為因應網路消費的興起與全球經濟的自由化，物流業須發展出更有效率的作業方式，來符合更高的顧客服務水準與品質。越庫作業是物流作業中的一種策略，其重點為讓供應商將不同品項由進貨月台送進物流中心後，將各品項指派至暫存區儲存格中，並由揀貨人員依不同訂單的需求於理貨區儲存格中揀貨，最後將完成的訂單移送至出貨月台並配送至顧客端。越庫作業的倉儲環境中，可減少品項入庫、儲存及檢驗的作業，因此使品項從進貨到出貨的時間最小化。 本研究針對配銷商型越庫（Distributor Cross Docking），使用播種式揀貨於理貨區為多區塊儲位佈置的環境下，對四項決策問題進行探討。分別提出播種品項選取法則、理貨區之貨車訂單儲位指派法則、播種點選取法則，及播種揀貨路徑規劃。透過模擬實驗數據，分析在此環境中的績效指標如總系統執行時間（Total System Time）、總揀貨時間（Total Picking Time）與總揀貨距離（Total Picking Distance），並找出各績效指標下的最佳因子組合，何者能對揀貨作業有最顯著的績效提升。 透過模擬實驗及統計分析，歸納出於總系統執行時間與總揀貨時間下的最佳因子組合為『靠近暫存區由中間且沿走道兩側依序指派』的訂單指派法則搭配『最大指數型出貨訂單緊急值總和』的品項選取法則搭配『需求數量最多的單一儲存格優先播種』之播種點選取法則，以及『合併最大間隙策略』之播種揀貨路徑規劃。而於總揀貨距離績效下的最佳因子組合為『靠近暫存區由中間且沿走道兩側依序指派』搭配『最短估計播種揀貨旅行距離』搭配『需求數量最多的單一儲存格優先播種』，以及『合併最大間隙策略』。 ;With the rise of on-line shopping market and the liberalization of global economy, the logistics industry has to develop a more efficient model to meet higher customer service level and product quality. Cross docking is one of the strategies of logistics operations. After suppliers transport different kinds of products to distribution center through receiving dock, products will be assigned to different temporary storage area. Then pickers pick the items to tally area according to the orders. Finally, the finished orders will be moved to shipping dock and transported to the customers. In the environment of cross-docking warehouse, the operation of inbound process, storage and examination can be eliminated, therefore the time that products go through from inbound to outbound can be minimized. This study focuses onDistributor Cross Docking.We investigate put-to-light picking problem under the environment with multi-block tally area storage location layout. We proposefour decision problems“the selection rules of items in temporary area”, “the assignment rule of order storage location in tally area”, “the selection rules of picking location” and “picking route planning.” The data is obtained by simulation software, Arena. We analyze the performance index such as total system time（TST）, total picking time（TPT）and total picking distance（TPD）in our environment. Finally, the best combinations of rules under different performance index are discovered. It’s observed that the best combination under performance index, TST and TPT, would be NTA× ESOTVMSQCLG. And the best combination under performance index of TPD would beNTASTDMSQCLG, whichcan make the best enhancement in the system.
