Please use this identifier to cite or link to this item:

Title: 機器學習為基礎的現貨與期貨動態採購模型
A Machine Learning-Based Dynamic Purchasing Model of Spot and Futures
Authors: 薛名皓
Hsueh, Ming-Hao
Contributors: 莊皓鈞
Chuang, Hao-Chun
Hsueh, Ming-Hao
Keywords: 機器學習
Machine learning
Data analysis
Numerical simulation
Dynamic procurement
Raw material futures
Date: 2022
Issue Date: 2022-05-02 15:00:56 (UTC+8)
Abstract: 對於需要持續優化獲利的企業而言,能否在原物料價格隨機變動下進行成本最佳化的採購規劃,對其營運和財務績效的管理甚為重要。此決策問題在原物料有現貨和期貨可供選擇時又更為複雜。一般而言,企業可以採用以歷史價格先行預測價格走勢,並依據預測值訂定未來一段期間採購計畫的策略。有別於此種先預測後決策的傳統思維,本研究提出一個新的動態採購最佳化模型,從機器學習觀點運用現貨和期貨的價格數據,以預估最佳採購量而非價格預測為訓練目標,進而求得各個輸入特徵的最佳係數解。模擬分析結果顯示,此動態採購模型的成本表現顯著優於依據價格預測值所做的採購決策。除了模擬實驗,我們使用近兩年的布蘭特原油現貨與期貨價格進行實證分析,再次驗證本研究提出的模型優於依據價格預測值進行決策的模式。本文提出的理論模型有著線性規劃的高運算效率,並可用在多種須考量現貨和期貨價格的原物料採購情境,如金屬、穀物、原油、天然氣等,故同時具有實務價值。
For enterprises that need to continuously optimize profits, it is very important to optimize the procurement planning under the random fluctuations of raw material prices, especially in the management of their operational and financial performance. This decision problem is more complicated when raw materials can be purchased through the spot and futures markets. Generally speaking, enterprises can adopt a strategy of predicting price trends in advance based on historical prices and then formulating procurement plans for a period of time in the future based on the predicted values. Different from the traditional thinking of making predictions before making decisions, this study proposes a new dynamic procurement optimization model, which uses the price data of spot and futures from the perspective of machine learning to estimate the optimal procurement volume instead of price prediction. The simulation results show that the procurement decision of this dynamic procurement model is significantly better than the procurement decision based on price forecasts. In addition to the simulation experiments, we use the spot and futures prices of the Brent crude oil in the past two years to conduct empirical analysis and do verify that the model proposed in this study is superior to the decision-making model based on price forecasts. The theoretical model proposed in this paper has high computational efficiency of linear programming and can be used in a variety of raw material procurement scenarios where spot and futures prices must be considered, such as metals, grains, crude oil, natural gas, etc. Thus, it has a practical value at the same time.
Reference: Beutel, A. L., & Minner, S. (2012). Safety stock planning under causal demand forecasting. International Journal of Production Economics, 140(2), 637-645.
Geman, H. (2005). Energy commodity prices: Is mean-reversion dead?. The Journal of Alternative Investments, 8(2), 31-45.
Geman, H., & Nguyen, V. N. (2005). Soybean inventory and forward curve dynamics. Management Science, 51(7), 1076-1091.
Goel, A., & Gutierrez, G. J. (2011). Multiechelon procurement and distribution policies for traded commodities. Management Science, 57(12), 2228-2244.
Mandl, C., & Minner, S. (2020). Data-driven optimization for commodity procurement under price uncertainty. Manufacturing & Service Operations Management.
Schwartz, E. S. (1997). The stochastic behavior of commodity prices: Implications for valuation and hedging. The Journal of finance, 52(3), 923-973.
Secomandi, N., & Kekre, S. (2014). Optimal energy procurement in spot and forward markets. Manufacturing & Service Operations Management, 16(2), 270-282.
Shrestha, G. B., Pokharel, B. K., Lie, T. T., & Fleten, S. E. (2008). Management of price uncertainty in short-term generation planning. IET generation, transmission & distribution, 2(4), 491-504.
Thakkar, A., & Chaudhari, K. (2021). Fusion in stock market prediction: a decade survey on the necessity, recent developments, and potential future directions. Information Fusion, 65, 95-107.
Description: 碩士
Source URI:
Data Type: thesis
Appears in Collections:[資訊管理學系] 學位論文

Files in This Item:

File Description SizeFormat
602901.pdf2120KbAdobe PDF0View/Open

All items in 學術集成 are protected by copyright, with all rights reserved.

社群 sharing