基于随机森林算法的JSQ6型车走行时间预测研究  

Travel Time Prediction of JSQ6-Type Vehicle Based on Random Forest Algorithm

作　　者：冀柯维 王莹[1] 张家瑞 周姗姗 李苏 JI Kewei;WANG Ying;ZHANG Jiarui;ZHOU Shanshan;LI Su(School of Traffic and Transportation,Beijing Jiaotong University,Beijing 100044,China;Transport Control Center,China Railway Special Cargo Logistics Co.,Ltd,Beijing 100055,China)

机构地区：[1]北京交通大学交通运输学院,北京100044 [2]中铁特货物流股份有限公司运输调度中心,北京100055

出　　处：《铁道运输与经济》2024年第12期145-151,共7页Railway Transport and Economy

基　　金：中铁特货物流股份有限公司科技研究开发计划项目(ZTTH-2022-WT-032)。

摘　　要：在铁路货物运输过程中,准确预测车辆的走行时间有助于调度人员提前制定运输组织方案和车辆装卸计划,也便于货主了解货物运到时间。为探讨商品汽车铁路运输专用车辆JSQ6型车走行时间预测问题,在分析JSQ6型车走行特点的基础上,从车辆空间状态和车辆自身状态角度分析,提炼出距到站距离、距到站车站数量、距到站技术站数量、车组中JSQ6型车数量、JSQ6型车编组方式和上报数据来源等影响JSQ6型车走行时间的因素,从提高JSQ6型车走行时间预测的准确度出发,设计了基于随机森林模型的车辆走行时间预测方法。结果表明,该模型的拟合度为0.908,均方差为41.565,相较于既有车辆走行时间预测的结果具有更高的准确度,更适合JSQ6型车的走行时间预测问题,可实现较好的预测效果。Accurately predicting the travel time of vehicles during railway freight transportation helps dispatchers develop transportation organization plans and vehicle loading and unloading plans in advance and facilitates shippers to be informed of the arrival time of goods.To explore the problem of predicting the travel time of JSQ6-type vehicles that specialize in transporting commercial cars through railways,this article analyzed the travel characteristics of JSQ6-type vehicles and extracted factors that affected the travel time of JSQ6-type vehicles,such as distance from the destination station,number of stations from the destination station,number of technical stations from the destination station,number of JSQ6-type vehicles included in the vehicle set,JSQ6-type vehicle grouping method,and source of reported data from the perspectives of vehicle spatial state and vehicle self-state.To improve the travel time prediction accuracy of JSQ6-type vehicles,a travel time prediction method of vehicles based on a random forest model was designed.The results show that the goodness of fit of the model is 0.908,and the mean square error is 41.565.Compared with the prediction of the travel time of existing vehicles,it has higher accuracy and is more suitable for the travel time prediction of JSQ6-type vehicles,achieving better prediction results.

关 键 词：铁路货物运输 车辆走行时间 影响因素 预测模型 随机森林 

分 类 号：U292.4[交通运输工程—交通运输规划与管理]