Research On Models And Algorithims For Estimating Urban Multi-modal O-D Demands Based On Multi-Data Sources

O-D (Origin-Destination) demands are the most important basic data for urban transportation planning and management. At present, there are two ways to obtain O-D demands:one is the survey statistics, which method requires a lot of manpower and material resources; another one is built the mathematical model of O-D demands estimation, and this method is convenient, high efficiency and low cost. With the development of the traffic management technology, a large number of real-time traffic data can be obtained effectively, and how to build the O-D demands estimation mathematical model and algorithm with these traffic data, is an important problem of traffic management. In this paper, making the urban multi-mode transportation system as the research object, through the statistical analysis of the multiple traffic data, draw on the priori information that for multi-mode O-D demands estimation. And on the basis of these prior information, we build the static and dynamic multi-mode O-D demands estimation model, and design the solving algorithm. Specifically, the research of this paper mainly from the following three aspects:(1) This paper analyzes the structure characteristics and the acquisition methods of existing urban traffic data, focusing on the traffic survey data, road test data, and bus IC card data. And based on the data characteristics of multi-modal O-D demands, we can get the priori information of O-D demands, road observed flow and passenger travel between bus stations, which used for the static and dynamic O-D demands estimation problem.(2) Using bi-level programming method, this paper build the model for estimating urban multi-modal O-D demands based on multi-data sources. The optimization goal of the upper model is to minimize the the error square which among the estimation value and priori of O-D demands, the estimation value and priori of road traffic flow, and the estimation value and prior of passenger travel between bus stations. The lower model is urban multi-modal traffic assignment model, which established in the condition of known the multi-modal O-D demands value, and through solving the lower model, we can get the approximate relationship between O-D demands and road traffic of different traffic modal. The solution algorithm is also presented. Finally, this paper analyzes and validates the convergence of the model and the calculation results, through a simple example. (3) On the basis of the urban multi-modal static O-D demands based on multi-data sources, this paper uses bi-level programming method, and build the model for estimating urban multi-modal dynamic O-D demands based on multi-data sources. And this paper also gives the solution algorithm of this model. The optimization goal of the upper model is to minimize the the error square which among the estimation value and priori of dynamic O-D demands, the estimation value and observased value of dynamic road traffic flow, and the estimation value and prior value of passenger travel between bus stations. The lower model is urban multi-modal dynamic traffic assignment model, which established in the condition of known the multi-modal O-D demands value, and through solving the lower model, we can get the approximate relationship between O-D demands and road traffic information of different traffic modal. Finally, this paper analyzes and validates the convergence of the model and the calculation results, through a numerical example.