Research On Modeling Of Multimodal Traffic Network Assignment Based On Hypergraph

With the rapid development of multi-mode transportation system,travelers have more possibilities in making travel choice,but the complex multi-mode transportation network also makes it more difficult to predict the traffic flow accurately.In order to reduce the difficulty of prediction,the hypergraph theory is used to construct the multi-mode traffic network which can simplify multi-mode traffic network modeling.In order to improve the accuracy of prediction,the cost function is defined from various dimensions,and the impact of capacity limitation and path overlap on traffic distribution is analyzed under multi-mode conditions,and the corresponding multi-mode traffic distribution model is constructed.The main research includes four parts.A multi-modal transportation network model is constructed based on hypergraph.The composition structure and travel characteristics of the multi-modal transportation system are discussed,and the characteristics of the multi-modal network structure are clarified.Based on this,this paper proposes a new method for network construction based on hypergraph which considering the shortcomings of existing network modeling methods.The modeling foundations of Traffic Assignment Modeling which based on Multimodal Hypergraph Networks is constructed.Firstly,by comparing the advantages and disadvantages of commonly used network storage structures,a hyper-edge linked list which saves more storage and more in line with the hypergraph network topology is designed.Then,the multimodal generalized travel cost is defined from the four dimensions including time,cost,comfort,and transfer penalty.Combining with the hypergraph storage structure,the definition of route generalized cost is proposed.Finally,based on the depth-first traversal algorithm,combining the constraints of effective paths and the storage characteristics of hypergraphs,an effective path traversal algorithm for hypergraph networks is proposed.A multi-modal traffic assignment model is proposed which considering capacity constraints and overlapping paths.Under multi-mode conditions,the impacts of capacity constraints and overlapping paths on traffic allocation are analyzed,and corresponding multimode stochastic user equilibrium conditions are proposed.Equivalent non-linear programming models are constructed,and the solution algorithm of the model is designed by combining the ALM algorithm and the MSWA algorithm.The example verifies the effectiveness of the proposed model and algorithm through the comparison experiments.The parking rate and the frequency of public transportation are taken as examples for sensitivity analysis to reflect the impact of changes in generalized travel costs on traveler choice.The research results provide a reference for multi-modal transportation network modeling and traffic assignment.