Coalition Formation For Roadside Units And Relay Vehicles Selection In Vehicular Networks

With the significant increase of vehicle wireless mobile devices, Vehicular Networks(VANET) appears,which can better organize and manage wireless network consisting of vehicles. VANET aims at improving traffic safety conditions, enhancing transport efficiency and facilitating people travelling. Increasing researchers focus on VANET in industry and academia as VANET has been a hotspot of ITS. The information transfer mechanism in VANET is important to supporting various kinds of applications that is related to traffic safety and entertainment.Achieving efficient information transfer mechanism is a huge challenge in VANET, considering fast moving vehicles and data redundancy associated with dense traffic. Waild, et al proposed a novel cooperation protocol depending on the content sharing possibilities of the underlying communications between vehicles in the opposite direction. Then Coalition Formation Games(CGS) is proposed to dividing RSUs into different coalitions for the coordination of RSUs. But a member must contribute to each member belonging to the same coalition. And the coalition formation based on graph partition theory results in the cooperation among RSUs is not flexible. In addition, the cooperation protocol is inefficient on one-way road because the content sharing possibilities of the underlying communications between vehicles in the opposite direction is low.Aiming at problems above, this paper proposes a Coalition Formation based on Genetic Algorithm(CFGA) strategy by analyzing characteristics of the cooperation protocol. A diversity of cooperation among RSUs can be represented with appropriate genetic representation. Qualified solutions that accord with road topology can be generated by initialization, Selection and Crossover. Local searching of solution space can be implemented by margin gene mutation. Considering the limitation of the cooperation protocol on one-way road, this paper proposes the strategy that vehicles share information backwards and selects (Multipoint Relay)MPR vehicles based on set cover to reducing redundancy forwarding data.