面向边缘智能的移动算力网络关键技术研究  

Research on Key Technologies of Mobile Computing Force Network for Edge Intelligence

作　　者：郑远鹏 张天魁[1] 庞博 陈杲 Zheng Yuanpeng;Zhang Tiankui;Pang Bo;Chen Gao(Beijing University of Posts and Telecommunications,Beijing 100876,China;China Unicom Research Institute,Beijing 100191,China)

机构地区：[1]北京邮电大学,北京100876 [2]中国联通研究院,北京100191

出　　处：《邮电设计技术》2023年第5期88-92,共5页Designing Techniques of Posts and Telecommunications

摘　　要：随着人工智能计算与应用的不断发展,基于深度学习的智能服务对算力资源需求也逐渐提高,因此利用移动算力网络为边缘智能提供泛在异构、高效可靠的算力服务支撑。针对边缘智能和移动算力网络的研究现状和关键技术进行了讨论,提出了面向边缘智能的移动算力网络设计实例,给出了面向智能任务的网络部署方案和多维资源分配方案。With the continuous development of artificial intelligence computing and applications,the demand for computing resources in intelligent services based on deep learning has gradually increased.Therefore,using mobile computing force networks to provide ubiquitous,heterogeneous,efficient and reliable computing force services for edge intelligence is essential.It discusses the research status and key technologies of edge intelligence and mobile computing force network,proposes a design example of a mobile computing network for edge intelligence,and provides a deployment scheme for intelligent tasks and a multidimensional resource allocation scheme.

关 键 词：边缘智能 移动算力网络 结合部署 资源分配 

分 类 号：TN915[电子电信—通信与信息系统]