面向集中供给与稀疏需求场景的网约车调度优化算法

doi:10.16097/j.cnki.1009-6744.2025.05.025

交通运输系统工程与信息 ›› 2025, Vol. 25 ›› Issue (5): 280-290.DOI: 10.16097/j.cnki.1009-6744.2025.05.025

面向集中供给与稀疏需求场景的网约车调度优化算法

王江锋^*1,2，宋之凡¹，李云飞¹，齐崇楷¹，闫学东^1,3，李青山⁴

1. 北京交通大学，交通运输学院，北京100044；2.北京交通大学，唐山研究院，河北唐山063000；3. 西南交通大学，交通运输与物流学院，成都611756；4.中国雄安集团交通有限公司，河北雄安070001

收稿日期:2025-06-05 修回日期:2025-08-02 接受日期:2025-08-21 出版日期:2025-10-25 发布日期:2025-10-25
作者简介:王江锋(1976— )，男，河北保定人，教授，博士。
基金资助:
河北省创新能力提升计划项目(244X0801D)；河北省省级科技计划项目(236Z0802G)。

Optimization Algorithm for Ride-hailing Dispatch Under Concentrated Supply and Sparse Demand

WANG Jiangfeng^*1,2, SONG Zhifan¹, LI Yunfei¹, QI Chongkai¹, YAN Xuedong^1,3, LI Qingshan⁴

1. School of Traffic and Transportation, Beijing Jiaotong University, Beijing 100044, China; 2. Tangshan Research Institute, Beijing Jiaotong University, Tangshan 063000, Hebei, China; 3. School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 611756, China; 4. China Xiong'an Group Transportation Co Ltd, Xiong'an 070001, Hebei, China

Received:2025-06-05 Revised:2025-08-02 Accepted:2025-08-21 Online:2025-10-25 Published:2025-10-25
Supported by:
The Project of Hebei Province Innovation Capacity Enhancement Plan (244X0801D)；Hebei Provincial Science and Technology Program Project (236Z0802G)。

摘要/Abstract

摘要： 针对部分新兴城区网约车服务在空间分布上存在“集中供给和稀疏需求”的结构性失衡现象，本文引入激励策略优化网约车调度决策，解决供需分布的空间错配。基于强化学习方法提出一种嵌入激励策略的网约车调度优化算法，包括调度环境构建、可行决策筛选、激励策略嵌入和调度决策优化这4个模块。具体而言，基于Markov决策过程构建涵盖订单匹配、空驶调度和充电管理等任务的调度环境，通过订单抽成下调与优先派单保障机制设计“直接经济刺激”“间接规则优化”相结合的联合激励策略，提出以平台收益最大化为目标的Actor-Critic调度优化算法。以雄安新区为例进行实证分析，结果表明，本文提出的算法能够有效提升“稀疏需求”订单的服务率。在网约车投放规模固定的情况下，分别评估订单抽成下调，优先派单保障以及联合激励策略，结果显示，在平台收益不下降且全局订单服务率有所提升的前提下，“稀疏需求”订单服务率分别提高了24.70%、2.53%和26.09%。随着网约车投放规模增加，当投放数量达120辆时，“稀疏需求”订单服务率和全局订单服务率分别达到61.20%和81.55%。对平台抽成比例与优先派单保障期的敏感性分析进一步表明，当抽成比例设置为24%，保障期设置为4个调度周期时，系统在全局订单服务率与平台收益之间实现了较优平衡。

关键词: 城市交通, 激励策略, 强化学习, 网约车, 调度算法

Abstract: This study addresses the structural imbalance of “concentrated supply and sparse demand” in ride-hailing services across emerging urban areas by introducing incentive strategies to optimize dispatch decisions and alleviate spatial mismatches between supply and demand. A reinforcement learning-based dispatch optimization algorithm is proposed, integrating incentive mechanisms into four key modules: environment construction, feasible decision selection, incentive embedding, and policy optimization. Specifically, a Markov Decision Process-based environment is developed to model tasks such as order matching, idle vehicle repositioning, and charging management. A joint incentive strategy combining direct financial incentives (via reduced platform commission) and indirect regulatory adjustments (via prioritized dispatch guarantees) is designed, and an Actor-Critic algorithm is used to maximize platform revenue. Empirical analysis using data from Xiong’an New Area in Hebei Province demonstrates that the proposed method significantly improves the service rate of sparse-demand orders with sparse demand. Under fixed fleet size, the service rates of sparse-demand orders respectively increased by 24.70%, 2.53%, and 26.09% through commission reduction, prioritized dispatch, and joint incentives, while maintaining or increasing overall service rate and platform revenue. As the ride-hailing vehicle fleet expands, the sparse-demand and overall order fulfillment rates reach 61.20% and 81.55% respectively, when the fleet size reaches 120 vehicles. Sensitivity analysis further indicates that a 24% commission rate and a 4-period dispatch guarantee yield a favorable balance between platform revenue and global service performance.

