預測週期和預測方式對長鞭效應的影響

Abstract

隨著全球化貿易日益頻繁，供應鏈管理這個議題也就越燒越熱。各家企業都希望藉由有效的供應鏈管理來做到在正確的地方以及準確的時間將正確數目與品質的產品送達顧客手中。在供應鏈中有一種常見的現象，此種現象稱之為「長鞭效應」，所謂的長鞭效應就是供應鏈下游需求有些微的改變，會造成供應鏈上游訂購量及存貨量較大的波動，且是越往上游波動起伏越大。而本篇論文即是在探討預測週期和預測方式對長鞭效應所造成的影響。 本篇論文所探討的供應鏈形式為一直線的供應鏈形式，其中包含了終端需求、零售商、批發商、製造商和供應商。在論文中使用兩種預測週期和三種預測方式組合成六種情境模式，經由各情境模式的構建來揣摩供應鏈各層級的需求預測與訂購行為。接下來模擬六種情境模式，並由模擬當中收集相關的資料，最後使用實驗設計來做資料分析。 由本論文的結果可知，預測的使用確實會對長鞭效應造成影響，而且使用不同的預測週期和預測方式的組合，會對長鞭效應造成不同程度的影響。

Due to the frequent and fierce global trades nowadays, issues of the supply chain management are getting more and more popular. Enterprises want to deliver their customers products of the right quantity and quality in the right place and on the right time by utilizing the supply chain management practices. There is one common phenomenon in the supply chain called “Bullwhip Effect”. The so-called Bullwhip Effect is that tiny changes of the demands downstream will cause enormous fluctuation upstream. The closer to the upstream, the bigger fluctuation. This thesis discusses the influence of forecast periods and forecast methods on the Bullwhip Effect. The model of the supply chain used here is a straight-line and simplified one, including one retailer, one wholesaler, one manufacturer and one supplier. With the combinations of two forecast periods and three forecast methods, it produces six scenarios. These scenarios are employed to simulate the possible demand forecasts and behaviors within each layer in the supply chain. Design of Experiment (DOE) is then adopted to analyze the data collecting through simulation. The experimental results indicate that the forecast indeed affect Bullwhip Effect. In addition, the combinations of forecast periods and methods also influence the Bullwhip Effect to certain extent.