基于马尔科夫链-BP神经网络模型对公路运量的预测研究  

Prediction of Highway Traffic Volume Based on MarkovChain-BP Neural Network Model

作　　者：裴同松[1] 裴彧 PEI Tongsong;PEI Yu(Department of Electrical and Information Engineering,Hebei Jiaotong Vocational and Technical College,Shijiazhuang 050091,Hebei,China;Department of Civil Engineering,Hebei Jiaotong Vocational and Technical College,Shijiazhuang 050091,Hebei,China)

机构地区：[1]河北交通职业技术学院电气与信息工程系,河北石家庄050091 [2]河北交通职业技术学院土木工程系,河北石家庄050091

出　　处：《重庆交通大学学报（自然科学版）》2021年第2期35-41,共7页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：河北省高校百名优秀人才计划项目(BR206);河北省科技基金项目(16217604D);2019年河北省高等学校科学技术研究项目(Z2019067)。

摘　　要：选取河北省某地区1998—2017年公路运量数据为例,采用BP神经网络模型进行预测并用马尔科夫链修正预测值,将公路运量实际值与BP神经网络预测值及马尔科夫链修正值作对比分析并预测了2018—2019年的公路运量数据。使用马尔科夫链修正后的BP神经网络预测模型可以将公路客运量和货运量的平均相对误差分别下降至2.07%和2.14%。修正后的模型不仅可以准确的对公路运量做出预测,而且可以为未来公路运输发展提供有利意见。Taking the highway traffic volume data of a region in Hebei Province from 1998 to 2017 as an example,BP neural network model was used to predict and Markov chain was used to correct the predicted value.The actual value of highway traffic volume was compared with the predicted value of BP neural network and the modified value of Markov chain,and the highway traffic volume data from 2018 to 2019 was predicted.The BP neural network prediction model modified by Markov chain can reduce the average relative error of highway passenger volume and freight volume to 2.07%and 2.14%,respectively.The modified model can not only accurately predict the road traffic volume,but also provide favorable opinions for the future development of highway transportation.

关 键 词：交通运输工程 公路运量 马尔科夫链 BP神经网络 预测 

分 类 号：U491.14[交通运输工程—交通运输规划与管理]