Research On Urban Multi-Modal Public Transit Network Fast Construction Method And Assignment Model

Vigorously developing urban multi-modal public transit system is an inevitable requirement for alleviating urban traffic congestion,meeting different levels of transit demand,and improving the overall efficiency of the public transit system.It is also an important means of implementing the ‘Public Transit Priority' strategy and promoting the construction of ‘Transit Metropolis'.In view of the backwardness of the construction technology of public transit assignment network,and the lack of comprehensive consideration of the characteristics of multi-modal public transit system,the impact of urban road traffic operation status and the characteristics of passenger's transit route choice behavior for the present transit assignment models,this dissertation conducts the “Research on Urban Multi-modal Public Transit Network Fast Construction Method and Assignment Model”,to provide a quantitative basis for formulating and evaluating public transit planning,management and control schemes,based on National 973 Program of China “Transit-Oriented Multimode Transportation Network Coordination Mechanism and Coupling Theory”(2012CB725402)and National Natural Science Foundation Key Project of China “Key Technologies for Fundamental Theory and Effectiveness Improvement of Modern Urban Multi-Mode Public Transit System”(51338003).Firstly,the urban road network database structure for storing network topology relationship and basic attribute information is designed,and the technologies for automatically extracting,generating and correcting road network based on OSM data are developed.On the basis of fully analyzing the characteristics of multi-modal public transit system,the public transit network database for storing the attribute information and location information of transit station and line is designed,and based on the Amap Open Platform,the technology of automatically generating data files of transit station and line is developed,and then the automatic matching of the public transit network and the road network is realized.Based on the theory of super-network,the multi-modal public transit super-network model used for assignment is formulated.Secondly,taking the early peak as analysis period,the calculation method of traffic generation and attraction weights of residential and non-residential land is proposed,based on multidimensional land use data.Then,choosing fuzzy C-means clustering method as land clustering algorithm,the location and transit trips of transit demand's generation and attraction points in every traffic zone is calculated,and OD matrix expansion method is presented correspondingly.Thirdly,with full analysis of the waiting process of passengers at the transit station under crowded conditions,the calculation model of the waiting time at the transit station is established based on the theory of stochastic processes.Regarding the type of station platform and the process of getting on and off the transit vehicle for passengers,the stop time calculating model for the transit vehicle at the station is established.Considering the consumption of time,money and comfort feelings,the impedance calculation models of travel link,transfer link,boarding and alighting link are formulated.By analyzing the impact of different fare pricing rules on the impedance of each link,an integrated impedance calculation model for transit travel path with consideration of route fare correction is established.Fourthly,adjacency list is chosen as storage structure for public transit super-network used for transit assignment,and the improved Dijkstra algorithm is applied to calculate the shortest transit travel path between each O-D pair.Based on the judgment rules of transit feasible path and effective path defined in advanced,the depth-first searching algorithm is chosen as the transit effective path searching algorithm by considering the calculation time and the searching results.The capacity limitation assignment,the capacity limitation-multipath assignment and the stochastic user equilibrium assignment models and algorithms for the transit network are studied,and the genetic algorithm is designed to calibrate parameters in assignment models.Finally,research results of this dissertation are transformed into software modules.Taking Nanjing as an example,the basic traffic database and multi-mode public transit network are established based on TranStar software platform.Based on metro card data and structure data of metro network,the statistical approach of metro lines' sectional passenger flow is studied.Then,transit assignment result of single traffic zone center is compared with result of multiple generation and attraction centers,verifying that the latter has the advantage of improving the accuracy of transit assignment;and the Yangtze River Bridge closure scheme,consistent with the actual situation,is set up to verify the influence of no entry measures on passenger flow based on TranStar;and the results of capacity limitation,capacity limitation-multipath and stochastic user equilibrium assignment models are compared,finding that the latter two methods performs better than the first one;and optimal parameters of transit assignment model are obtained by genetic algorithm.