计算机科学
Lyapunov优化
服务器
隐藏物
移动边缘计算
强化学习
分布式计算
最优化问题
水准点(测量)
边缘计算
任务(项目管理)
计算机网络
GSM演进的增强数据速率
人工智能
算法
李雅普诺夫方程
经济
管理
李雅普诺夫指数
地理
混乱的
大地测量学
作者
Nianxin Li,Xiumin Zhu,Yumei Li,Lingling Wang,Linbo Zhai
标识
DOI:10.1109/ispa-bdcloud-socialcom-sustaincom57177.2022.00023
摘要
With the development of Internet of Things (IoT) networks and Mobile Edge Computing (MEC), many computing-intensive applications have been developed in large quantities. Due to the heterogeneity of tasks, different application services are required to perform each task. Caching application services and related data in edge servers is challenging. Hence, we study the service cache placement and task offloading problem in IoT networks. Since IoT devices and edge servers with limited storage resources can only cache a few services at the same time, we formulate the service cache placement and task offloading of IoT devices problem to minimize task service delay with long-term energy constraint of IoT devices, which is a mixed integer nonlinear programming problem. To solve this problem, an online Deep Reinforcement Learning guided by the Lyapunov optimization framework algorithm (LYADRL) is proposed. We first build a virtual queue model to decouple the problem by Lyapunov optimization technique to transform the problem into a single time slot optimization problem. Then, we use Deep Reinforcement Learning techniques to find the optimal edge service caching and task offloading policies for each time slot. Simulation results show that our algorithm can reduce the service delay compared with other benchmark algorithms.
科研通智能强力驱动
Strongly Powered by AbleSci AI