摘要
针对车联网数据卸载策略在选择边缘服务器时忽略负载均衡的问题,提出一种基于移动云服务的车联网任务卸载策略。该策略基于一种新型网络架构,采用强化学习实现卸载任务指派,将任务卸载的问题转化为车联网服务收益的问题,以通信资源和计算资源构建约束条件,结合整个计算资源系统任务处理时延最低和系统可靠性的要求,寻找可用的服务器节点,实现对卸载任务的最优指派。实验仿真表明,所提出的方法能够减少系统的负载率和任务完成时间,降低了最大链路带宽占用率,从而提升了任务卸载的效率。
Aiming at the problem of ignores load balancing when selecting edge servers for data offloading strategy in Internet of vehicles(IoV), an IoV task offloading strategy is proposed based on mobile cloud services. The proposed strategy adopts reinforcement learning to realize the offloading task assignment based on a new network architecture, and the problem of task offloading is transformed into the revenue issue of Io V services. Specifically, the communication and computing resources constructs as the constraint conditions with the combination of the requirements of the lowest task processing delay and system reliability in the whole computing resource system, and available server nodes are found to realize the optimal assignment of offloading tasks. Experimental simulation shows that the proposed method can reduce the load rate and task completion time of the system, reduce the occupancy rate of the maximum link bandwidth, and thus improve the efficiency of task offloading.
作者
彭科
黄焘
程旭
李强
李陶然
PENG Ke;HUANG Tao;CHENG Xun;LI Qiang;LI Taoran(China Southern Power Grid Electric Vehicle Service Co.,Ltd.,Shenzhen 518000,China)
出处
《移动通信》
2022年第8期113-119,共7页
Mobile Communications
关键词
车联网
移动云服务
任务卸载
增强学习
Internet of vehicles
mobile cloud services
task offloading
reinforcement learning
