Deep reinforcement learning-based optimization for task offloading and resource management in vehicular network

doi:10.3785/j.issn.1008-973X.2026.12.012

Journal of ZheJiang University (Engineering Science)

2026, Vol. 60

Issue (12): 2663-2672 DOI: 10.3785/j.issn.1008-973X.2026.12.012

Deep reinforcement learning-based optimization for task offloading and resource management in vehicular network

Xuxiu ZHANG(

),Lixuan YIN,Jian XIE,Yi DING,Han YAN,Lin WANG*(

)

School of Electrical Engineering, Dalian Jiaotong University, Dalian 116028, China

Download:

HTML

PDF(1877KB) HTML
Export: BibTeX | EndNote (RIS)

Abstract

A deep reinforcement learning (DRL)-based method was proposed for task offloading and resource management optimization in vehicular network in order to enhance real-time performance and computational efficiency in autonomous driving task processing. A mobile edge computing (MEC) system was targeted. Roadside units (RSUs) were configured as computing and caching nodes in order to achieve dynamic task offloading and content caching optimization. Vehicle task demand variations were predicted by an improved Hawkes model, and corresponding caching strategies were generated. The task offloading and resource allocation problem was formulated as a 0-1 optimization model and converted into a Markov decision process (MDP). An improved proximal policy optimization algorithm was applied to obtain the solution. The simulation results showed that average computation delay was significantly reduced and request response rate was increased in highly dynamic vehicular environments. A 15.1% improvement in total reward was achieved compared with the baseline. The effectiveness of the method in optimizing real-time computation and caching management of vehicular network in complex traffic scenarios was validated.

Key words： vehicular network mobile edge computing (MEC) deep reinforcement learning (DRL) task offloading resource management roadside unit (RSU)

Received: 27 September 2025 Published: 21 October 2026

CLC:

TP 181

Fund: 国家自然科学基金资助项目（U24B20159）；辽宁省重点研发计划资助项目（2022020594-JH1/108）.

Corresponding Authors: Lin WANG E-mail: zhangxuxiu@163.com;wanglin_jl@163.com

	Service
	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Xuxiu ZHANG
	Lixuan YIN
	Jian XIE
	Yi DING
	Han YAN
	Lin WANG

Cite this article:

Xuxiu ZHANG,Lixuan YIN,Jian XIE,Yi DING,Han YAN,Lin WANG. Deep reinforcement learning-based optimization for task offloading and resource management in vehicular network. Journal of ZheJiang University (Engineering Science), 2026, 60(12): 2663-2672.

URL:

https://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2026.12.012 OR https://www.zjujournals.com/eng/Y2026/V60/I12/2663

基于深度强化学习的车联网任务卸载与资源管理优化

为了提升自动驾驶中任务处理的实时性与计算效率，提出基于深度强化学习（DRL）的车联网任务卸载与资源管理优化方法. 研究以移动边缘计算（MEC）系统为对象，将路侧单元（RSU）配置为计算与缓存节点，实现车辆任务的动态卸载与内容缓存优化. 通过改进的霍克斯模型预测车辆任务需求的变化，据此生成缓存配置策略，将任务卸载与资源分配问题建模为0-1优化模型并转化为马尔可夫决策过程（MDP），采用改进的近端策略优化算法进行求解. 仿真结果表明，该算法在高动态车辆环境下可以显著降低平均计算延迟，提升请求响应率，且相较于未改进的算法，获得的总奖励提升15.1%，验证其在复杂交通场景下优化车载网络实时计算与缓存管理的有效性.

关键词： 车联网, 移动边缘计算（MEC）, 深度强化学习（DRL）, 任务卸载, 资源管理, 路侧单元（RSU）

Fig.1 Mobile edge computing system for roadside unit-assisted vehicular network

Fig.2 Structural and optimization flowchart of actor network and critic network

Tab.1 Simulation parameter configuration of vehicular MEC system

Tab.2 Performance comparison of EUPPO algorithm modules

Fig.3 Impact of different weight factor

$ \tau $

Fig.4 Impact of different environment parameters on response rate

Fig.5 Comparison of cache hit rate under different cache capacities

Fig.6 Comparison experiment of different learning rates and rewards　　　　

Fig.7 Comparison of average rewards across training episodes for different algorithms

Fig.8 Comparison of delay and vehicle cache hit counts

Fig.9 Reward comparison of different optimal algorithms


[1]	Marwein P S, Sur S N, Gao Xiaozhi, et al Recent survey on Internet of vehicles: architecture, applications, challenges, and its solutions[J]. Journal of Testing and Evaluation, 2024, 52 (1): 731- 753 doi: 10.1520/JTE20230095

[2]	Song Tao, Li Xiuhua, Yu Yingheng, et al Low-power secure caching strategy for internet of vehicles[J]. Wireless Networks, 2024, 30 (6): 5289- 5303 doi: 10.1007/s11276-023-03271-2

[3]	Zhu Yan Collaborative intelligent traffic planning in the internet of vehicles[J]. Sensors, 2024, 24 (4): 1303 doi: 10.3390/s24041303

[4]	Liu Lei, Lei Xiaoming, Wang Qiang A multi-objective computation offloading algorithm in MEC environments[J]. International Journal of Computational Science and Engineering, 2022, 25 (3): 298- 307

[5]	Yang Yang, Yang Qiang Efficient task offloading and resource allocation in MEC[J]. Journal of Cases on Information Technology, 2025, 27 (1): 1- 22 doi: 10.4018/jcit.371753

[6]	Al-Essa R I, Al-Suhail G A. A service of RSU communication in Internet of vehicles (IoV) in urban environment [C]//Computational Intelligence, Data Analytics and Applications. Cham: Springer, 2023: 291–303.

