南京师大学报(自然科学版)2024,Vol.47Issue(1) :121-132.DOI:10.3969/j.issn.1001-4616.2024.01.014

边端协同环境中的任务卸载和资源分配方法

Joint Task Offloading and Resource Allocation Method Based on Multi-Objective Optimization

张俊娜 赵豪 李天泽 赵晓焱 王亚丽
南京师大学报(自然科学版)2024,Vol.47Issue(1) :121-132.DOI:10.3969/j.issn.1001-4616.2024.01.014

边端协同环境中的任务卸载和资源分配方法

Joint Task Offloading and Resource Allocation Method Based on Multi-Objective Optimization

张俊娜 1赵豪 1李天泽 1赵晓焱 1王亚丽1
扫码查看

作者信息

  • 1. 河南师范大学计算机与信息工程学院,河南 新乡 453007
  • 折叠

摘要

将终端任务卸载至边缘计算环境弥补了云计算距离较远而产生较大延迟的缺陷,同时还降低了设备能耗.但从资源方面来讲,边缘服务器的各类资源并不像云服务器那么充足,因此,任务卸载和资源分配的联合优化成为边缘计算的研究热点之一.已有的任务卸载和资源分配联合优化研究通常假设任务卸载至单个边缘服务器,默认每个终端设备产生一个任务,即使有研究多服务器的,也通常忽略服务器间的负载均衡.为此,本文在一个多边缘服务器多用户多任务的边端系统中,提出了一种权衡时延、能耗和负载均衡指标(即效益)的任务卸载和资源分配方法,其通过优化任务卸载决策、服务器计算资源分配和终端设备发射功率,实现任务卸载效益最大化.最后,为了验证所提方法的有效性,进行了充分的对比实验.实验结果表明,与对比方法相比,所提出的方法在提升卸载效益和实现服务器间负载均衡方面有良好的性能.

Abstract

Offloading terminal tasks to the edge computing environment makes up for the large delay caused by the distance of cloud computing,and reduces the power consumption of equipment.But in terms of resources,all kinds of resources of edge server are not as sufficient as that of cloud server.Therefore,the joint optimization of task offloading and resource allocation becomes one of the research hotspots of edge computing.Existing studies on joint optimization of task offloading and resource allocation generally assume that tasks are offloading to a single edge server;By default,each terminal device generates one task.Even when multiple servers are considered,load balancing between servers is often ignored.Therefore,in a multi-edge server,multi-user and multi-task edge system,this paper proposes a task offloading and resource allocation method that balances delay,energy consumption and load balancing index(i.e.,benefit).By optimizing task offloading decision,server computing resource allocation and terminal device transmission power,the benefit of task offloading can be maximized.In order to verify the effectiveness of the proposed method,a full comparison experiment is carried out.The experimental results show that,compared with the comparison method,the proposed method has good performance in improving the unloading efficiency and realizing the load balancing between servers.

关键词

边缘计算/任务卸载/资源分配/负载均衡/强化学习

Key words

edge computing/task offloading/resource allocation/load balancing/reinforcement learning

引用本文复制引用

基金项目

国家自然科学基金项目(62072159)

河南省科技攻关项目(232102211061)

河南省科技攻关项目(222102210011)

出版年

2024
南京师大学报(自然科学版)
南京师范大学

南京师大学报(自然科学版)

CSTPCD北大核心
影响因子:0.427
ISSN:1001-4616
参考文献量21
段落导航相关论文