Prophet框架下生产物流运输节点的阻塞预警方法  

Congestion Warning Method for Production Logistics Transportation Nodes under Prophet Framework

作　　者：曹小华[1] 刘道凡 涂圣才 CAO Xiaohua;LIU Daofan;TU Shengcai(School of Logistics Engineering, Wuhan University of Technology, Wuhan 430063, China)

机构地区：[1]武汉理工大学物流工程学院,武汉430063

出　　处：《武汉理工大学学报（交通科学与工程版）》2020年第2期255-259,264,共6页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家重点研发计划(2018YFC1407405);中央高校基本科研业务费资助(195218005)资助。

摘　　要：针对生产物流阻塞异常的问题,提出一种基于Prophet框架的生产物流运输节点的阻塞预警方法.建立表征阻塞状态的阻塞预警模型,通过预测出各运输路径的流入/流出速度及累积流入/流出量以实现对该物流运输节点的阻塞预警.在对原始物流数据进行预处理的基础上,提出了基于Prophet框架的物流运输节点的流速度预测方法,实现了对运输路径的流入/流出速度的有效预测.实验结果表明,该预测方法性能优良,比经典的SARIMA和LSTM预测方法更适用于物流运输节点的流入/流出速度预测,从而实现对生产物流阻塞异常的预警.In order to solve the problem of abnormal congestion in production logistics,a congestion early warning method for production logistics transportation nodes based on Prophet framework was proposed.Firstly,a congestion warning model was established to characterize the congestion status.By predicting the inflow/outflow speed and cumulative inflow/outflow volume of each transportation route,the congestion warning for this logistics transportation node was realized.Then,on the basis of preprocessing the original logistics data,a flow velocity prediction method of logistics transportation nodes based on the Prophet framework was proposed,which realizes the effective prediction of the inflow/outflow velocity of the transportation path.The experimental results show that the prediction method has excellent performance,and is more suitable for the inflow/outflow speed prediction of logistics transportation nodes than the classical SARIMA and LSTM prediction methods,thus realizing the early warning of abnormal congestion of production logistics.

关 键 词：生产物流 运输节点 阻塞预警 时序预测 PROPHET 

分 类 号：U675.7[交通运输工程—船舶及航道工程]