摘要
海量的光通信网络传输数据导致光通信信道负载增加,数据传输负载预测难度加大,提出基于LSTM模型的光通信网络数据传输负载预测方法。构建光网络通信数据传输信道模型,计算数据传输速率,结合传输效率计算结果与联合特征选择方法获取数据传输负载序列。构建LSTM预测模型,采用粒子群算法对LSTM预测模型进行优化,将获取的数据传输负载序列输入到LSTM模型中,实现数据传输负载预测。实验结果表明,该方法的光通信网络数据传输负载预测精度平均值为98.9%,预测时间平均值为0.64 s,服务拒绝率最高仅为14.5%。
Massive data transmission in optical communication network leads to the increase of optical communication channel load and the difficulty of data transmission load prediction. A data transmission load prediction method based on LSTM model is proposed. The data transmission channel model of optical network communication is constructed, the data transmission rate is calculated, and the data transmission load sequence is obtained by combining the transmission efficiency calculation results with the joint feature selection method. Build the LSTM prediction model,use particle swarm optimization algorithm to optimize the LSTM prediction model, and input the obtained data transmission load sequence into the LSTM model to realize the data transmission load prediction. The experimental results show that the average prediction accuracy of data transmission load in optical communication network is 98. 9%, the average prediction time is 0. 64 s, and the highest service rejection rate is only 14. 5%.
作者
相富钟
赵庆海
XIANG Fuzhong;ZHAO Qinghai(Shandong University of Technology,Zibo 255000,China;TaiShan University,Tai'an 271000,China)
出处
《激光杂志》
CAS
北大核心
2023年第2期154-158,共5页
Laser Journal
基金
山东省科学技术发展项目(No.20SJG022)。
关键词
LSTM模型
光通信网络数据
传输负载预测
传输信道
联合特征选择
LSTM model
optical communication network data
transmission load prediction
transmission channel
joint feature selection
