限制理論之需求拉動補貨策略於半導體新產品存貨管理之應用

Abstract

存貨管理是影響公司績效重要的一部分，而限制理論以需求拉動補貨策略搭配緩衝管理之方法經過去中、外文獻驗證為一個有效的管理方法，但較少在定義於不同產品之使用時機與補貨策策略參數之定義，以及新產品使用之方法。本研究根據舊產品之需求資訊與產品特性建構產品導入需求拉動補貨策略之方法，首先將定義本研究之產品需求型態因子及產品特性因子，分別為描述產品需求特性及產品特性之因子，接著將產品進行限制理論之需求拉動補貨策略模擬，其模擬結果可得到產品之策略參數組合及對應的績效指標，再來將產品分群，本研究分群可分為兩階段，階段一利用統計方法將產品依需求資訊而分為需求型態群，接著根據產品模擬結果之績效指標將產品分群而得到最後的分群結果，而針對屬於適合之產品群再利用變異數分析找出其合適的補貨策略參數，最後使用決策樹分析建構產品群之產品特性及需求特性之分類模型，利用此分類模型管理者可以快速判斷產品使否適合限制理論之需求拉動補貨策略及合適的補貨策略參數，並且新產品亦可根據產品特性分類方法找出對應的產品群，而都可得到良好的效果。本研究分析台灣某晶圓代工廠所提供之資料，根據本研究方法可得知不同需求型態之產品與限制理論之需求拉動補貨策略之參數是有關係的，並且不同的需求型態產品亦可找出產品特性規則，可利用在新產品存貨管理利用上，因此本研究成果可作為相關企業實務參考之依據，使企業能做好存貨管理增加競爭優勢，並提升企業管理效能及利益。

Inventory control is an important part of company performance. Theory of constraint is an effective method which combines demand-pull replenishment and buffer management to manage inventory, and it has been verified by literatures in past years. But there are limited literature researching on the timing and parameters of demand-pull replenishment policy, and application rule of new product. This study construct product application rule of theory of constraint replenishment policy by demand information and product characteristics of old products. This study first identifies demand pattern factors and product characteristics factors which describe demand property and product characteristics. And then it will use simulation of demand-pull replenishment policy on product historical demand data. Simulation results include parameter set of demand-pull replenishment policy and performance indicators. Next step will group the products which contain two steps; First step using statistical method grouping product which based on demand information. Second step will use performance indicators of simulation results and obtain the final grouping results. ANOVA analysis will use to identify suggestible replenishment policy parameters of individual product cluster. The final procedure is to construct product's classified model by demand factors and product characteristics which managers may have good decision ability on product's replenishment parameters, regardless product are old or new. This study use practical demand data on 37 products provided by a Taiwan wafer foundry company to demonstrate the proposed research methodology and found that demand pattern of product have relationship with demand-pull replenishment policy. Also we can find out product's demand pattern by its product characteristics by the model no matter the product is old or new. The study results can be a reference in practice for wafer relative industries for inventory management.