Research On Routing Protocol Based On Social Metrics In Mobile Social Networks

With the high-speed development of the mobile communication technology and network, wireless mobile devices with high-capacity are rapidly spreading on a global scale and promote to the creation of distributed mobile social networks. However, routing in such environments is difficult because of the rapid changes of the social relationship graph’s topology. Efficient data transmission mechanism over the temporal relationship graph has become a hot topic in research area, which is also the problem to be solved in this paper. However, most existing routing protocols can be classified categories with their own drawbacks. Some of the methods are based on duplications, which can cause network resource wasting. Other methods are based on network architecture, and sociability in mobile social networks is ignored. To cure the above problems, in this paper, we propose two routing protocols:High-Efficiency Routing protocol based on Social metrics in mobile social networks (HERS) and an Advanced routing Protocol based on Parameters Optimization in the Weighted mobile social networks (APPOW).In this paper, we firstly analyze the dual characteristics of distributed mobile social network, and combine the social metrics with machine learning algorithms. Then we propose a novel routing protocol-HERS. For introducing HERS, we present the system model of mobile social networks, the social metrics of dual social networks, i.e., egocentric betweenness, similarity and contact strength, the pseudo-code and the concrete implementation process of the algorithm in this paper. We conduct simulations on real trace datasets, i.e., Cambridge dataset and Infocom05 dataset, and the results demonstrate that HERS outperforms the state-of-the-art SimBet Routing in delivering messages and significantly reduces the average end-to-end delay.Though HERS has achieved great performance in efficient data transmission, it still has two limitations. Firstly, HERS algorithm is based on the analysis of binary network, but in real life, most of the networks are weighted networks; Secondly, we adopt common LR approach to learning parameters, but the method in the pairwise comparison approach is not optimal. So next, we solve the above two problems and design APPOW routing protocol. APPOW merely uses the local information including LinkRank, cosine similarity and contact strength to decide the next relay node. APPOW are better than HERS. Firstly, APPOW uses the weighted network structure, which can better confirm to reality. Secondly, the concept of LinkRank is proposed to represent a node’s rank in the short term. The introduction of LinkRank leads to the edge’s value changing from binary to real-valued in the relationship graph. Thirdly, we use a pair-wise learning algorithm to derive the weights of social metrics. As a result, on the one hand, the method to determine weights undergoes a change, from artificial to automatic. On the other hand, the weight values are more reasonable. In order to verify the effectiveness and efficiency of the algorithm, we realized the APPOW algorithm on the Eclipse platform and compared with other related algorithms, the actual running results show that the algorithm can achieve more effective data transmission.To sum up, the two routing protocols, i.e., HERS and APPOW, proposed by us can outperform state-of-the-art methods. Comparing to HERS, APPOW is more effective and more practical according to theoretical analysis and experimental results.