Introducing a new algorithm based on collaborative game theory with the power of learning selfish node records to encourage selfish nodes in mobile social networks

Mobile social networks are a type of delay tolerant network consists of a large number of mobile nodes with social characteristics. The pattern of data transmission and delivery across these networks is due to intermittent, storage, transport and forward connections. Therefore, relay nodes play an important role in these networks. One of the problems with these networks is that there is selfishness in relay nodes. Many algorithms have been proposed to detect and counteract these selfish nodes, but these methods have low detection rates. This paper presents an algorithm for finding selfish nodes and how to persuade them to collaborate based on the game theory. The simulation results show that the proposed method improved the correct detection rate compared to the other methods. First, it applies the Nash equilibrium to identify the selfish nodes and then using the credentials in the relay phase, it forces the selfish nodes to cooperate. The parameters and how the nodes are applied will increase the learning power of the system. The governing rules are applied to the inference system after the learning process. According to the results obtained at the datacenter of the Tabriz University, the improvement has been dramatic, with more than 98% of packets being transferred to the destination in due time and less than 2% of the them has been deleted from the network or their time to live was over. We tried to use the closest nodes to the destination with highest remaining energy to send the data.