摘要
随着互联网技术的不断发展,AR,VR和自动驾驶等技术日渐成熟,丰富的应用为网络带来了爆炸式的流量增长,同时大量异构终端设备被接入到网络中,网络复杂度也在急剧上升。简单的路由策略已经无法适应复杂的网络环境和差异化的业务需求,学术界对于动态复杂网络的路由问题研究一直保持着高度关注。综合路由策略研究中各类算法的最新进展,从传统动态路由算法与智能路由算法等角度阐述了近年来各领域中路由算法的研究工作,对不同路由方法的应用场景及优劣进行了分析。结合人工智能领域的相关研究与发展方向,讨论了智能路由算法在未来研究中的机遇与挑战,并给出了智能路由算法未来的研究方向。
With the continuous development of the Internet,technologies such as AR,VR and autonomous driving are becoming more and more mature.Abundant applications have brought explosive traffic growth to the network.At the same time,as plenty of heterogeneous terminal devices are connected to the Internet,the network complexity is also increasing significantly.Simple routing strategies can no longer be applicable to the complex network environment and meet differentiated business requirements.The routing problems of dynamic networks have attracted the great interest of the research community.By synthesizing the latest progress of various routing algorithms,the advantages and disadvantages of application scenarios and various algorithms are analyzed from the perspectives of traditional and intelligent routing algorithms.The opportunities and challenges of intelligent routing algorithms are discussed combined with related research and development directions of artificial intelligence,and the future research directions are given.
作者
王志浩
郗海龙
许萌签
刘晓东
潘宁
肖哲
WANG Zhihao;XI Hailong;XU Mengqian;LIU Xiaodong;PAN Ning;XIAO Zhe(The 54th Research Institute of CETC,Shijiazhuang 050081,China;Navy North Sea Fleet General Staff,Qingdao 266000,China)
出处
《无线电工程》
北大核心
2022年第10期1755-1764,共10页
Radio Engineering
基金
通信网信息传输与分发技术重点实验室基金项目(SXX21641X033)
中国电科集团科技创新基金资助项目(SCX21641X002)。
关键词
智能路由
深度学习
启发式算法
有监督学习
强化学习
intelligent routing
deep learning
heuristic algorithm
supervised learning
reinforcement learning
