Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/92325
標題: 應用動態線性模型預測物流中心即時貨車需求量之研究
Applying Dynamic Linear Model to Forecast Real-Time Truck Demand for Logistics Centers
作者: Huang-Cheng Shiu
徐煌盛
關鍵字: Logistics Management;Demand Prediction;Dynamic Linear Model;物流管理;需求預測;動態線性模型
引用: Afzal, M. K. & Kim, Y. J., 2014. Demand Determinants for Urban Freight Consolidation Center -A Case of Korea, International Journal of Transportation, 2(2), pp. 89-108. Angulo, A., Nachtmann, H. & Waller, M. A., 2004. Supply chain information sharing in a vendor managed inventory partnership, Journal of Business Logistics, 25(1), pp. 101-120. Azoury, K. S., Miyaoka, J., 2014. Sequential learning versus no learning in Bayesian regression models, Naval Research Logistics, 61(7), pp. 532-548. Bourland, K. E., Powell, S. G. & Pyke, D. F., 1995. Exploiting timely demand information to reduce inventories, European Journal Operation Research, 99(2), pp. 239-253. Chen, H., Chiang, R.H.L. & Storey, V.C., 2012. Business Intelligence and Analytics: From Big Data to Big Impact, MIS Quarterly, 36(4), pp. 1165-1188. Garrido, R. A. & Mahmassani, H. S., 2000. Forecasting freight transportation demand with the space-time multinomial probit model, Transportation Research Part B: Methodological, 34(5), pp. 403-418. Heizer, J & Render, B., 2014. Principles of Operation Management. New York: Pearson. Jeseke, M., Gruner, M. & Weib, F., 2013. DHL, Big Data in Logistics. Germany: Solution & Innovation. Lewis, C. D., 1982. Industrial and Business Forecasting Methods. London: Butterworth Scientific. Oracle Corporation, 2006. Supply Chain Best Practice: Demand Planning Using Point-of-Sale Data. Unite State: Oracle. Pole, A., West, M. & Harrison, J., 1994. Applied Bayesian Forecasting and Time Series Analysis. New York: Chapman & Hall. Provost, F. & Fawcett, T., 2013. Data Science for Business. United States: O'Reilly Media. Sakauchi, T., 2011. Applying Bayesian Forecasting to Predict New Customers' Heating Oil Demand. Wisconsin: Marquette University. Sharma, S., 1996. Applied Multivariate Techniques. New York: John Wiley & Sons. Stevenson, W. J., 2013. Operations Management. New York: McGraw-Hill Waller, M. A., Fawcett, S. E., 2013. Data Science, Predictive Analytics, and Big Data: A Revolution That Will Transform Supply Chain Design and Management, Journal of Business Logistics, 34(2), pp. 77-84. Wang, H. L. & Shi, Z. K., 2012. Hierarchical regression model for analyzing truck freight demand, Applied Mechanics and Materials, vol. 178-181, pp. 2752-2756. West, M., Harrison, J. & Migon, H., 1985. Dynamic Generalized Linear Models and Bayesian Forecasting, Journal of American Statistical Association, 80(389), pp. 73-83 Williams, B. D., Waller, M. A., Ahire, S. & Ferrier, G. D., 2013. Predicting retailer orders with POS and order data: The inventory balance effect, European Journal of Operational Research, 232(3), pp. 593-600. Yelland, P. M. & Lee, E., 2003. Forecasting Product Sales with Dynamic Linear Mixture Models, Sun Microsystems Laboratories. Zhu, J., 2013. POS Data and Your Demand Forecast, Information Technology and Quantitative Management, 17, pp. 8-13 Harvilicz, D. (2014, May 21). Big data for Hollywood. Retrieved October 15, 2014, from http://www.huffingtonpost.com/dave-harvilicz/big-data-for-hollywood b5362220.html Karl. (2014, April 15). Pregnant? Big Data is Watching you. Retrieved October 14, 2014, from http://www.abc.net.au/science/articles/2014/04/15/3985934.htm Rijmenam, M. (2014, May 23). Why UPS spend over $1 Billion on Big Data Annually. Retrieved October 21, 2014, from http://www.bigdata-startups.com/BigData-startup/ups-spends-1-billion-big-data-annually/ Skorupa, J. (2013, December 17). Top POS Prediction for 2014. Retrieved October 23, 2014, from http://risnews.edgl.com/retail-news/Top-POS-Predictions-for-201490104 洪敬浤&修瑞瑩(民99年2月14日)。年關宅配塞車爆退貨潮。聯合報,民103年10月17日,取自:http://city.udn.com/54543/3862463
摘要: 
With the development of information technology, the amount of global data has been explosively growing; the traditional tool of analysis is no longer applicable. Consequently, the new concept of Big Data is generated. Big Data has three characteristics: Volume, Velocity and Variety. Therefore, we need to use specific algorithms to analyze Big Data. The successful examples of applying Big Data to businesses include Target, Hollywood Movies and FBI. Logistics management is one of the most important issues in supply chain management. To confront the keen competition of global logistics, logistics service providers also need the new analytical skills. Efficiently managing trucks makes a logistics company more competitive. Imprecise prediction will result a company in loss by losing a large number of orders.

