微型出貨中心訂單履行問題之研究

倉庫主要功能為訂單履行，而在倉庫當中揀貨作業為成本最高且最耗時的工作。然而，當今電子商務發展蓬勃，除了訂單型態的轉變之外，消費者的要求與期待也日益漸增，尤其是業者的配送速度，絕大多數的消費者都希望業者能夠在下單的當天即出貨，在競爭激烈的市場上，唯有滿足，甚至超出消費者的期待才能在市場上有一席之地。
為了達到速配目標，發展微型出貨中心為一解決方案。微型出貨中心能夠從本質上縮短商品與消費者之間的距離，再透過一些揀貨作業效率提升的方法，例如批次揀貨、路徑規劃與批次指派等，縮短揀貨處理總花費的時間，進而大幅降低商品配送至消費者這個過程所耗費的時間。
本研究採用群組技術、分群演算法與演化演算法進行揀貨問題的探討，提出一有效機制來提升揀貨效率，分別處理訂單批次、路徑規劃以及批次指派問題。最後利用實驗設計，在實驗案例的情境下，經由實驗證實本研究所提出之方法能夠有效地改善揀貨效率。尤其是執行批次揀貨和路徑規劃，除了能夠減少二分之一以上的總到訪揀貨點個數之外，還能夠減少約10~20%的總行走距離。由於減少至少一半重複到訪揀貨點以及行走路徑的浪費，因此，揀貨總完工時間得以大幅下降，進而實現速配此一目標。

The major task of warehouses is order picking which is time-consuming and costly. The boom of e-commerce has rapidly risen, in addition to the innovation in the types of orders, the demands and expectations of consumers, particularly in the demands of fast delivery. Most consumers wish that merchants can ship the ordered goods at the same day when the order list is completed. In the modern e-commerce marketing, satisfying and even passing the expectations of consumers is the only pathway to be a survival in the market.
In order to achieve fast distributions, the construction of micro-fulfillment centers becomes one solution to this end. The micro-fulfillment centers can shorten the distances between the ordered goods and consumers. In addition, they are capable of improving the picking efficiency by means of batch handling, route planning and batch assignment. Improving the time costs spent in picking can largely reduce the delivery time for consumers.
This paper adopted grouping algorithm, clustering algorithm and evolutionary algorithm in picking. This study proposed a novel design to improve the picking efficiency by means of batch handling, route planning and batch assignment. Our experimental results showed that the proposed designs in this study can significantly promote the picking efficiency. Batch handling and route planning can remarkably reduce the total amount of visited picking locations and shorten the total walking distance. Thanks to these reductions, the goal of fast distributions is achieved.