中文摘要：

在电子商务在线订单拣选系统中,订单到达时间和订购商品等信息未知。针对拣选设备容量、员工人数等资源有限约束情况,研究在何时、对多少订单进行分批优化,以保证在订单完成期限前以最短的时间拣出最多的订单。构建考虑订单完成期限的在线订单分批混合整数规划模型,以最小化平均有效订单服务时间,采用改进的固定时间窗订单分批启发式规则求解模型,定义剩余操作时间在[前置时间,配送准备时间]内的订单为紧急订单,构建综合考虑紧急程度和相似度因素的在线订单分批算法。采用某配送中心14：00~18：00时间段内以泊松分布（λ=17）随机生成的订单进行数据实验,将实验结果与传统固定时间窗在线订单分批算法进行比较。研究结果表明,考虑完成期限时,系统拣选配送的订单数量更多,总服务时间和平均有效订单服务时间更短,且出现延迟订单的数量更少,延迟时间更短。拣选员工人数的增多在不同程度上提高配送率,且考虑完成期限时,配送率提高幅度要大于传统算法;但随着人数的增加,配送率的提高幅度呈降低趋势。

英文摘要：

In the e-commerce on-line order picking system, customer orders＇ arrival time and goods cannot be informed in advance. With the constraints of picking equipment capacity and pickers＇ number, the order batching optimization approach, which focuses on the batehing time and batching strategy, is proposed to pick out maximum orders in the shortest service time before the due time. The on-line order batehing mixed-integer programming model considering orders＇ due time is established to minimize the valid average service time of distributed orders. To solve this problem, the improved fixed time window order batching algo- rithm is proposed. The order is identified as the urgent one if its remained operation time is between lead time and distribution setup time. Based on orders＇ different urgent level, we propose the on-line order batching rules while taking into account urgent degree and similar degree. Through a series of experiments where the orders are generated from 14：00 to 18：00 based on Poisson distribution （A = 17）, we compare the results with ones of traditional on-line order rules. Several enlightening findings are dis- covered： If considering orders＇ due time, the number of distributed orders is bigger, the batches＇ total service time and the dis- tributed batches＇ valid average service time are shorter, and the number of delayed orders is smaller and delayed time is shorter. Meanwhile, if considering orders＇ due time, with the increase of the number of order pickers, the delivery rate improves in different .degree, and the increase of delivery rate is larger than the one of traditional rules. However, the increasing of delivery rate is a progressive decline.