延迟敏感约束下的多址边缘网络中资源联合分配  

Joint Resource Allocation Strategy in Multiple Access Edge Networks Under Delay Sensitive Constraints

作　　者：葛志辉[1] 刘子萌 王哲[2] 李陶深[1] GE Zhi-hui;LIU Zi-meng;WANG Zhe;LI Tao-shen(School of Computer and Electronic Information,Guangxi University,Nanning 530004,China;School of Artificial Intelligent,Guangxi University for Nationalities,Nanning 530005,China)

机构地区：[1]广西大学计算机与电子信息学院,南宁530004 [2]广西民族大学人工智能学院,南宁530005

出　　处：《小型微型计算机系统》2023年第6期1283-1289,共7页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61762010)资助;广西科技计划项目(桂科AD20297125)资助.

摘　　要：在移动边缘计算(Mobile Edge Computing,MEC)中,传统的基于平均性能约束的多址接入和计算卸载方法,无法满足用户设备的时延敏感型业务需求.本文针对多接入点(Access Point,AP)和多移动边缘主机(Mobile Edge Host,MEH)的边缘网络,基于李雅普诺夫随机优化与匹配理论,在用户设备随机到达和计算任务随机生成的场景下,提出一种边缘网络资源联合优化分配策略,包括用户与AP-MEH接入策略和计算优化卸载策略,旨在最小化平均时延约束和失效概率约束下的用户平均功耗.仿真结果表明,所提出的边缘网络联合优化分配策略在降低用户设备功耗中是有效的,且较传统的队列稳定性约束下的资源分配策略在能耗优化方面实现了更高的系统增益.In mobile edge computing(MEC),the traditional multiple access and computing offload methods based on average performance constraints can not meet the delay sensitive service requirements of user equipment.Based on Lyapunov stochastic optimization and matching theory,this paper proposes a joint optimal allocation strategy of edge network resources in the scenario of random arrival of user equipment and random generation of computing tasks,including users and AP-MEH access strategy and computing optimization unloading strategy,Under the multiple access point(AP)and multiple Mobile Edge Host(MEH)edge network scenario.It aims to minimize the average power consumption of users under the constraints of average delay and out-of-service probability.Simulation results show that the proposed edge network joint optimization allocation strategy is effective in reducing user equipment power consumption,and achieves higher system gain in energy consumption optimization than the traditional resource allocation strategy under queue stability constraints.

关 键 词：多址接入边缘计算 随机优化 资源分配 平均队列 失效概率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]