基于边端协同的多用户通信计算资源联合管理方法研究  

Joint Communication and Computation Resources Allocation for Edge-end Collaborative Multi-User Networks

作　　者：李嘉盛 姜宁[1] 贺家乐 郭凤仙 LI Jiasheng;JIANG Ning;HE Jiale;GUO Fengxian(State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]北京邮电大学网络与交换技术全国重点实验室,北京100876

出　　处：《移动通信》2025年第3期78-85,共8页Mobile Communications

基　　金：国家科技重大专项项目“5G-A通感融合组网关键技术研究与验证”(2024ZD1300100);国家自然科学基金项目“面向差异化服务的通信-感知-计算融合无线组网理论与方法”(62371067)。

摘　　要：移动边缘计算是针对通信和计算需求的下一代网络关键技术,通过联合边缘服务器的计算处理功能和无线网络的无线传输功能,可以显著减少计算密集型业务的系统延迟。然而随着毫米波技术的发展,无线系统中的波束变窄,且多用户间干扰严重,无线链路质量受限于波束赋形性能,从而导致基站与多用户间卸载数据传输效率低、算力资源管理与无线链路质量耦合等问题。为此,提出了一种基于边端协同的多用户通信计算资源联合管理方法以降低系统延迟,具体来说,利用基于交替方向乘子法优化框架和连续凸逼近迭代算法,联合优化多用户计算任务卸载比率、基站波束赋形矩阵和移动边缘计算服务器处算力资源分配,将耦合的优化问题拆分成更易处理的子问题求解,仿真结果表明该算法可以在特定的通信和计算预算下显著降低系统最大时延。Mobile edge computing(MEC)is a key technology for next-generation networks,designed to address both communication and computation demands.By combining the computation processing capability of MEC servers with the wireless transmission functions of the network,MEC can significantly reduce system latency for computation-intensive tasks.However,with the development of millimeter-wave technology,wireless beams become narrower,and severe interference among multiple users leads to degraded wireless link quality,limiting beamforming performance.This results in low data offloading efficiency between base stations and users,as well as challenges in managing computation resources and wireless link quality.To address these issues,we propose a joint management method for communication and computation resources based on edge-end collaboration to reduce system latency.Specifically,we utilize an optimization framework based on the alternating direction method of multipliers and a continuous convex approximation iterative algorithm to jointly optimize the user task offloading ratio,base station beamforming matrix,and computation resource allocation at MEC servers.By decomposing the coupled optimization problem into more manageable subproblems,the proposed method effectively reduces the maximum system latency under given communication and computation budgets,as demonstrated by simulation results.

关 键 词：计算卸载 波束赋形 边端协同 边缘计算 

分 类 号：TN929.5[电子电信—通信与信息系统]