基于强化学习的在线订单配送时隙运能分配

Capacity allocation of delivery time slot for online orders based on reinforcement learning

为解决在线订单配送效率低、时隙运能分配不均衡和顾客满意度不高的问题,考虑价格和交付期对消费者选择行为的影响建立Logit模型,采用强化学习结合时隙运能分配特点对到达的订单群进行运能分配.算例模拟结果证明:采用强化学习能使每个时隙每辆车的运能分配均衡,且分配方法符合消费者的行为偏好;消费者对时隙价格偏好程度越高商家收益就越低.结论验证了采用强化学习解决时隙运能分配问题的可行性和有效性.

In order to solve the lower efficiency of online order delivery,the unbalanced capacity allocation of time slots and the lower customer satisfaction,the Logit model is established considering the influence of the price and lead time on the selection behavior of consumers. Considering the character of capacity allocation of time slot,the orders are assigned to the vehicles by the reinforcement learning. The example simulation results show that： the capacity of every time slot and every vehicle can be balanced by the reinforcement learning and the allocation method accords with the behavioral preference of consumers;the more attention consumers take to the price of time slot,the lower profit retails can get. The conclusion verifies feasibility and effectiveness of adopting the reinforcement learning to solve the capacity allocation of time slot.