Study On Section Passenger Flow Uncertainty Of Urban Rail Transit

The predictive value of section flow is the basis to compile the plan of station operation in urban rail transit,and the accuracy of the prediction result directly affects the configuration of the rail transit infrastructure and auxiliary facilities.However,under the influence of numerous uncertain natural factors and social factors,there are differences in section flow between the prediction result and the actual data.Transfer station is as an important support of rail transit network,and its influence factors is more complex so that it is more difficult to grasp accurate prediction.And based on the accurate prediction of passenger flow,the effective management measures by rail transit administrators provide the rail transit operation a powerful guarantee of safety and order,while sometimes the decisions are far from achieving the desired result.The cause of the deviation is section flow has the characteristics of time variation and uncertainty,which means different sections in the face of external shocks have different sensitivities.This leads section flows have different volatilities.For this reason,we can't use the same prediction indexes and decision strategies when to deal with different section flows.Therefore,it is of great significance to research on the uncertainty of section flow in rail transit.The outcomes not only could provide exact prediction of section flow with theoretical support,but also could offer a reference for the stage of preliminary planning and design as well as post operation management.Hence,based on large quantity statistical data of section flow,the GARCH family models that could precisely depict the time variation and uncertainty of time series is applied to build the uncertainty model of section flow.To be specific,the research work mentioned in the paper is mainly about:(1)The basic statistical characteristics of section flow is analyzed in detail.Then the generalized error distribution is applied to describe the sequence which has the characteristics of leptokurtosis and heavy tails,volatility cluster and asymmetry,thus to provide a basis for subsequent modeling and parameters calibration.(2)In the paper,the TGARCH model has been built to solve the problem that time series of section flow is asymmetric to volatility.Moreover,in order to make up for the deficiency of GARCH model which can't describe the changes of susceptibility at different times,dummy variables are introduced.The asymmetry of time series based on passengers,thus,could be portrayed.As a result,the model could better reflect the time-varying characteristics of section flow.(3)The VAR model and variance decomposition are both used to analyze the correlation between transfer flow and section flow,which provides transfer stations with theoretical support for their planning and decision-making accordingly.