外卖智能派单的多主体利益协调机理研究
The Mechanism of Multi-agent Cooperation on the Food Delivery Platform
熊浩 1鄢慧丽2
作者信息
- 1. 海南大学国际商学院,海口 570228
- 2. 海南大学国际旅游与公共管理学院,海口 570228
- 折叠
摘要
通过研究外卖智能派单中多主体的利益目标相互影响机理,以及动态订单分派的多目标在线优化机理,为破解骑手的"外卖算法困局"提供一定的理论参考.研究表明:第一,平台利润目标需要结合订单的抽成规则和支付给骑手的提成规则;第二,顾客的订单履约时间构成比较复杂,需要考虑商家备餐时间、骑手到商家行驶时间、骑手从商家到用户的行驶时间等多种因素;第三,派送服务质量目标可以为顾客优先分配服务水平高的骑手,既能提升服务质量,也能提高具有高服务水平的骑手的收益;第四,引入骑手等待时间和骑手空驶时间等目标,可以进一步提高骑手的工作效率;第五,外卖智能派单不仅是订单分配与配送路径的集成优化问题,而且是量纲不同、重要程度不确定的多目标优化问题,可以利用各个目标的预设值与实际值的差距动态调整权重,建立基于插值法的自适应邻域搜索启发式算法.
Abstract
This paper attempts to reveal the objectives of the multi-agent in the smart delivery system of the food delivery platforms and the multi-objective online optimization mechanism of dynamic smart delivery to provide a theoretical refer-ence for solving the"algorithm trap"of smart delivery.The results show that first,the payment rule from the sellers and the payment rule to the riders for each order should be considered in the profit objective of the platform.Second,the composi-tion of the delivery time of the customer's order is relatively complex and needs to consider many factors,such as the meal preparation time of the seller,the driving time of the rider to the seller,and the driving time of the rider from the seller to the customer.Third,the objective of delivery service quality can give higher priority to the riders with higher service levels to be assigned to the customers,which can improve the service quality and enable the riders with higher service levels to im-prove revenue.Fourth,the rider's work efficiency can be further enhanced by introducing the objectives of the rider's wait-ing and idle times.Fifth,smart delivery is related to order-rider matching and vehicle routing problems,which are multi-objective optimization problems.The units of these objectives are different,and the importance of these objectives is uncer-tain.So,a novel adaptive neighborhood search heuristic algorithm is based on an interpolation method and a dynamic weight adjustment method,which adjusts the objective weights according to the gap between the preset value and the actual value.
关键词
外卖智能派单/外卖骑手/协调/多目标优化/算法Key words
smart food delivery/food delivery rider/cooperation/multi-objective optimization/algorithm引用本文复制引用
基金项目
国家自然科学基金项目(71761009)
国家自然科学基金项目(72061010)
海南省自科高层次人才项目(722RC646)
海南省自科高层次人才项目(720RC570)
教育部人文社科研究项目(22XJA630008)
海南省哲学社会科学课题(HNSKYB23-11)
海南省教育厅重点项目(Hnky2020ZD-6)
海南省教育厅重点项目(Hnky2019ZD-10)
出版年
2024
NETL
NSTL
万方数据