[7]	An H Y, Choi W S, Choi S G Performance analysis of RSU relaying in the vehicular network[J]. The Journal of Korean Institute of Communications and Information Sciences, 2022, 47 (3): 522- 529 doi: 10.7840/kics.2022.47.3.522

[8]	Kang Zhidong, Li Dan, Guo Wei, et al. Multi-objective optimization technique for RSU deployment [C]//Data Mining and Big Data. Singapore: Springer, 2022: 118–132.

[9]	Yang Tengfei, Yang Jian Deep reinforcement learning method of offloading decision and resource allocation in MEC[J]. Computer Engineering, 2021, 47 (8): 37- 44

[10]	Li Ming, Mao Ning, Zheng Xiao, et al. Computation offloading in edge computing based on deep reinforcement learning [C]//Proceedings of International Conference on Computing and Communication Networks. Singapore: Springer, 2022: 339–353.

[11]	Ren Dan, Gui Xiaolin, Zhang Kang Adaptive request scheduling and service caching for MEC-assisted IoT networks: an online learning approach[J]. IEEE Internet of Things Journal, 2022, 9 (18): 17372- 17386 doi: 10.1109/JIOT.2022.3157677

[12]	Mohammed L B, Anpalagan A, Jaseemuddin M Energy and latency efficient caching in mobile edge networks: survey, solutions, and challenges[J]. Wireless Personal Communications, 2023, 129 (2): 1249- 1283 doi: 10.1007/s11277-023-10187-9

[13]	Li Chang, Song Mei, Du Shoushang, et al Adaptive priority-based cache replacement and prediction-based cache prefetching in edge computing environment[J]. Journal of Network and Computer Applications, 2020, 165: 102715 doi: 10.1016/j.jnca.2020.102715

[14]	Moradi A, Rezaei F Intelligent and efficient task caching for mobile edge computing[J]. Cluster Computing, 2024, 27 (10): 14095- 14112 doi: 10.1007/s10586-024-04658-2

[15]	Zhao Wei, Cheng Yu, Liu Zhi, et al A trust-based malicious RSU detection mechanism in edge-enabled vehicular Ad Hoc networks[J]. Journal of the Asiatic Society of Bangladesh, Science, 2025, 39 (2): 242- 249 doi: 10.3329/jasbs.v51i1.82796

[16]	Zhu Xin, Zhang Bo, Chen Zhi, et al Notice of violation: a 3D printed multi-mode conical dielectric resonant antenna with stable omnidirectional radiation patterns for 5G vehicle picocell applications[J]. IEEE Transactions on Vehicular Technology, 2024, 73 (1): 1- 10 doi: 10.1109/tvt.2022.3196909

[17]	Luo Quyuan, Li Changle, Luan T H, et al Collaborative data scheduling for vehicular edge computing via deep reinforcement learning[J]. IEEE Internet of Things Journal, 2020, 7 (10): 9637- 9650 doi: 10.1109/JIOT.2020.2983660

[18]	Zhao Liang, Zhang Engang, Wan Shaohua, et al MESON: a mobility-aware dependent task offloading scheme for urban vehicular edge computing[J]. IEEE Transactions on Mobile Computing, 2024, 23 (5): 4259- 4272 doi: 10.1109/TMC.2023.3289611

[19]	Li Zhenyu, Zhang Zhixuan Deep reinforcement learning-based joint optimization model for vehicular task offloading and resource allocation[J]. Peer-to-Peer Networking and Applications, 2024, 17 (4): 2001- 2015 doi: 10.1007/s12083-024-01693-z

[20]	Abuthahir S S, Selvin Paul Peter J Multi-dimensional task offloading using deep learning for vehicular edge computing networks: a survey[J]. IEIE Transactions on Smart Processing and Computing, 2024, 13 (1): 1- 11 doi: 10.5573/ieiespc.2024.13.1.1

[1]	Shuxu ZHAO,Qi ZHU,Xiaolong WANG. Reliability-enhanced task deployment method in mobile edge computing environments[J]. Journal of ZheJiang University (Engineering Science), 2026, 60(8): 1697-1708.

[2]	Qingqing YANG,Runpeng TANG,Yi PENG. Joint waveform and phase shift design in integrated sensing and communication systems[J]. Journal of ZheJiang University (Engineering Science), 2026, 60(4): 906-914.

[3]	Xingxing LI,Fan YANG,Jie HUANG,Xianzhi LAI,Fenghang YAO,Jieliang CAI,Ni ZHANG. Federated reinforcement learning for resource management in vehicular networks with differentiated interferences[J]. Journal of ZheJiang University (Engineering Science), 2026, 60(10): 2207-2214.

[4]	Xianzhi LAI,Fan YANG,Jie HUANG,Chengbo YU,Xingxing LI. Heterogeneous hierarchical task offloading and resource optimization for Smart Internet of Vehicles[J]. Journal of ZheJiang University (Engineering Science), 2026, 60(10): 2236-2246.

[5]	Zhong CHEN,Xiao XU,Hai-wei WANG,Hong-hao LUO,Xuan CHEN. Optimization strategy for unloading power tasks in residential areas based on alternate edge nodes[J]. Journal of ZheJiang University (Engineering Science), 2021, 55(5): 917-926.

[6]	Xing-cheng HUA,Peng LIU. Near-data processing based on dynamic task offloading[J]. Journal of ZheJiang University (Engineering Science), 2019, 53(12): 2348-2356.

Viewed

Full text

Abstract

Cited

Shared

Discussed