| The delay of high-speed railway trains has always affected the quality and efficiency of passenger transportation in China.It is the key for dispatchers to study and master the regularity of delay recovery and improve the ability of delay recovery,resulting in making more reasonable dispatching decisions in time and maintaining the usual train operation rule.As an important part of railway transportation system,the studying of delay recoveries of off-line trains can help to improve the quality of high-speed railway operation and satisfaction of passengers.It provides a basis for deepening the theoretical research on dispatching of high-speed railways,so as to further guide dispatching practical work with the help of theories.In this thesis,based on the operating performance data of high-speed rail off-line trains,the regularity of delay recovery of off-line trains is studied,and the delay recovery ability of trains is improved by optimizing the supplement time assignment.The main work includes:1.This thesis reviews and summarizes the existing research from four aspects: the regularity of off-line train delays,the propagation mechanism of train delays,the recovery regularity and prediction of train delays,and the optimization of supplement time assignment.The research status is analyzed and the deficiencies are found.2.The basic theory of train delays is described,and the method of numerical statistics is used to compare and analyze the delay recovery rules of main-line trains and off-line trains.The results show that the overall delay recovery level of off-line trains is higher than that of the local trains,and their delay recovery regularities are different.Therefore,it is necessary to separately discuss the delay recovery regularities of off-line trains.3.Based on the operation data of the off-line trains of Wuhan-Guangzhou high-speed railway(WG-HSR),the time-space and time-length features of delay recovery are explored.Firstly,the distribution regularities of train delay recovery in each station,section and each time period is analyzed.The research results show that the delay recovery rate of sections is significantly higher than that of stations.Then,the function fitting method is used to construct the delay recovery time distribution model.Through K-S test,the probability distribution models of the delay recovery time of off-line trains at each station,section and each time period are constructed.4.Based on the supplement time data of off-line trains,the setting and utilization of supplement time are analyzed.Respectively calculate the supplement time configuration of each station and section,and calculate the supplement time utilization rate.The analysis results show that the average supplement time of off-line trains at stations is higher than that in sections,and the utilization level of supplement time of sections is higher than that of stations.5.Based on the support vector regression(SVR)algorithm,an off-line trains delay recovery prediction model is constructed.Using the delay recovery data of off-line trains,the independent variables and dependent variable of the model are determined through Spearman correlation analysis.Subsequently,the model is trained and tested based on the SVR algorithm,and the prediction model for the delay recovery of off-line trains is constructed.The prediction results show that the predicted values of this model coincide with the actual values relatively high,and when the allowable error is 1 minute,the model prediction accuracy can reach 95.10%,which has a good effect.Finally,the model is compared with the random forest model and the GBDT model.The comprehensive results show that the SVR model has the best prediction effect.6.With the goal of maximizing the expectation of delay recovery time,the supplement time assignment optimization model is constructed to realize the redistribution of the supplement time in stations and sections of off-line trains.First,the model is described in detail,and the SQP algorithm is chosen to solve this model.Then,the stations and sections between Hengyang East Station and Changsha South Station of WG-HSR are used as the research objects,and the supplement time of off-line trains are re-arranged.After the optimization of the model,the expectation value of delay recovery of the off-line trains increases by 11.52%,indicating that this model can effectively optimize the supplement time assignment of the off-line trains,thereby improving the overall delay recovery capability of trains. |
