網路購物公司配送中心搬遷方案之研究

Abstract

近年來，民眾使用網路購物的比率大為上升。隨著網路購物市場商機增加，各公司為了滿足消費者的需求，開始在各地興建倉儲配送中心(distribution center)，並將舊有的倉儲配送中心搬遷、擴張，以符合市場之需求。當消費者在網路上進行購買行為時，倉儲配送中心之供應鏈(supply chain)系統即會進行商品配送，以期在特定時間內，滿足消費者的訂單需求。 本研究即探討，當網路購物公司之倉儲配送中心需要進行擴廠搬遷時，如何制定出一套搬遷流程以降低這段時期所衍生的成本。由於新、舊倉分別負責不同時期之市場需求，所以除了搬遷運輸成本外，往來於新、舊倉之間的側向緊急運輸(lateral emergency transshipment)補貨成本、訂單完成率之高低、搬遷過程中造成的商品擺放錯誤成本等等，皆為本研究制定搬遷政策時所關注的重點。 搬遷政策(moving policy)的制定上，有將各商品直接排序之方法，如：體積大小、價錢高低、出貨頻率高低等等；亦有依市場需求變動和歷史銷售資料所制定之改良搬遷方案。搬遷政策的制定除了降低運輸成本外，該政策能夠使搬遷過程迅速準確、流程方便簡潔、不降低消費者的服務水準，亦為衡量政策優劣的標準。 本研究將建構一確定型整數規劃模型，以描述此問題並進行最佳每日商品搬遷數量之求解；而系統模擬模型的建構，則將對整個搬遷流程有更詳細的描述。在缺貨商品數量、訂單完成率、複雜度的衡量上，也可以有準確的衡量。 各績效衡量項目之間，會有不同的成本結構；所處的環境也可能會有不同，如：每日運輸趟次的限制。本研究將探討在不同成本結構或不同環境下，搬遷政策的選擇上是否會有改變。藉由不同情境和成本下之系統模擬，面對此決策問題的公司，即可彈性的在不同狀況下，選擇出適宜的搬遷方法。而針對此問題所撰寫之程式，亦將介紹於附錄中。

This research studies the logistics activities when a retailer is in the transition of moving its current distribution center (DC) to a new location. Because of expanding business, an online store is required to build a more spacious DC to sustain the increasing customer orders. During the early stage of the transition period, customer orders are still fulfilled by the current DC. At the same time, inventory has to be gradually transported to the new DC. If a product is moved to the new DC but an order of that product arrives, emergency shipping is required to move the product back to the current DC and an extra cost incurs, Given the inventory levels of products, the customer forecasts and the capacity of trucks, the online retailer has to determine a logistics plan of moving items to the new facility so that the cost is minimized including shipping cost, the penalty of shortage, and the penalty of complicated moving operations. A deterministic integer programming model is formulated. Although this model can obtain an optimal solution, it is time consuming to solve a real-world problem in a large scale. In additions, demand orders in fact are uncertainty and the objective such as complexity of packing items in a truck is not simple to quantify. Therefore, a system simulation model is constructed, which can authentically reflect the results of practical moving policies without the drawbacks as mentioned in the mathematical model. Moving policies’ performance can be measured by different criterion, such as order fulfill rate, the number of emergency shipments, complexity. This research compares the simulation results them to provide the decision makers the most suitable moving policy under each condition.