Key words: urban traffic, incentive strategies, reinforcement learning, ride-hailing, dispatch algorithm

中图分类号:

U121

王江锋, 宋之凡, 李云飞, 齐崇楷, 闫学东, 李青山. 面向集中供给与稀疏需求场景的网约车调度优化算法[J]. 交通运输系统工程与信息, 2025, 25(5): 280-290.

WANG Jiangfeng, SONG Zhifan, LI Yunfei, QI Chongkai, YAN Xuedong, LI Qingshan. Optimization Algorithm for Ride-hailing Dispatch Under Concentrated Supply and Sparse Demand[J]. Journal of Transportation Systems Engineering and Information Technology, 2025, 25(5): 280-290.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.tseit.org.cn/CN/10.16097/j.cnki.1009-6744.2025.05.025

http://www.tseit.org.cn/CN/Y2025/V25/I5/280

参考文献

[1] OH S, KONDOR D, SESHADRI R, et al. Exploring the operational characteristics of ride-sourcing in an urban area[J]. Research in Transportation Business & Management, 2022, 43: 100827.

[2] MAO C, LIU Y, SHEN Z J. Dispatch of autonomous vehicles for taxi services: A deep reinforcement learning approach[J]. Transportation Research Part C: Emerging Technologies, 2020, 115: 102626.

[3] 牛振宁,安琨,马万经.车网互动场景下电动网约车运营与充放电动态调度策略[J]. 交通运输系统工程与信息, 2024, 24(4): 50–59. [NIU Z N, AN K, MA W J. Electric vehicle ride-hailing operation and charging-discharging dynamic scheduling strategy in vehicle-to-grid scenario[J]. Journal of Transportation Systems Engineering and Information Technology, 2024, 24(4): 50–59.]

[4] ZHANG A, KANG J E, KWON C. Generalized stable user matching for autonomous vehicle co-ownership programs[J]. Service Science, 2020, 12(2/3): 61–79.

[5] ERDMANN M, DANDL F, BOGENBERGER K. Combining immediate customer responses and car-passenger reassignments in on-demand mobility services [J]. Transportation Research Part C: Emerging Technologies, 2021, 126: 103104.

[6] AL-KANI L, NASCIMENTO J, POWELL W B. Approximate dynamic programming for planning a ride-hailing system using autonomous fleets of electric vehicles[J]. European Journal of Operational Research, 2020, 284(3): 1088–1106.

[7] FENG S, DIJAN P, KE J, et al. Coordinating ride-sourcing and public transport services with a reinforcement learning approach[J]. Transportation Research Part C: Emerging Technologies, 2022, 138: 103611.

[8] YANG H, SHAO C, WANG H, et al. Integrated reward scheme and surge pricing in a ridesourcing market[J]. Transportation Research Part B: Methodological, 2020, 134: 126–142.

[9] GARC N, NAZERZADEH H. Driver surge pricing[J]. Management Science, 2022, 68(5): 3219–3235.

[10] VALADKHANI A H, RAMEZANI M. Dynamic ride-sourcing systems for city-scale networks, Part II: Proactive vehicle repositioning[J]. Transportation Research Part C: Emerging Technologies, 2023, 152: 104159.

[11] KE J, QIN X, YANG H, et al. Predicting origin-destination ride-sourcing demand with a spatio-temporal encoder-decoder residual multi-graph convolutional network[J]. Transportation Research Part C: Emerging Technologies, 2021, 122: 102858.

[12] 陈垚, 柏赟, 张安英,等.基于滚动时域优化的共享自动驾驶汽车动态调度方法[J]. 交通运输系统工程与信息, 2022, 22(3): 45–52, 73. [CHEN Y, BAI Y, ZHANG A Y, et al. Dynamic fleet management of shared autonomous vehicles with rolling horizon optimization[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(3): 45–52, 73.]

