Identification And Improvement Of Subway Service Quality Based On Smart Card Data

To achieve sustainable development and improve the competitiveness of the industry,subway operating companies are paying more attention to improving service levels while ensuring operational safety.With the development of intelligent transportation,the rapid and accurate knowledge of subway service levels has also become a hot topic of research in recent years.This thesis proposes an automatic evaluation method of subway service quality based on the smart card data,which breaks the lag and subjectivity of the previous evaluation system combining questionnaire survey and expert evaluation.The model corresponds to two different level respectively:OD pairs and network.A multi-stage train control model based on travel time distribution is established to enhance the quality of subway services and its service sustainability level.The main work of the thesis consists of the following three parts.（1）Based on the smart card data generated in the subway system during pre-operations,the classification model set of passenger travel time distribution is established.By means,variance,kurtosis and skewness,k-means clustering analysis method and K-S test method are used to classify passenger travel time in different operating periods and different ODs pair.By the clustering results,the operation characteristics of the stations in periods are inferred through the distribution of the classification in same period and the data characteristics.（2）Based on the evaluation results of station pairs,the thesis proposes a service quality weight index(₆))as a measurement tool for the overall level of subway operations and passenger satisfaction.The feasibility and rationality of the proposed method is verified by using the Chengdu Metro smart card data as an example.By linking key operational indicators with subjective indicators of passenger satisfaction,the operating lines are visually assessed in the form of scores from the dual perspectives of passengers and companies,in both directions of time and space,effectively improving the automation ratio of the subway service level evaluation system.（3）To improve the quality and sustainability of subway services,this thesis establishes a multi-stage train operation framework model based on the pseudo-spectral method using the train travel time distribution method to solve the optimal control strategy for train operation control.The solution is validated using Shenzhen Metro Line 5 as an example,and the results show that the train energy consumption is 2.78%less than the original energy consumption in the optimal train operation solution.