Modeling Passenger Travel Behavior And Guidance Optimization For Disruption Management In The Urban Rail Transit Network

The urban rail transit system is a unique public transit mode with huge capacity,green and sustainability,high reliability,relatively high speed.It plays a backbone role in solving the problems of traffic jams,traffic pollution in urban areas.Till the end of 2015,there are 25 cities with operating 110 subway lines and 3293 miles in total in the Mainland of China.Meanwhile,some huge cities,like Beijing and Shanghai,own over 550 kilometers of subway lines separately.With the growth in the scale of the rail transit network,the complexity of the network structure,the diversities of the transfer organizations and the train running modes,the urban rail transit is stepping into the era of networked operation.However,urban rail transit is relatively more closed and massed system,the safety of which is especially quite important.Moreover,with the network scale becoming larger,more and more daily disruption incidents occur.In this situation,once some disruptions occur in the partial of the rail network,it not only would make a significant impact on the normal local operations,but also may make a huge shock on the operation in the nearby stations and subway lines.Even worse,it may cause the on-ground traffic to break down.Therefore,it is urging to study how to recognize,management and control this kind of crises.This thesis focuses on the problems of disruption management,including the passenger travel behavior modeling and optimization of guiding passengers under the alternative disruption scenarios.And this thesis aims at revealing the interactive mechanism among the disruption incidents,disruption management strategies and the passenger travel behavior.And then,the model-driven and data-driven methods will be adopted to support our study,more contents are listed below:(1)Macroscopic method of recognizing the impact of operational disruption incidents.Firstly,a simulated-attack method based on the complex network theory is proposed to explore the significant influence of different types of operational disruptions on the macro connectivity efficiency of urban rail network.The method can help the operators to find the key positions in the network,and then help them to make targeted enhancement in operational measures and investment.Secondly,a data-driven Bayesian prediction method is adopted to estimate the extent,time-spatial scope of the disruption impact on passenger flow demand.Moreover,the disruption index is proposed.A large-scale historical operational data is adopted to carry out the comprehensive,quantitative and dynamic identification of the impact of the operational disruption.(2)Modeling passenger travel behavior under alternative disruption scenarios.This paper is aimed at revealing the decision-making mechanism of passenger travel behavior under the condition of complex station closure and line/segment disruption.Firstly,the passenger behavior optimization model under the condition of alternative closure scenarios is constructed,and then a mixed solution algorithm based on passenger flow simulation is proposed.Secondly,in the line/segment disruption scenarios,a mathematical optimization model of rescheduling the train timetable is constructed.And then,an algorithm to find K-shortest paths in time-expended rail network is proposed,which is the important basis of passenger’s decision-making in segment disruption.Finally,the decision-making mechanism of passenger travel behavior under segment disruption is proposed,which is also integrated with the above-mentioned models.(3)Bi-programming based optimization and modeling of passenger inducement.The bi-level programming model of optimizing passenger inducement is proposed,considering the scope and content of the inducement information dissemination.The lower level is to model the passenger travel behavior under the inducement information and alternative disruption scenarios.The upper layer model depicts the decision variables of the scope of the induced information,and designs a hybrid intelligent algorithm for the integration of a passenger flow simulation and a genetic algorithm.The main innovations of this paper are as follows:(1)the improved complex network method is proposed to quantitatively estimate the influence of different types of disruption,and then a data-driven Bayesian method to recognize the impact of alternative disruption scenarios on the macro passenger flow demand.(2)In this paper,the fine behavior model under the complex and uncertain station closure scenarios is proposed.This is the first systematic definition and description of the station closure problem at home and abroad.The study also expands the boundaries of disruption management and belongs to the original theoretical innovation.(3)This thesis puts forward a trinity modeling method integrated with the timetable rescheduling,time-expended k-shortest paths finding algorithm and passenger travel behavior selection model.In this way,the estimation of passenger travel delay time under line/segment disruption is transformed into the scheduled path finding problem.And it is more refined to describe the passengers’ decision-making behavior in the urban rail network.It belongs to the original theoretical innovation in a certain degree.(4)This paper also puts forward the bi-level programming model and hybrid intelligent solution algorithm for passenger flow inducement optimization in urban rail transit network.This study fills the blank of the theatrical study of urban rail transit passenger flow inducement optimization,which promotes the research of urban rail passenger flow from system development to scientific theoretical study.Our study deepened the theoretical depth of the existing research and belongs to the innovation both in the application method and in the original theory.