| Decentralized mobile social network (D-MSN) is one of the hot spots in mobilenetwork applications. As a critical issue in D-MSN, data transmission gets a lot of attention.For traditional mobile ad hoc network routings depending on the end-to-end connections,they are no longer adaptive to this new network. Although some routing protocolsproposed for the delay tolerant network can be applied to some extent, they didn’t take fulladvantage of nodes’ social attributes that makes them perform poorly in data transmissiondelay. Recently, a few routing protocols take nodes’ social characteristics intoconsideration and improve routing efficiency. However, most of them ignored nodes’selfishness. All the above reasons limit the applications of D-MSN. In this dissertation, weanalyze and present the entire routing scheme by the ant colony optimization theory etc.and combining the features of D-MSN.Firstly, we analyzed the classic routing protocols that applicable for the D-MSN. Thenwe extracted the key factors in the data transmission process and three general routingmodels were grouped. On this basis, the distribution functions of the data transmissiondelay on each model were analyzed by the theory of probability. After the analysis of theproblem of data transmission in such network environment, we proposed an ant colonyoptimization based routing protocol. This protocol is inspired by the real ants finding food,and the optimal path between two nodes can be explored by sending artificial ants. In theexploring process, we make the artificial ant explore and update pheromone simultaneously,which solving the requirement of the end-to-end connections in the original ant-basedrouting protocols. In the data transmission process, the nodes selects the next hop is notonly dependent on the result of artificial ants explore, but also consider the social attributesof the next hop, which optimizes the data transmission efficiency to some extent. Inaddition, we proposed a repeated game based routing incentive mechanism in this networkenvironment. To detect the selfish behavior of nodes, the central bank is introduced. Thedata forwarding process is abstracted as an infinitely repeated game. By introducing punishment mechanism to selfish behavior, we analyzed the model and get the conditionsfor nodes’ cooperation. We proved that the selfish nodes will converge to cooperationnodes under this condition.Simulation results show that the D-MSN delay assessment model presented in thisdissertation is accurate. What’s more, the proposed routing protocol and incentivemechanism can improve the efficiency of the data transmission and suppress the nodes’selfish behavior. The above work optimized the D-MSN environment and improved theexperience of the mobile users, which has positive significance for the further developmentof the D-MSN applications. |
PDF Full Text Request