Abstract: In the context of customer can choosing different delivery timeslots and methods,providing an alternative delivery option set for the rejected demands is one of the key decisions to balance customer satisfaction and cost.Accordingly,the customer's substitution behavior of choosing a delivery option was analyzed based on the nested logit model.Then,the multi-objective optimization model to design the alternative delivery options set was constructed to minimize the distance and maximize the cumulative transfer rate.It showed that the fast elitist Non-dominated Sorting Genetic Algorithm (NSGA-Ⅱ) was superior in the distance and the number of vehicles for the case of more customers and their distribution showing a combination of dispersion and aggregation;as the time window threshold increased,the total revenue gradually decreased with the significant decline of attended-home-delivery revenue and the slow increment of reception-box-delivery revenue;there was no linear relationship between the minimum constraint of the cumulative transfer rate and the income;by decreasing the time window threshold and increasing the minimum constraint of the cumulative transfer rate,the total revenue could be improved.

Key words: urban delivery, customer substitution, attended home delivery, reception box delivery, time slot, dynamic vehicle routing, multi-objective optimization

摘要： 在顾客可选末端交付方式和时间窗等服务选项的情境下,为被拒绝的配送需求提供备选服务选项集是供应商平衡顾客满意度和服务成本的重要措施之一。首先使用嵌套Logit模型分析顾客对配送服务选项的替代选择行为,然后构建考虑期望距离最小和顾客替代选择累计转移率最大的城市配送备选服务选项集多目标优化模型。仿真发现:在配送点分布呈分散和聚集结合的情形下,非支配排序遗传算法更适用于配送点数量更多的算例,可以有效降低车辆行驶距离、车辆数量等成本因素;随着时间窗偏差阈值增大,总收益逐渐下降,其中送货上门服务收益下降趋势明显,自提柜服务收益则震荡上升;顾客替代选择累计转移率最小约束值对收益的影响不存在线性关系;在时间窗偏差阈值较小、顾客替代选择累计转移率最小约束值较大的区域,往往能获取更高的总收益。

关键词: 城市配送, 替代选择, 送货上门交付, 自提柜交付, 时间窗, 动态车辆路径, 多目标优化