移动边缘网络下服务缓存与资源分配联合优化策略  

Joint optimization strategy of service cache and resource allocation in mobile edge network

作　　者：龙隆[1] 刘子辰[1] 陆在旺 张玉成[1,2] 李蕾 LONG Long;LIU Zichen;LU Zaiwang;ZHANG Yucheng;LI Lei(Institute of Computing Technology,Chinese Academy of Sciences,Beijing 100190,China;Institute of Computing Technology,University of Chinese Academy of Sciences,Beijing 100190,China)

机构地区：[1]中国科学院计算技术研究所,北京100190 [2]中国科学院大学计算技术研究所,北京100049

出　　处：《通信学报》2023年第1期64-74,共11页Journal on Communications

基　　金：国家重点研发计划基金资助项目(No.2022YFD2001004)。

摘　　要：针对边缘服务器的计算与存储资源有限,过多的任务卸载将使其计算能力与负载能力不匹配,从而导致任务处理时延增加的问题,研究了基于多用户、边缘服务器以及云服务器组成的三层网络架构下的任务卸载与存储资源联合优化问题以降低系统整体时延。由于该问题为混合整数非线性问题,因此提出了一种服务缓存与资源分配联合优化策略。首先,将原问题的连续与离散变量进行解耦为2个子问题,即服务缓存决策问题以及计算资源与通信资源联合优化问题。然后,通过重构线性化方法、松弛法以及凸优化方法对2个子问题进行交替优化迭代获得近似最优解。仿真结果表明,所提策略在低复杂度情况下能够获得近似最优解,并且与其他策略相比时延降低10%左右。Aiming at the problem that computing and storage resources of edge node in a mobile edge computing(MEC)system were limited, and excessive task offloading would cause the mismatch between the computing capacity and the load capacity of the edge server, resulting in the increase of task processing delay, the joint optimization of task offloading and storage resources under the three-layer network architecture composed of multi-user, edge server and cloud server was studied to reduce the overall system delay. For the problem was a mixed integer nonlinear problem, a joint optimization strategy of service caching and resource allocation was proposed. First, the continuous and discrete variables of the original problem were decoupled into two sub-problems, namely, the service cache decision problem and the joint optimization problem of computing resources and communication resources. Then, the linear reconstruction, relaxation method and convex optimization method were used to alternately optimize the two sub-problems to obtain the near-optimal solution. Simulation results demonstrate that the proposed strategy can obtain a near-optimal performance with low complexity, and can reduce up to 10% of the task duration compared with other strategies.

关 键 词：移动边缘计算 多级网络 服务缓存 资源分配 

分 类 号：TN929.52[电子电信—通信与信息系统]