Identification And Prediction Of Urban Network Traffic Bottleneck

Traffic congestion has become urgent and important urban problem, and it is often caused by the fixed and dynamic bottleneck of network, if not to take corresponding traffic planning and management measures, it will lead to congestion spread to the entire network, even form a large congestion areas which is hard to control. Therefore, identification the fixed and dynamic bottleneck of networks is the important problems need to be resolved of dynamic traffic planning and formulation the congestion network management and control strategy.In this paper we mainly studies the content of identification (prediction) of the fixed and dynamic bottleneck of networks, which the identification of fixed traffic bottlenecks is mainly used in road network planning and improving planning. The dynamic traffic bottleneck recognition and prediction is mainly used in congestion network management and control, the predictive results can be directly applied to formulate the strategy of management and control.First, analysis the mechanism of the traffic bottlenecks formation and study the influence factors of traffic bottlenecks forming, and based on the formation and evolution regular pattern of traffic bottlenecks determine the temporal and spatial distribution characteristics and existing form of traffic bottleneck.Second, on the analysis of the mechanism of traffic bottlenecks formation, select the important parameters which have been quantified to recognize the fixed traffic bottleneck. It combines with dynamic and static recognition factor by using fuzzy reasoning method and identifies (predict) the fixed traffic bottleneck of network in the planning and improving network.Then, the paper analyzes the causes of dynamic bottleneck and defines its concept, extract the traffic information that identify the dynamic bottleneck by the application of data mining technology. Application the association rules mining method extract the information of bottleneck identifying, it can be used to recognise the real-time congestion areas of network, and puts forward recognition method and recognition process of the traffic bottleneck of network. Finally, considering time and space evolution factors of the traffic flow changes, takes the spatial and temporal parameters as input variables of the prediction model and the next moment section status as prediction object, establishes dynamic bottleneck prediction model based on the BP neural network.