面向直播的边缘计算任务卸载方案研究  

Research on Edge Computing Task Offloading Scheme for Live Broadcasting

作　　者：谢松 王薇 XIE Song;WANG Wei(College of Network Security,Changchun University,Changchun 130022,China)

机构地区：[1]长春大学网络安全学院,吉林长春130022

出　　处：《软件工程》2024年第1期27-31,共5页Software Engineering

基　　金：2024年度吉林省教育厅科学研究重点课题项目(JJKH20240749KJ);中国职业技术教育学会2023年度重点课题立项(ZJ2023A022)。

摘　　要：文章研究分析了直播中多用户、多服务器场景下存在的直播用户的体验质量(Quality of Experience,QoE)不高的问题,为提升用户的QoE,将能耗和时延作为决策目标,设计一种经改进的NSGA-Ⅱ(非支配排序遗传算法),即L-NSGA-Ⅱ,用线性排名的方式进行父代的选择加速算法收敛。实验表明,与LUA、NSGA-Ⅱ和Random算法策略相比,所提方案的平均延迟降低约9.1%,用户QoE提升约4.39%。该方案应用于直播场景中,在减少延迟、提升吞吐量和降低能源开销方面表现出较好的效果。This paper studies and analyzes the problem of low Quality of Experience(QoE)for live streaming users in multi-user and multi-server scenarios.To improve the QoE of users,this paper proposes to design an improved NSGA-Ⅱ(Non-Dominant Sorting Genetic Algorithm),namely L-NSGA-Ⅱ,taking energy consumption and delay as decision-making objectives.The algorithm uses linear ranking to accelerate the convergence of the parent selection algorithm.The experiment shows that compared with the LUA,NSGA-Ⅱ,and Random algorithm strategies,the proposed scheme reduces the average latency by about 9.1%and improves user QoE by about 4.39%.This scheme has shown good performance in reducing latency,improving throughput,and reducing energy consumption when applied to live streaming scenarios.

关 键 词：直播 边缘计算 卸载 用户体验 用户分配 L-NSGA-Ⅱ 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]