Abstract: To make the order acceptance or rejection decision efficiently in the context that the customer can choose the last-mile delivery modes and time slots, the framework of dynamic order acceptance in urbandelivery is constructed, which is composed of pre-routing,assessment of delivery requests, adjustment of order acceptance strategy and global optimization. An order valuation method based on the threshold of time window deviation is proposed. Then three different algorithms for adjustingorder acceptance strategy are designed including accepting all orders, allocating delivery modesstatically and allocating service options dynamically. The simulation results show that the dynamic allocation of service options could obtain higher profit than the other algorithms and spent less time especially in the example with more delivery requests; as the time slot range is increased, the revenue of reception box service and the total distance are gradually decreasing, however the revenue of attending home delivery service and the profit are increasing; the threshold of time window deviation has a significant effect on the profit of distribution service, but there is no trend; the threshold of time window deviation with higher revenue of attending home delivery often make a higher profit. The results may support the decisions of allocating the last-mile delivery modes and time slots in different distribution areas.

Key words: urban delivery, last-mile delivery, time slot allocation, dynamic order acceptance, insertion algorithm