OD Forecast Of Urban Rail Transit Passenger Flow Based On Passenger Flow Trend

In the background of the continuous expansion of the urban rail transit network and the rapid growth of passenger traffic,the operation is facing huge challenges.This paper studies the short-term demand of passenger flow OD in the actual work of the road network operation department.Considering the available data of the operation department,time series data analysis and machine learning modeling methods are used to predict the passenger flow OD of the road network in one or more days in the future,providing a quantitative basis for operational decision-making.The main research contents are as follows:(1)This paper analyzes the characteristics of urban rail,the influencing factors of the passenger flow OD of the road network and the characteristics of passenger flow OD under the network operation condition.According to the actual requirements and available data conditions,Based on the analysis of the characteristics of the passenger flow OD prediction problem and the summary of the characteristics of the existing passenger flow prediction methods,the framework and process of the short-term passenger flow OD prediction method are proposed,which consists of the extraction of passenger flow characteristics,the mapping of passenger flow characteristics and the timing prediction of passenger flow.(2)Aiming at the problem of feature extraction of passenger flow,the characteristics of passenger flow OD in multiple time dimensions are defined and expressed on the basis of the analysis of the changing rules and characteristics of inbound passenger flow and inbound passenger flow.Further,according to the data characteristics of inbound passenger flow data and passenger flow destination data,the corresponding trend feature extraction methods are proposed,which is that the trend characteristics of inbound passenger flow were extracted by STL time series decomposition method.the trend characteristics of inbound passenger flow is extracted by clustering analysis method and the corresponding model is built.(3)Based on the results of passenger flow feature extraction,a prediction framework combining passenger flow feature mapping and time series analysis is proposed.The random forest model is used to construct the mapping relation model between the OD structure feature and date attribute of passenger flow in the segment of passenger flow feature mapping.In the OD time series prediction link,the research has formed the inbound passenger flow time series prediction method under the phase trend and cycle trend characteristics,and the OD structure time series prediction method for passenger flow under the OD structure characteristics and finally formed the station passenger flow OD vector prediction,and combined to form the passenger flow OD matrix of road network.(4)Taking urban rail transit of a city as a case study object,the research method proposed in this paper is used for verification.