基于改进BP网络的时延预测建模与仿真  

Modeling and Simulation of Time Delay Prediction Based on Improved BP Network

作　　者：李世峰 LI Shi-feng(School of Physics and Technology,Tangshan Normal University,Tangshan 063000,China)

机构地区：[1]唐山师范学院物理科学与技术学院,河北唐山063000

出　　处：《唐山师范学院学报》2022年第6期56-59,共4页Journal of Tangshan Normal University

基　　金：唐山师范学院博士基金项目(2019A03)。

摘　　要：针对远程控制中的时延预测需求,考虑典型BP网络学习方法导致的局部最优问题,对权值调整规则进行了改进。基于改进后的学习方法,提出了专用于Internet时延的预测模型。仿真结果表明,改进后的Internet时延预测模型在预测精度和稳定性方面的性能有所提高,且预测误差均值低于10%,满足实际需求。According to the time delay prediction requirement of the remote control,a model for Internet time delay prediction was proposed.The local optimization problem of the learning method of typical BP network was considered and the weight adjustment algorithm was improved.According to the improved learning method,a model for Internet time delay prediction was proposed.The simulation results show that the performance of the prediction model has improved in the prediction accuracy and the stability.The average error is less than 10%which satisfies the actual demand well.

关 键 词：Internet时延 时延预测 BP网络 建模与仿真 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]