动态车联网中的任务卸载策略

吉林大学学报(理学版) ›› 2026, Vol. 64 ›› Issue (5): 1162-1172.

动态车联网中的任务卸载策略

郑先锋^1,2, 王丽艳^1,2, 李汶蔚³, 冯浩⁴

1. 重庆移通学院大数据与计算机科学学院, 重庆 401520； 2. 公共大数据安全技术重庆市重点实验室, 重庆 401420; 3. 重庆邮电大学通信与信息工程学院, 重庆 400065； 4. 山西财经大学信息学院, 太原 030006

收稿日期:2025-09-26 出版日期:2026-09-26 发布日期:2026-09-26
通讯作者: 郑先锋 E-mail:cqtlzxf@163.com

Task Offloading Strategy in Dynamic Internet of Vehicles

Zheng Xianfeng^1,2, Wang Liyan^1,2, Li Wenwei³, Feng Hao⁴

1. School of Big Data and Computer Science, Chongqing College of Mobile Communication, Chongqing 401520, China; 2. Chongqing Key Laboratory of Public Big Data Security Technology, Chongqing 401420, China; 3. School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China; 4. School of Information, Shanxi University of Finance and Economics, Taiyuan 030006, China

Received:2025-09-26 Online:2026-09-26 Published:2026-09-26

摘要/Abstract

摘要： 为解决动态车联网（Internet of Vehicles, IoV）环境中的任务卸载决策问题, 首先, 提出一个由车辆、多个边缘服务器和云服务器组成的三层任务卸载网络架构; 其次, 提出以最小化任务卸载时延和能耗、提高服务质量为目标的任务卸载决策问题（task offloading decisions problem with the aim of minimizing latency, energy consumption, and enhancing service quality, TOD-LEQ）；再次, 将TOD-LEQ建模为Markov决策过程（Markov decision process, MDP）模型; 最后, 为应对动态变化环境对车联网中任务卸载的影响, 提出一种基于元学习的分布式强化学习任务卸载（meta learning-based distributed reinforcement learning task offloading, ME-DRO）算法, 该算法可在多线程中执行, 以寻求最优的任务卸载决策. 仿真实验结果表明, ME-DRO算法在时延、服务质量和收敛速度方面明显优于其他基线算法.

关键词: 车联网, 任务卸载, Markov决策过程, 元学习

Abstract: To address the task offloading decision problem in a dynamic Internet of Vehicles (IoV) network environment, a three-tier task offloading network architecture consisting of vehicles, multiple edge servers, and cloud servers is proposed. Secondly, the task offloading decisions problem with the aim of minimizing latency, energy consumption, and enhancing service quality (TOD-LEQ) in this network is defined. Thirdly, the TOD-LEQ is modeled as a Markov decision process (MDP) model. Finally, in order to cope with the impact of dynamically changing environments on task offloading in telematics, a meta learning-based distributed reinforcement learning task offloading (ME-DRO) algorithm is proposed, which can be executed in multiple threads to seek the optimal task offloading decision. Simulation results show that the ME-DRO algorithm significantly outperforms other baseline algorithms in terms of latency, quality of service and convergence speed.

Key words: Internet of Vehicles, task offloading, Markov decision process, meta learning

中图分类号:

郑先锋, 王丽艳, 李汶蔚, 冯浩. 动态车联网中的任务卸载策略[J]. 吉林大学学报(理学版), 2026, 64(5): 1162-1172.

Zheng Xianfeng, Wang Liyan, Li Wenwei, Feng Hao. Task Offloading Strategy in Dynamic Internet of Vehicles[J]. Journal of Jilin University Science Edition, 2026, 64(5): 1162-1172.

[1]	王子恒, 唐菁敏, 宋耀莲, 虞贵财. 车辆边缘计算中的资源分配与任务卸载比例优化[J]. 吉林大学学报(理学版), 2026, 64(2): 411-0420.
[2]	周冬杨, 胡红琼, 李汶蔚, 冯浩. 基于图神经网络的服务迁移策略[J]. 吉林大学学报(理学版), 2026, 64(1): 123-0133.
[3]	冷嵩, 李志强, 高金梁, 刘衍珩. 基于车联网的中心导航云性能分析模型[J]. 吉林大学学报(理学版), 2025, 63(4): 1150-1156.
[4]	董立岩, 齐竞则, 刘元宁, 冯嘉辉. 基于偏好和虚拟适应度的两阶段依赖任务卸载算法[J]. 吉林大学学报(理学版), 2024, 62(4): 923-932.
[5]	汪雨竹, 彭涛, 朱蓓蓓, 崔海. 基于元学习的小样本知识图谱补全[J]. 吉林大学学报(理学版), 2023, 61(3): 623-630.
[6]	万聪, 王英. 基于加权元学习的节点分类算法[J]. 吉林大学学报(理学版), 2023, 61(2): 331-337.
[7]	吴玲云, 秦贵和, 于赫. 基于随机森林的车载CAN总线异常检测方法[J]. 吉林大学学报(理学版), 2018, 56(3): 663-668.