[13] WANG Y, SUN H, LV Y, et al. Reinforcement learning-based order-dispatching optimization in the ride-sourcing service[J]. Computers & Industrial Engineering, 2024, 192: 110221.

[14] DING H, LI J, ZHENG N, et al. Dynamic dispatch of connected taxis for large-scale urban road networks with stochastic demands: An MFD-enabled hierarchical and cooperative approach[J]. Transportation Research Part C: Emerging Technologies, 2022, 142: 103792.

[15] ASHKROF P, GHASEMI F, KUCHARSKI R, et al. The implications of drivers’ ride acceptance decisions on the operations of ride-sourcing platforms[J]. Transportation Research Part A: Policy and Practice, 2025, 192: 104362.

[16] YAN P, YU K, CHAO X, et al. An online reinforcement learning approach to charging and order-dispatching optimization for an e-hailing electric vehicle fleet[J]. European Journal of Operational Research, 2023, 310(3): 1218–1233.

面向集中供给与稀疏需求场景的网约车调度优化算法

Optimization Algorithm for Ride-hailing Dispatch Under Concentrated Supply and Sparse Demand

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	韩艳, 袁昌银, 唐心恬, 关宏志. 考虑心理因素的自动驾驶车用户停车选择行为分析[J]. 交通运输系统工程与信息, 2025, 25(5): 72-82.
[2]	尹超英, 周悦, 徐震宇, 邵春福, 王晓全, 齐欣. 建成环境与主观感知对共享单车地铁换乘出行的非线性影响[J]. 交通运输系统工程与信息, 2025, 25(5): 169-178.
[3]	胡立伟, 龚麒, 赵雪亭, 周泽禹, 陈家乐, 潘江雄, 杨灿, 马思月. 多车联动式掉头区域交通冲突风险识别研究[J]. 交通运输系统工程与信息, 2025, 25(5): 179-192.
[4]	陈文强, 薛盼盼, 王涛, 顾玉磊. 无信号灯控制路段人行横道车辆让行行为研究[J]. 交通运输系统工程与信息, 2025, 25(5): 215-225.
[5]	孟冉, 杨洋, 宋子龙. 城市轨道交通分时定价策略的公益性水平评价[J]. 交通运输系统工程与信息, 2025, 25(5): 248-260.
[6]	唐春艳, 代馨怡, 张积昱. 适老化定制公交线路服务优化研究[J]. 交通运输系统工程与信息, 2025, 25(5): 271-279.
[7]	马飞虎, 陈晓燕, 孙翠羽, 田星彤. 高速公路收费站动态车道配置的在线优化方法[J]. 交通运输系统工程与信息, 2025, 25(5): 333-342.
[8]	翁剑成, 王亦昊, 陈旭蕊, 高润鸿, 胡松. 基于寻路实验的综合客运枢纽引导标识效用评估与诊断方法[J]. 交通运输系统工程与信息, 2025, 25(5): 343-352.
[9]	谭德明, 陈可沛, 吴大维, 李延欢, 胡四新, 张彩平. 低碳视角下轨道交通与土地利用供需协同发展研究[J]. 交通运输系统工程与信息, 2025, 25(4): 13-23.
[10]	张鹏羽, 李正中, 张翕然, 岳晓辉. 不同时期下城市轨道交通客流的时空影响机制研究[J]. 交通运输系统工程与信息, 2025, 25(4): 24-33.
[11]	俞诚成, 戴一楚, 杨超, 徐雷, 袁泉. 出行即服务促销策略对用户生命周期提升的因果推断研究[J]. 交通运输系统工程与信息, 2025, 25(4): 34-43.
[12]	代亮, 杜鹏飞, 黄自彬, 杨朋博. 基于深度强化学习的城市交通信号分层协同控制方法[J]. 交通运输系统工程与信息, 2025, 25(4): 63-72.
[13]	王海涌, 张丹, 王孟琳, 田爱爱. 异质交通流下交叉口信号及车辆轨迹融合控制模型[J]. 交通运输系统工程与信息, 2025, 25(4): 96-103.
[14]	王祥, 任浩, 谭国真, 李健平, 王珏, 王妍力. 大语言模型协同强化学习的自动驾驶决策方法[J]. 交通运输系统工程与信息, 2025, 25(4): 137-146.
[15]	宋翠颖, 丁杰, 张春波. 模块化公交车辆调度研究综述[J]. 交通运输系统工程与信息, 2025, 25(4): 175-192.