Dynamic Division of Road Network Sub Areas Integrating the Importance and Similarity of Intersections

Aiming at the high complexity of urban traffic road network and insufficient traditional sub-district division methods, based on the improved Pearson Correlation Coefficient, a combined feature index to measure the similarity of time series of traffic flow features was constructed; based on the improved Louvain calculation, a dynamic division method of regional road network is designed. Firstly, the Pearson correlation coefficient is introduced to construct the traffic flow feature similarity model, and it is combined with the importance of the intersection to propose an improved intersection correlation model(Ipearson model). Secondly, modularity Q, control efficiency and algorithm running time are used as evaluation criteria and take Ipearson model as the edge weight, an improved Louvain algorithm(Prolouvain algorithm) is proposed to improve the operation efficiency of the algorithm and make it suitable for the sub-division of large-scale traffic network. The simulation results show that the proposed road network sub-division method can effectively combine the characteristics of traffic flow, which fully verifies the validity and accuracy of the model proposed in this study.