摘要
移动边缘云以分布式计算方式将计算能力以及存储资源下沉至网络边缘,而边缘云网络覆盖的局限性以及用户移动的不确定性使得用户对实时服务提出较高的要求。边缘云服务迁移可以为移动中的用户提供连续服务,而迁移系统的普适性和迁移决策的高效性成为迁移过程的一大挑战。因此论文提出了一种无模型的深度强化学习算法来解决迁移决策问题。实验结果表明,该方法可建立通用的系统架构且在服务时延、迁移开销以及两者之间的折衷优化方面优于现有的方法。
Mobile edge cloud sinks computing power and storage resources to the edge of the network in a distributed comput-ing way.However,the limitation of the coverage of the edge cloud network and the uncertainty of users'mobility make users put for-ward higher requirements for real-time services.Edge cloud service migration can provide continuous services for users on the move,but the universality of migration system and the efficiency of migration decision become a major challenge in the migration process.The experimental results show that this approach establishes a common system architecture and is superior to existing ap-proaches in terms of user service latency,migration overhead,and trade-off optimization between them.
作者
李苗
LI Miao(College of Computer Science and Technology,China University of Petroleum(East China),Qingdao 266580)
出处
《计算机与数字工程》
2023年第6期1333-1337,共5页
Computer & Digital Engineering
关键词
移动边缘云
用户移动
服务迁移
深度强化学习
优化目标
mobile edge cloud
mobile user
service migration
deep reinforcement learning
optimization goal
