The Study Of Dynamic Group-finding In Public Bus Networks

At present, the study of the public bus networks is a hotspot issues. The public busnetwork is an important part of mobile network. It is established by the mobile wirelessnetwork and the communication often occurs between the vehicle and vehicle or vehiclewith the roadside. Vehicles can send messages to each other by the public bus networks.For the public bus networks the problem of routing is mainly studied. So they can findthe best method to forward messages. People pay close attention on the social network.Social network as a large and complex network has a very strong community effect. Thestrong community structure can be shown in social network. People have been studied alot of community detect algorithms from many different aspects. Group (community)structure detection presents an insight of the potential organization and functionalproperties, and benefit the data packet propagation in the various networks.Public bus networks are large and complex networks. The character of the publicnetworks is that the rout of buses is fixed following bus’ schedules. So in this thesis weuse this character and find that the average number of neighbors of bus changes in aregular way across the running time. Thus we present a dynamic group-findingalgorithm in public bus networks. The proposed algorithm includes two phases. Tobegin with, time is divided into some time interval based on the average number ofneighbors of bus, and the time interval is extracted from the bus mobility trace. A groupdetection algorithm based on spectral clustering is then proposed for each time interval.Finally, we use the Newman’s modularity to evaluate the quality of communitystructure divided with our algorithm. A large number of simulation experiments havebeen carried over a realistic bus trace data. The simulation results show that ourdynamic group-finding algorithm well capture group evolution in public bus networks.