Dynamic network zoning method based on community detection

As one important step of area-coordination in traffic management systems, the technology of network zoning is indispensable. On this topic, this paper introduces our work to reduce the burden of engineers on this problem. Firstly, inspired by the pros and cons of previous researches, we found three principles of zoning. Then, one model is proposed to measure the relationship between intersections for automatic identification of hierarchical characteristic of traffic network and adapt to the alternating busy-idle traffic environment. Under one certain traffic scene, a higher score of model means that two intersections have closer relationship and the related link has a higher coordination priority. Hence, using this model, we can rank the coordination priority of links in the network. Lastly, considering traffic network is one kind of directed spatial network, one network zoning method based on community detection is detailed to handle the characteristics of traffic network. We use a series of experiments on the traffic network in Tianhe area of Guangzhou to illustrate the effectiveness of this method and can be employed to handle large scale traffic networks.