Research On Urban Dynamic Network Traffic Assignment And Relevant Problem

Intelligent Transportation System (ITS), as an advanced integrated traffic operation and management system, is being deeply developed and applied in every country in the world nowadays. Urban dynamic network traffic assignment theory is one of the most crucial technologies in ITS, and the integrating control system with guidance system is the direction of the urban traffic management's development, which is one of the most important researches in ITS. Based on these, this paper mainly researches the dynamic network traffic assignment problem based on road traffic flow model, and the combined problem of dynamic traffic assignment and traffic signal control in saturated urban transportation network. The main works of this thesis are as follows:Firstly, we study the inequality equivalent to the dynamic user optimal principle (DUO). By extending the previous results of static configuration, two inequalities equivalent to the instantaneous DUO principle and the ideal DUO principle respectively and the equivalent proofs are provided. Furthermore we discuss the inequalities' application, such as the equivalent DUO constraints in formulating the relative practical problems, the stop criterion and the minimal route time calculating.Secondly, adopting the idea of combination for the road traffic flow model and the network traffic assignment model, we probe into the dynamic network traffic assignment problem based on road traffic flow model and mainly do the following three works:Towards the classic model-LWR equations, we formulate a dynamic user optimal assignment problem based on the Cell Transmission Model by using link variables. Based on CTM, traffic flow dynamic propagation on the urban network is simulated. Then a method for estimating the link travel time satisfying FIFO principle is developed. Furthermore DTA satisfying the DUO principle is completed.Extending the studies above, for the different types of vehicles driving on the dynamic urban network, we probe into the multiple user class dynamic network traffic assignment problem based on traffic flow model. The distinction between user classes is typically based on vehicle travel speed. The asymmetric interactions in the multiclass traffic flow are taken into account. Besides, dynamic network user optimal assignment guarantees the FIFO principle is hold for the same class user. Since the LWR model has some shortcomings in describing traffic flow dynamic phenomena, we further adopt a more advanced continuum traffic flow model-Speed Gradient model to describe traffic flow propagation on dynamic network and give a method to estimate actual link travel time, which provide an elementary work for the combination of road traffic flow model with the network traffic assignment model.Finally, we also research the combined problem of dynamic traffic assignment and traffic signal control on congested network and its algorithm. Considering the combination of dynamic traffic assignment and traffic signal control, the generalized bi-level programming model of dynamic network traffic signal optimal problem is proposed, and especially analyzes two procedures for determining the equilibrium queuing delays on saturated link. For solution, a chaotic optimal algorithm is developed for the bi-level model, and furthermore a numerical example is given to illustrate advantages of the proposed model and performance of solution algorithm.