基于日志信息的光通信网络流量预测研究  

Research on traffic prediction of optical communication network based on log information

作　　者：尹益燕 黄振华 刘翠焕[1] YIN Yiyan;HUANG Zhenhua;LIU Cuihuan(Hebei engineering and Technology Institute,Shijiazhuang 050091,China)

机构地区：[1]河北工程技术学院,石家庄050091

出　　处：《激光杂志》2023年第1期164-168,共5页Laser Journal

基　　金：河北省教育厅基金项目(No.Z2020229)。

摘　　要：为了明确当前光通信网络流量预测过程中存在的一些问题,提出了基于日志信息的光通信网络流量预测方法。首先获取光通信网络服务的日志信息获取网络流量历史数据,并光通信网络流量数据进行预处理,建立光通信网络流量预测训练样本,然后引入灰色模型和神经网络建立光通信网络流量组合预测模型,以实现未来一段时间的光通信网络流量预测,最后的光通信网络流量预测实例分析结果表明,本方法根据日志信息可以挖掘光通信网络流量变化趋势,不但光通信网络流量预测精度超过92%,可以为光通信系统管理者提供有用的参考信息。In order to clarify some problems in the current optical communication network traffic prediction,a traffic prediction method based on log information is proposed.Firstly,the log information of optical communication network service is obtained,the network traffic history data is obtained,the optical communication network traffic data is preprocessed,and the optical communication network traffic prediction training samples are established.Then,the grey model and neural network are introduced to establish the optical communication network traffic combination prediction model,so as to realize the optical communication network traffic prediction in the future.Finally,the analysis results of optical communication network traffic prediction example show that this method can mine the change trend of optical communication network traffic according to the log information.Not only the accuracy of optical communication network traffic prediction is more than 92%,but also can provide useful reference information for optical communication system managers.

关 键 词：光通信系统 日志信息 网络流量 遗传算法 训练样本构建 

分 类 号：TN929[电子电信—通信与信息系统]