基于朋友关系变迁演化的社区划分算法

机会网络中的节点大多是由人携带的具有短距离无线通信能力的设备，因此节点具有人的社会属性，机会网络也具有社区特性。而现有的社区划分算法多数来源于社会网络，不能直接用于节点稀疏、网络拓补结构不断不化、节点的计算和存储资源有限的机会网络中。本文通过提出节点的朋友关系，利用节点自身的相遇历史信息，将网络划分为不同的朋友社区。在两种不同数据集下的实验结果表明朋友社区的划分可有效减低路由开销，同时保持较高的传输成功率。

Community division algorithm based on the evolution of friendships

Nodes that have short-distance wireless communication capabilities in the opportunistic network are mostly carried by human. Therefore, the nodes have the social properties and the opportunistic networks have community features. Some existing algorithms for community structure detecting are mostly from social networks, which cannot be directly used in opportunistic networks as nodes in those networks have limited storage and compute resources. Besides, the number of nodes are sparse and the topological structures are changing all the times. In this article, we proposed the friends relationship metric between nodes. Then, we use it to divide the networks into different friends communities by using the node''s encounter histories itself. Experiments are conducted under two different sets of data, and the experimental results showed that routing overhead has reduced effectively and the transmission rate still keep a high value by establishing friends communities.