基于VMD-BP神经网络模型的铁路车站月度客流发送量预测研究
|
摘要
为确定合理的列车开行对数,首先分别以BP神经网络(BP Neutral Network)和极限学习机(Extreme Leaning Machine, ELM)构建基准模型;在此基础上,结合经验模态分解算法(Empirical Mode Decomposition, EMD)和变分模态分解算法(Variational Mode Decomposition,VMD)适合处理非线性、非平稳数据的特点,对客流发送数据进行处理,然后分别与基准模型组合形成组合模型,并提出了月度客流发送量的概念,建立了滚动的月度客流发送量预测机制。最后,以兰州西站衔接各线的客流日发送数据为原始数据,以各基准模型和组合模型对月度客流发送量进行了预测,并对预测结果和指标加以分析对比。研究结果表明:(1)以客流发送量作为原始数据,VMD算法分解效果优于EMD算法;(2)对比各基准模型和组合模型,VMD-BP神经网络组合模型预测精度最高,适用于铁路车站月度客流发送量预测问题。
In order to determine the reasonable pairs of trains, the benchmark model was established firstly based on BP neural network and ELM(Extreme Leaning Machine) respectively. On this basis,EMD(Empirical Mode Decomposition) algorithm and VMD(Variational Mode Decomposition) algorithm that were suitable for dealing with the nonlinear and non-stationary data were used to process the passenger flow transport volume. Then combined with the benchmark models, combined models were formed. The concept of monthly passenger flow transport volume was proposed, and the rolling monthly passenger flow transport volume forecast mechanism was established. Finally, taking the daily passenger flow transport volume of Lanzhou West Railway Station as the original data, monthly passenger flow transport volume was forecasted by each benchmark model and combined model, and the forecast results and indicators were analyzed and compared. The results show that:(1) taking the passenger flow transport volume as the original data, the decomposition effect of VMD algorithm is better than EMD algorithm;(2) comparing benchmark models with combined models, the VMD-BP neural network combined model has the highest forecast accuracy and is suitable for forecasting the monthly passenger flow transport volume at railway stations.
In order to determine the reasonable pairs of trains, the benchmark model was established firstly based on BP neural network and ELM(Extreme Leaning Machine) respectively. On this basis,EMD(Empirical Mode Decomposition) algorithm and VMD(Variational Mode Decomposition) algorithm that were suitable for dealing with the nonlinear and non-stationary data were used to process the passenger flow transport volume. Then combined with the benchmark models, combined models were formed. The concept of monthly passenger flow transport volume was proposed, and the rolling monthly passenger flow transport volume forecast mechanism was established. Finally, taking the daily passenger flow transport volume of Lanzhou West Railway Station as the original data, monthly passenger flow transport volume was forecasted by each benchmark model and combined model, and the forecast results and indicators were analyzed and compared. The results show that:(1) taking the passenger flow transport volume as the original data, the decomposition effect of VMD algorithm is better than EMD algorithm;(2) comparing benchmark models with combined models, the VMD-BP neural network combined model has the highest forecast accuracy and is suitable for forecasting the monthly passenger flow transport volume at railway stations.
引文
[1]骆晨,刘澜.基于多次修正残差灰色模型的铁路客流预测[J].华东交通大学学报,2013(4):19-23.
[2]陈淑燕,王炜.交通量的灰色神经网络预测方法[J].东南大学学报(自然科学版),2004,34(4):541-544.
[3]李海军,张玉召,朱昌锋.基于灰色模型及月度比例系数法的铁路客流预测方法[J].西部交通科技,2012(3):61-64,75.
[4]邹巍,陆百川,邓捷,等.基于遗传算法与小波神经网络的客流预测研究[J].武汉理工大学学报(交通科学与工程版),2014,38(5):1148-1151,1157.
[5]郭雅楠.基于EMD的BP神经网络在公交客流预测中的应用[D].天津:天津大学,2016.
[6]潘杉.基于时间序列的铁路客流量预测[D].广州:华南理工大学,2017.
[7]周璐婕,董昱.基于GA-BP神经网络的列控车载设备故障诊断方法研究[J].铁道科学与工程学报,2018,15(12):3257-3265.
[8]王长鹏,齐俊,蔡永宁,等.基于改进PSO-BP神经网络在径流量预测中的应用[J].城市勘测,2019(2):199-203.
[9]贺坤,李星野.基于经验模态分解和遗传算法改进的神经网络模型的风速时间序列预测[J].电子测量技术,2018,41(8):146-149.
[10]陆利军,廖小平.基于EMD-BP神经网络的游客量预测研究[J].统计与决策,2019,35(4):85-89.
[11]梁洪卫,张旭,邹岱峰.基于VMD-BP神经网络模型的天然气管道工况检测研究[J].化工自动化及仪表,2017,44(7):633-636,655.