The traditional prediction methods guess the future situation through historical data or experience, which cannot respond real time demand. This study intends to create a new truck demand forecasting model. At the first, applying logistic regression model to predict the probability of manufacturer making order to logistics centers. After that, we continue to predict the demand of freight by combining historical and POS data. Eventually, we use dynamic linear model to update the parameters. We consider two situations: multiple manufacturers with a single destination and multiple manufacturers with multiple destinations.

The results find our new prediction method is better than the traditional prediction method. In the situation of multiple manufacturers with a single destination, the accuracy of logistic regression model is above 90%, and the MAPE of freight demand prediction model is between 15% and 16%. In the situation of multiple manufacturers with multiple destinations, the accuracy of logistic regression model is above 80%, and the MAPE of freight demand prediction model is under 20%. The results prove that, compared to the dynamic models, our new prediction method can improve the accuracy of prediction.

隨著資訊科技的發展,全球的資料量呈現爆炸性的成長,傳統分析資料的工具已不再適用,進而衍生出大數據(Big Data)的觀念,而所謂大數據為需處理的資料過於龐大,必須透過特定的演算法來運算,大數據的資料具有三大特性:大量、高速、多變。使用大數據成功的案例非常多,像是Target、好萊塢電影、美國聯邦調查局。在供應鏈管理中,物流管理為最重要的環節之一,面對全球物流中心的激烈競爭,物流業也非常迫切需要大數據的分析工具,有效率地管理好貨車是競爭優勢來源之一,如果因為需求預測不準確,會導致企業失去大量訂單進而虧損。

傳統的需求預測方法大多為利用過去的歷史資料或經驗來猜測未來可能發生的情況,此種預測方法無法反應即時的需求狀況,本研究將發展一套新的貨車需求量預測模型,首先利用邏輯斯迴歸模型預測製造商向物流中心下訂單要求運輸服務的機率,接著利用結合歷史訂單與POS累積銷售資料之模型預測貨物配送需求量,最後再以動態線性模型來更新模型裡的參數,並以多個製造商且單一與多個配送點兩種情況來說明。

從數值案例結果發現,本研究所提出的預測方法比傳統預測方法的準確度高,在多個製造商且單一配送點的部分,邏輯斯迴歸預測模型高達90%以上的正確率,貨物配送需求量預測模型的平均絕對百分比誤差皆在15~16%;而在多個製造商且多個配送點的部分,邏輯斯迴歸模型高達80%以上的正確率,貨物配送模型的平均絕對百分比誤差皆在20%以下,證明使用本研究所提出的動態線性模型預測方法確實能夠比傳統的預測方法更好。
URI: http://hdl.handle.net/11455/92325
Rights: 同意授權瀏覽/列印電子全文服務,2015-07-30起公開。
Appears in Collections:企業管理學系所

Files in This Item:
File Description SizeFormat Existing users please Login
nchu-104-7102023028-1.pdf1.63 MBAdobe PDFThis file is only available in the university internal network   
Show full item record
 
TAIR Related Article

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.