以品類管理為基礎之紡織成衣業CPFR銷售預測模式

A CPFR Sales Forecasting Model Based on Category Management for the Textile and Apparel Industry

黃文俊(Wen-Chun Huang)

指導教授 ： 陳銘崑

國立臺北科技大學/管理學院/商業自動化與管理研究所/碩士(2008年)

https://doi.org/10.6841/NTUT.2008.00093

摘要

因應紡織成衣業面對全球化競爭，以及在WTO體制下從2005年起配額制度取消等因素，使得紡織成衣業急需轉型，公司與產地間的合作需要更加的密切、產業全球運籌的能力需要提升、並能夠快速反應客戶以及市場需求、掌握並預測市場趨勢，因此紡織成衣業需要一個新的商業模式導入來協助他們解決問題。
而CPFR (Collaborative Planning, Forecasting and Replenishment)是由VICS協會於1998年所提出。其主要強調透過供應鏈上中下游協同合作以及資訊分享的概念來進行協同預測，以提高預測之準確性，並提升供應鏈上流程的處理效率。
本研究參考過去文獻與產業實務，以紡織成衣業為例，提出一以品類管理為基礎的CPFR模型，其能夠在建立協同商業計畫步驟時將產品先做分類，並針對各品類發展其適合的CPFR協同銷售預測模式。其中在產品分類階段，本研究使用自我組織映射網路(Self Organization Map, SOM)來針對各品類的特性進行分類的動作，而後使用倒傳遞類神經網路(Back-propagation network, BPN)來對各個品類發展不同的銷售預測模式，以提升銷售預測的準確度；同時以實際企業案例作驗證，希望能作為紡織成衣業施行之參考。

關鍵字

CPFR ； 品類管理 ； 銷售預測模式 ； SOM ； BPN

並列摘要

In order to have reacted to cancel the quota restrictions by WTO since 2005, and the globalization of competition, the textile and apparel industry needs the closer cooperation between the company and the place of the production, the better ability to management the global logistics, the quicker response to the customer and the demand from the market, and to handle and forecast the trend to the market. Therefore, the textile and apparel industry must have a new business model to help it.
The concept of CPFR (Collaboration Planning, Forecasting, and Replenishment) proposed by VICS (Voluntary Inter-industry Commerce Standards) in 1998 had been adopted in many enterprises and it had proved it`s performance. The main spirit of CPFR is to emphasize the sharing of the information among participating partners, and through these information to deal with the Collaboration forecasting. Following this process it can improve the accuracy of forecasting and the efficiency in supply chain.
In this study we proposed a CPFR sales forecasting model for the textile and apparel industry. This CPFR process is based on a category management process and takes the former literatures and industrial practices as references. Through this process, we analysis different kinds of categories by using the SOM(Self Organization Map) classification algorithm and develop the different kinds of sales forecasting models by using BPN(Back-propagation network) algorithm . Furthermore, we take the real case for testing and verifying. The result of simulation can offer references to enterprise when making practical decisions.

並列關鍵字

CPFR ； Category Management ； SOM ； BPN ； Sales Forecasting

參考文獻

[11] 高鵬虎，「CPFR商業流程導入之研究以國內行動電話維修業為例」，國立台灣科技大學工業管理系，碩士論文，2003。

[1] 王建信，「需求鏈管理協同預測模式之系統動態模擬─以主機板產業為例」，台北科技大學商業自動化與研究所，碩士論文，2007。

[2] 王靜詩，「採用RFID的CPFR商業流程模式之探討－以液晶電視產品為例」，大同大學資訊經營學所，碩士論文，2006.

[8] 吳綺芳，「整合神經網路與灰色理論之預測模型設計」，元智大學工業工程與管理學系，碩士論文，2005。

[14] 張怡筠，「CPFR中滾動式預測的關鍵因子探討」，國立清華大學資訊系統與應用研究所，碩士論文，2005。