Study On Combined Models Of Demand Forecasting And Algorithms For Urban Mixed Traffic

Traffic demand forecast is the essential link in the urban traffic planning and the important basis to make the city traffic planning scheme. Considering the characteristics of the mixed traffic in China and the actual traffic network, this thesis mainly studies combined models and algorithms for traffic demand forecast.Chapter 1 briefly introduces the importance of the urban traffic, existing problems, the current situation and the developing trend of traffic demand fore-casting combined models. Moreover, the basic knowledge and main contents which are researched are presented.In Chapter 2, assuming that the influence of the link travel function on the impedance is symmetric for the typical mixed traffic, the author researches a symmetrical double-constrained combined model of OD matrix and traffic assignment. The equivalence of the first-order condition and the existence and uniqueness of solution of the model are proved by optimization theory. Meanwhile, the algorithm and a example are given.Based on the deviation of travelers on path impedance and the variability of traffic demand, combining with the characteristics of two-way mixed traffic, Chapter 3 discusses the stochastic user equilibrium assignment problem under the elastic demand. For the consideration of the traffic demand associated with the path impedance of the whole traffic network in the network and the characteristics of the urban mixed traffic, a bi-level programming model is built for trip distribution and stochastic user equilibrium in the final chapter. On this basis, a specific algorithm is proposed and an example is given to demonstrate its feasibility.