[12]李俊甫,吕红芳,何力,等.基于核极限学习机的电厂锅炉NOx排放建模与预测[J].工业控制计算机,2018,31(9):85-86,89.
[13]刘霖雯,刘超,江成顺.EMD新算法及其应用[J].系统仿真学报,2007,19(2):446-447.
[14]孟凡煦,张琰骏.基于VMD和ELM的超短期负荷预测方法[J].国外电子测量技术,2019,38(1):24-28.
[15]赵征,汪向硕,乔锦涛.基于VMD和改进ARIMA模型的超短期风速预测[J].华北电力大学学报(自然科学版),2019,46(1):54-59.
[16]DRAGOMIRETSKIY K,ZOSSO D.Variational Mode Decomposition[J].IEEE Transactions on Signal Processing,2014,62(3):531-544.
[17]王玲,程耕国,袁志强,等.关于风力发电系统短期风速预测仿真[J].计算机仿真,2018,35(11):98-102.
[18]吴华稳,甄津,王宇,等.基于径向基神经网络的铁路客货运量预测研究[J].铁道科学与工程学报,2014(4):109-114.
[19]成强.基于小波包与长短时记忆融合的铁路旅客流量预测模型[J].计算机系统应用,2018,27(7):121-126.
[20]李梅,李静,魏子健,等.基于深度学习长短期记忆网络结构的地铁站短时客流量预测[J].城市轨道交通研究,2018,21(11):42-46,77.
[21]林培群,陈丽甜,雷永巍.基于K近邻模式匹配的地铁客流量短时预测[J].华南理工大学学报(自然科学版),2018,46(1):50-57.
[2]陈淑燕,王炜.交通量的灰色神经网络预测方法[J].东南大学学报(自然科学版),2004,34(4):541-544.
[3]李海军,张玉召,朱昌锋.基于灰色模型及月度比例系数法的铁路客流预测方法[J].西部交通科技,2012(3):61-64,75.
[4]邹巍,陆百川,邓捷,等.基于遗传算法与小波神经网络的客流预测研究[J].武汉理工大学学报(交通科学与工程版),2014,38(5):1148-1151,1157.
[5]郭雅楠.基于EMD的BP神经网络在公交客流预测中的应用[D].天津:天津大学,2016.
[6]潘杉.基于时间序列的铁路客流量预测[D].广州:华南理工大学,2017.
[7]周璐婕,董昱.基于GA-BP神经网络的列控车载设备故障诊断方法研究[J].铁道科学与工程学报,2018,15(12):3257-3265.
[8]王长鹏,齐俊,蔡永宁,等.基于改进PSO-BP神经网络在径流量预测中的应用[J].城市勘测,2019(2):199-203.
[9]贺坤,李星野.基于经验模态分解和遗传算法改进的神经网络模型的风速时间序列预测[J].电子测量技术,2018,41(8):146-149.
[10]陆利军,廖小平.基于EMD-BP神经网络的游客量预测研究[J].统计与决策,2019,35(4):85-89.
[11]梁洪卫,张旭,邹岱峰.基于VMD-BP神经网络模型的天然气管道工况检测研究[J].化工自动化及仪表,2017,44(7):633-636,655.
[12]李俊甫,吕红芳,何力,等.基于核极限学习机的电厂锅炉NOx排放建模与预测[J].工业控制计算机,2018,31(9):85-86,89.
[13]刘霖雯,刘超,江成顺.EMD新算法及其应用[J].系统仿真学报,2007,19(2):446-447.
[14]孟凡煦,张琰骏.基于VMD和ELM的超短期负荷预测方法[J].国外电子测量技术,2019,38(1):24-28.
[15]赵征,汪向硕,乔锦涛.基于VMD和改进ARIMA模型的超短期风速预测[J].华北电力大学学报(自然科学版),2019,46(1):54-59.
[16]DRAGOMIRETSKIY K,ZOSSO D.Variational Mode Decomposition[J].IEEE Transactions on Signal Processing,2014,62(3):531-544.
[17]王玲,程耕国,袁志强,等.关于风力发电系统短期风速预测仿真[J].计算机仿真,2018,35(11):98-102.
[18]吴华稳,甄津,王宇,等.基于径向基神经网络的铁路客货运量预测研究[J].铁道科学与工程学报,2014(4):109-114.
[19]成强.基于小波包与长短时记忆融合的铁路旅客流量预测模型[J].计算机系统应用,2018,27(7):121-126.
[20]李梅,李静,魏子健,等.基于深度学习长短期记忆网络结构的地铁站短时客流量预测[J].城市轨道交通研究,2018,21(11):42-46,77.
[21]林培群,陈丽甜,雷永巍.基于K近邻模式匹配的地铁客流量短时预测[J].华南理工大学学报(自然科学版),2018,46(1):50-57.