS, 48(5), 790-800.

5. Johnson, S. M., (1954). OPTIMAL TWO- AND THREE-STAGE PRODUCTION SCHEDULES WITH SETUP TIMES INCLUDED. Naval Research Logistics Quarterly, 1(1), 61-68.