面向F5G的光接入网智能规划应用研究  

Research on Intelligent Network Planning of Optical Access Network for F5G

作　　者：李奥 韩赛 王光全[1] 邵岩[1] 王泽林[1] Li Ao;Han Sai;Wang Guangquan;Shao Yan;Wang Zelin(China Unicom Research Institute,Beijing 100048,China)

机构地区：[1]中国联通研究院,北京100048

出　　处：《邮电设计技术》2023年第2期70-73,共4页Designing Techniques of Posts and Telecommunications

摘　　要：针对在保障用户带宽需求的前提下,如何提前评估扩容升级并主动规划光接入网络的问题,提出了一种基于树模型的接入网扩容规划算法,即通过决策树ID3算法定量分析影响PON口峰值速率的关键因素,根据决策树的结果使用回归树CART算法预测PON口峰值速率区间。实验表明,基于该方案构建的2个模型准确率均达80%以上,可为F5G时代光接入网精准评估并规划扩容提供借鉴。Aiming at the problem of how to evaluate the capacity expansion and upgrade in advance and actively plan the optical access network on the premise of ensuring the bandwidth demand of users,an optical access network capacity expansion planning algorithm based on the tree model is proposed.The key factors affecting the peak rate of PON ports are quantitatively analyzed through the decision tree algorithm(ID3).The peak rate range of the PON ports is predicted by the regression tree algorithm(CART)according to the result of a decision tree.Experiments show that the accuracy of the two models based on the scheme is more than 80%,which can provide a reference for the accurate evaluation and planning expansion of optical access network in the F5G era.

关 键 词：人工智能 F5G网络 光接入网 扩容评估 

分 类 号：TN915.6[电子电信—通信与信息系统]