TFT-LCD面板運輸模式之研究

Abstract

TFT-LCD面板產業為接單組裝製造型態，TFT-LCD面板廠商於一個月前得到顧客之需求量預估值，並於得到顧客最後確定訂單才進行運輸作業之規劃。雖然此做法能夠即時地根據顧客訂單做出貨安排，卻缺乏考量整體物流流程的可變化性與彈性，造成整體物流效率不佳、出貨作業集中於月底，於運輸旺季時，往往無法訂到海運艙位、必須以較昂貴之空運運送。 本研究以顧客當月約略之需求量做中短期策略面之整合規劃，整合生產與運輸兩個流程，使整體物流作業更有效率，並且能夠提早預訂艙位，以避免旺季時訂不到艙位而以較昂貴之空運運送或延遲出貨。因此以數學規劃構建一模式，作為整體製造流程與運輸流程上之規劃工具；對於顧客之需求量，分為預估之需求量與臨時增加之需求量以表達預估之需求量與最後實際需求量之差異，並以滾動連續求解方式逼近實務上顧客需求量常常異動之情形。 經以國內某家TFT-LCD面板廠商提供之相關參數資料與顧客需求型態為案例分析，並以LINGO8.0軟體求解問題，結果顯示並非所有貨物都以海運運輸為最優，而是七成至九成之貨物以海運運送，其他則以空運運送；由於產能有限，廠商應對不同顧客地區之海運運輸量做不同之順序分配，於需求量非高峰時，優先分配給海運運輸班次較少之地區，其次為海運與空運運輸價差大之地區，最後為海運運輸班次頻繁且海運與空運運輸價差小之地區。於需求量尖峰之情形，應優先分配給海運與空運運輸價差大之地區，其次為海運運輸班次較少之地區，最後為海運運輸班次頻繁且海運與空運運輸價差小之地區；若臨時增加需求量偏高，需求量尖峰時其分配順序仍相同，但需分配更多數量之海運運輸量至海運與空運運輸價差大之地區；TFT-LCD廠商為因應臨時增加之需求量較多之情況，應具備較高之存貨水準維持對於顧客的供貨水準；該模式應用簡單且適用性高，若搭配實際之需求量資料其結果應符合實務界營運之參考。

The make-to-order production model is widely adopted by the TFT-LCD manufacturers. Although the preliminarily estimated number of orders can be obtained one month earlier than the delivery date, TFT-LCD manufacturers require the precise number of orders, which in most cases could not be reached until 7 days prior to delivery, to arrange production and transportation activities to meet the demand of the customers. The main drawback of this current practice for TFT-LCD manufactures is that insufficiencies of lead time to plan production and transportation have caused inefficiency in logistics performance, i.e. inadequate inventory or higher transportation costs. This study attempts to solve the aforementioned problems by employing an integer mathematical programming model to better schedule production and transportation activities on the basis of the estimated number of orders, coupled with a rolling-over approach to tackle the problem that in reality customers submit their finally determined orders in a short lead time. Utilizing the investigated information from one large TFT-LCD manufacturer in Taiwan as the input data for parameters and TFT-LCD demand patterns of the model, and employing LINGO 8.0 as the solving software for the model, a number of empirical studies based on four scenarios have been performed. The results showed that the total costs (including production and transportation) were minimized by delivering 70-90% of TFT-LCDs by sea and the rest by air, rather than entirely by sea. In respect of TFT-LCDs transported by sea during low ordering season, the products were firstly offered to the destinations where the frequency of voyages was lower, secondly assigned to those where the difference of costs between air transportation and sea transportation (thereafter transportation cost difference) was greater, and finally provided to the locations where more frequent voyages and flights were available and where transportation cost difference was lesser. However, during peak season, the first priority was to provide products to those locations where transportation cost difference was greater, and the next to deliver products to destinations where voyages were fewer. The same results could be obtained in situations where the final order number was relatively greater than that of the preliminary estimation, with a need for TFT-LCD manufactures to reserve a higher level of inventory and to transport more products by sea to destinations where transportation cost difference was greater.