Research On Passenger Transportation Demand Forecasting Of Urban Agglomeration Base On Neural Network

The formation and development of urban agglomeration has become a global typical socio-economic phenomenon,this trend brought population growth and substantial expansion of the regional scope within it, directly result the traffic contact between urban and regional become more frequent, reasonable transportation network is the material basis of healthy and rapid development of urban agglomeration, it is necessary to make targeted urban transportation planning, determine the direction of development and the reasonable scale of traffic network. As a major component, transportation demand forecasting is a important technique to determine it. Combined with the characteristics of urban agglomeration, a reasonable model for traffic demand forecasting is very important.With analysis and summary of existing studies, built traffic demand forecasting model according to characteristics of urban agglomeration. Firstly, elaborate the definition, the formation and spatial distribution patterns, describe the passenger mechanism of urban agglomeration. On this basis, describes and summarizes the characteristics of demand for urban transport demand, summed up the major factors impact the transport demand.In the urban agglomeration passenger trip generation forecast,listing the time series method,regression analysis and elasticity coefficient method,and the combined model of them. And pointed out their shortcomings. On this basis, proposed a model using a variety of related factors based on neural network.Finally,suggest a combine model of traffic distribution and mode split,simplified four-stage method of forecasting process, improve the accuracy, and modeled after the traditional double-constrained gravity model to improved, the introduction of the adjustment factor further improve the accuracy of the results.