移动边缘计算中的智能任务迁移技术:现状与展望

Intelligent Task Migration Technology in Mobile Edge Computing:Status and Prospects

与传统边缘计算相比,移动边缘计算(Mobile Edge Computing,MEC)技术能更好地解决移动设备资源受限问题,因而备受关注。而移动场景的高动态性,又对MEC的任务调度能力提出了挑战。为了应对这种挑战,将人工智能技术与MEC技术相结合已经成为一种新的发展趋势。首先,介绍了MEC技术的发展背景,然后详细说明MEC中的任务迁移技术和研究现状,最后展望了MEC和人工智能技术融合的发展方向,并对结合了深度强化学习技术的MEC技术进行仿真。结果表明,结合了深度强化学习的MEC系统在服务时延以及系统容量性能上都有着较大提升。

Compared with traditional edge computing,Mobile Edge Computing(MEC)technology can better solve the problem of resource constraints of mobile devices and has attracted much attention.The high dynamics of mobile scenarios also poses many challenge to the task scheduling capability of MEC.To meet this challenge,it has become a new trend to combine artificial intelligence technology with MEC technology.Firstly,the development background of MEC technology is introduced,and then the task migration techno-logy and research status in MEC are explained in detail.Finally,the development direction of the integration of MEC and artificial intelligence technology is prospected,and the MEC technology combined with deep reinforcement learning technology is simulated.The results show that the MEC system combined with deep reinforcement learning can greatly improve the service delay and system capacity performance.