基于MI-PSO-RBF神经网络的铁路客货运量预测研究  

Railway Passenger and Freight Volume Prediction Based on MI-PSO-RBF Neural Network

作　　者：薛锋[1,2,3] 吴林鸿 汪雯文 周琳 XUE Feng;WU Linhong;WANG Wenwen;ZHOU Lin(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 611756,Sichuan,China;National and Regional Joint Engineering Laboratory of Comprehensive Intelligent Transportation,Southwest Jiaotong University,Chengdu 611756,Sichuan,China;National Engineering Laboratory of Integrated Transportation Big Data Application Technology,Southwest Jiaotong University,Chengdu 611756,Sichuan,China)

机构地区：[1]西南交通大学交通运输与物流学院,四川成都611756 [2]西南交通大学综合交通运输智能化国家地方联合工程实验室,四川成都611756 [3]西南交通大学综合交通大数据应用技术国家工程实验室,四川成都611756

出　　处：《铁道运输与经济》2024年第9期123-135,共13页Railway Transport and Economy

基　　金：国家铁路局科研项目(市场委合2022-3/2024-2号);四川省哲学社会科学规划项目“四川统计发展”专项课题(SC23TJ038);四川循环经济研究中心项目(XHJJ-2405)。

摘　　要：准确地预测铁路客货运量对合理配置运输资源、提高铁路客货运组织工作效率有重要作用。为提高铁路客货运量的预测精度,提出一种基于MI-PSO-RBF神经网络的客货运量组合预测模型。本研究对铁路客货运量的影响因素及其内在关联进行分析,选取相关指标,利用互信息素法对指标进行筛选,构建影响因素指标体系。基于该指标体系,运用粒子群算法优化的RBF神经网络模型分别对铁路客货运量进行预测,并与传统的BP神经网络、RBF神经网络预测模型进行比较。结果显示,经过参数调整优化后的MI-PSO-RBF神经网络在铁路客运量及货运量的预测精度方面表现最佳,测试集R2分别达到了0.9481与0.9911,具有较高的精度及泛化能力,表明该组合预测模型能够进一步提升神经网络模型预测铁路客货运量精确度。Accurate prediction of railway passenger and freight volume is of positive significance to the rational allocation of transportation resources and the improvement of the organization efficiency in railway and freight transportation.To improve the prediction accuracy of railway passenger and freight volume,a combined passenger and freight volume prediction model based on the MI-PSO-RBF neural network was proposed.In this study,the influence factors of railway passenger and freight volume and their internal correlation were analyzed,and the mutual pheromone method was adopted in screening relevant indexes to construct the influence factors index system.Based on this system,this study applied the RBF neural network model optimized by the particle swarm algorithm in predicting railway passenger and freight volume respectively,and compared this model with the traditional BP neural network and RBF neural network prediction models.The results show that the MI-PSO-RBF neural network has the best performance in accurately predicting railway passenger and freight volume after parameter adjustment and optimization,and the R2 in the test sets reaches 0.9481 and 0.9911,respectively,demonstrating high accuracy and generalization ability.The proposed combined prediction model is important for improving the prediction accuracy of the railway passenger and freight volume by neural network models.

关 键 词：客货运量预测 互信息素 粒子群算法 RBF神经网络 影响因素法 

分 类 号：U239.5[交通运输工程—道路与铁道工程]