一种低开销的机会网络社区检测算法  

A Low-overhead Community Detect Algorithm in Opportunistic Networks

作　　者：任智[1] 黄希凯 谭永银 

机构地区：[1]重庆邮电大学移动通信技术重庆市重点实验室,重庆400065

出　　处：《微电子学与计算机》2016年第9期60-63,69,共5页Microelectronics & Computer

基　　金：国家自然科学基金项目(61379159);长江学者和创新团队发展计划基金资助项目(IRT1299)

摘　　要：针对CDTFS(Community Detection algorithm based on Threshold of Familiar Set)算法在社区检测阶段中存在通信冗余开销的问题,提出了一种低开销机会网络社区检测算法—NSDMDT(Network Structure Detection Mechanism based on Dynamic Threshold).NSDMDT算法优化了新节点加入朋友集合时子图的发送顺序和社区关系子图更新机制,减少社区检测时网络中的通信开销,进一步提高了网络的性能.理论分析和仿真结果表明在节点通信半径不同、网络中节点个数不同以及节点移动速率不同的三种情况下,NSDMDT算法网络开销均要低于CDTFS算法,同时在与朋友节点的累积相遇持续时间所占比例上也优于CDTFS算法.To solve the existing problem of redundant communication overhead of the community detection in the CDTFS algorithm, and a novel routing algorithm called NSDMDT （Network Structure Detection Mechanism based on Dynamic Threshold） is proposed. The algorithm optimizes the sending order when a new node ioining familiar set and the mechanism of updating community relationship sub graph, reduces the network overhead in the network community detection and improves the performance of the network. Theoretical analysis and simulation results show that NSDMDT algorithm outperforms an existing CDTFS algorithm in terms of control overhead when the communication radius is not the same or when the number of nodes in the network is different, and the ratio of cumulative duration with friend nodes.

关 键 词：机会网络 社区检测 子图 开销 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]