代工製造業產品需求預測模式

在資訊科技高速發展的產業競爭環境中，原物料訂購量、在製品掌控、快速生產出貨、存貨控管以及需求預測的準確與否都影響企業競爭優勢。其中需求預測，綜合了庫存、銷售、需求等資訊，對企業採購生產的影響甚大。現有企業採用需求預測如MRP、Demand Forecast，其預測數據來源多為內部存貨、採購、生管及銷售部門資料。本研究針對需求預測尋求外部可能資訊，考量環境面資料(如景氣領先指標、景氣同時指標、景氣對策信號)以及產業面資料(如外銷訂單指數、製造業存貨量指數、工業生產指數、製造業設備利用率、製造業成品存貨率、製造業新接訂單指數及製造業銷售值等等)，以倒傳遞類神經網路建構出不同面向的需求預測模式，搭配MAE、RMSE、Mean及Std Dev方法評估各面向對預測模式的相關性，以及預測模型的準確性，作為企業需求預測的參考。

關鍵字

原物料訂購量；在製品掌控；快速生產出貨；存貨控管；需求預測模式

並列摘要

During the violent industry competition with IT technology rapidly expanded, precisely dominating the raw material purchase amount, monitoring WIP(Work In Process), quickly manufacturing and delivering, inventory control and demand forecast impacts enterprise’s competitive advantage. Demand forecast, composited by inventory, sales, demand etc, impacts enterprise’s purchase and manufacture most. Nowdays the method adopted to forecast the demand such as MRP, demand forecast, all forecast with internal information from departments like inventory, purchase, manufacture and sales. With seeking for external information such as environment attributes (eg. Composite Leading Index, Composite Coincident Index, Monitoring Indicator) and industrial attributes (eg. Index of Export Orders, Index of Producer’s Inventory, Industrial Production Index, Manufacturing capacity Utilization Rate, Manufacturing Inventory-to-Sales Ratio, Manufacturing New Order Index and Manufacturing Sales), this research hope to construct a different demand forecast model for enterprises by Back-Propagation Neural Network, which evaluated by MAE, RMSE, Mean and Std Dev to verify its forecast precision.

並列關鍵字

raw material purchase amount ； WIP(Work In Process) ； quickly manufacturing and delivering ； inventory control ； demand forecast model

參考文獻

[2]Ronald E Miller, Peter D Blair, Input-Output Analysis: Foundations and Extensions, Englewood Cliffs, New Jersey: Prentice-Hall, 1985.

[3]Christopher M. Bishop, Neural Networks and Machine Learning, Berlin, New York, Springer, 1998.

[5] Min Qi, Guoqiang Peter Zhang, “An Investigation of Model Selection Criteria for Neural Network Time Series Forecasting”, European Journal of Operational Research, vol.132, pp.666-680,August 2001.

[6] I. Kaastra and S. B. Milton, “Forecasting futures trading volume using neural networks”, The Journal of Futures Markets, pp.853–970, 1995.

[7]Colombo J. C., Silverberg N., Gearing J. N., W-C C., Urban T., Baldridge G.,,” A neural network approach to mutual fund net asset value forecasting”, Omega , vol.24, pp. 205–219, Apr 1996.

國際替代計量

代工製造業產品需求預測模式

全文下載

主題瀏覽