Research On Context-Aware Routing Algorithms In Distributed Moblie Social Networks

As smart mobile devices widely used in people's daily life, mobile social networks have come to attract significant attention in academia and industry recently. At present, smart mobile devices which integrate a variety of sensors can access internal and external sources of context information. So it is show a good development opportunity for context-aware computing. Because mobile devices are usually carried by people, there are more and more close relationship between people and devices. People's social relations and their mobile behavior are called social context. Social context reflects the characteristics of the mobile node in some ways and affects the network node movement patterns significantly. Consequently, the context information (especially social context information) can be exploited to design efficient networking protocols in pervasive mobile environment.However, the efficient and intelligent routing for distributed mobile social networks is an important but highly challenging problem, because of characteristics such as limited resources, intermittent connection and distributed environments, etc. The limited energy of mobile node seriously affects the performance of networks. But through the use of context information related to energy in the network, we can effectively improve energy efficiency of nodes and prolong the lifetime of the network. The network connections are often intermittent or not available in sparse mobile social network, and it is difficult to guarantee end-to-end communication. Data transmission between networks nodes rely mainly on accidental encounter opportunity for nodes, so the way of storage-carry-forward is used to forwarding data. One of the key factors for data transmission in this network is how to predict encounter between nodes. Fortunately, social context information of networks imply the contact regularity and mobile pattern. This offers a new way to solve the routing problem in mobile social networks.The routing problem of distributed mobile social network is studied deeply in this thesis. Based on multidisciplinary crossover concept and relevant theoretical knowledge, the methods of social network analysis, machine learning and context-aware computing are used to study context-aware routing in different network environment. This research will not only promote the development of efficient and intelligent routing in distributed mobile social networks, but also help the development of next generation people-based intelligent networks for both academic sense and applied value. In summary, the major contributions of this dissertation can be summarized as follows:1. Energy saving is one of the important research content of mobile social networks. However, energy saving often lead to a decline in the performance of the network. It is difficult to balance energy efficiency and network performance. To address this problem, a novel context-aware energy-efficient routing (CAER) was proposed. By adopting the idea of cross-layer design in the scheme, network topology information, application-aware information in application layer and link-stability were comprehensive utilized to make routing decisions in network layer. Under the condition of guarantee the performance of routing protocols, we tried to improve energy efficiency of the networks as far as possible. The evaluation show that, comparing to other state-of-art algorithms, CAER performs more effectively on energy efficiency and lifetime of the network, at the same time it reduces the delay overhead of the network.2. Often there is no end-to-end path in sparse mobile social networks. The multidimensional context matching forwarding scheme (MCMF) was proposed by making full use of a variety of context information. Message carrier first obtained its neighbors at the present moment. Then social similarity matching was used to search relay candidates set in neighbors, and the discrete-time semi-Markov prediction model was used to determine the best relay node. At last the efficient data forwarding strategy was designed. Simulation results show that MCMF performs higher forwarding performance.3. Mobile social networks are a collection of social relations, and social relations and social activity of nodes affect profoundly the movement mode of the node. In view of person's daily activities and encounter mode, Social Context-aware for Opportunistic Routing (SCOR) was proposed. In this novel protocol, social context information of the network was exploited to predict the mobility patterns of nodes based on the back-propagation neural networks model. The routing scheme considered both the spatial and the temporal dimensions. If the recipient was present in the same connected region of the network as the sender, the message was delivered by synchronous method, or else the message was delivered by asynchronous method. The evaluation analysis and the simulation results indicate that for the social context-based routing algorithms in opportunistic networks, SCOR solution outperforms other state-of-art routing solutions thanks to its ability to maximize the delivery ratio and to minimize the network overhead.