摘要
车辆边缘计算(VEC)为处理计算密集、延迟敏感型任务提供了新的范式,然而边缘服务器在整合可再生能源方面的能力较差。因此,为了提高边缘服务器的能效,该文设计了一种面向绿色计算的车辆协同任务卸载框架。在该框架中,车辆配备能源收集(EH)设备,通过彼此间共享绿色能源和计算资源协作执行任务。为有效促进车辆的参与积极性,该文通过动态定价激励车辆,并综合考虑了车辆的移动性、任务优先级等。为了使卸载决策适应动态环境的变化,该文提出了一种基于双延迟深度确定性策略梯度(TD3)的任务卸载方法,以在最大化所有车辆平均任务完成效用的同时减少边缘端电网电力的使用。最后,仿真结果验证了该方法的有效性,相比基于深度确定性策略梯度(DDPG)和基于贪心原则(GPE)的方法在性能上分别提升了7.34%和37.47%。
Vehicular Edge Computing(VEC)has become a promising and prospective paradigm for computation-intensive and delay-sensitive tasks.However,edge servers are less capable of integrating renewable energy.Therefore,in order to improve the energy efficiency of edge servers,a green computing oriented vehicle collaborative task offloading framework is proposed.In this framework,vehicles equipped with Energy Harvest(EH)devices cooperate to perform tasks by sharing green energy and computing resources with each other.To effectively enhance the participation enthusiasm of vehicles,dynamic pricing is adopted to motivate vehicles,and the mobility and task priority are also considered comprehensively.In order to adapt the offloading decisions to the dynamic environment,a Twin Delayed Deep Deterministic policy gradient(TD3)based task offloading method is proposed to maximize the average task completion utility of all vehicles while reducing the use of grid power.Finally,simulation results verify the effectiveness of the proposed method,and the performance achieves 7.34%and 37.47%improvement respectively compared with Deep Deterministic Policy Gradient(DDPG)based method and Greedy Principle Execution(GPE)method.
作者
张红霞
吕智豪
席诗语
刘佳敏
郭加树
张培颖
ZHANG Hongxia;LÜZhihao;XI Shiyu;LIU Jiamin;GUO Jiashu;ZHANG Peiying(Qingdao Institute of Software,China University of Petroleum(East China),Qingdao 266000,China;College of Computer Science and Technology,China University of Petroleum(East China),Qingdao 266000,China)
出处
《电子与信息学报》
EI
CAS
CSCD
北大核心
2024年第1期175-183,共9页
Journal of Electronics & Information Technology
基金
山东省自然科学基金(ZR2020MF006,ZR2022LZH015)。
关键词
车辆边缘计算
任务卸载
能源收集
车辆协同
动态定价
Vehicular Edge Computing(VEC)
Task offloading
Energy Harvest(EH)
Vehicle collaboration
Dynamic pricing
