ISSN 1000-1239 CN 11-1777/TP

计算机研究与发展 ›› 2016, Vol. 53 ›› Issue (6): 1365-1375.doi: 10.7544/issn1000-1239.2016.20148412

• 信息处理 • 上一篇    下一篇



  1. (哈尔滨工程大学计算机科学与技术学院 哈尔滨 150001) (
  • 出版日期: 2016-06-01
  • 基金资助: 

Message Dissemination Based on Interest Matching for Opportunistic Social Networks

Zhang Yushu, Wang Huiqiang, Feng Guangsheng, Lü Hongwu   

  1. (College of Computer Science and Technology, Harbin Engineering University, Harbin 150001)
  • Online: 2016-06-01

摘要: 机会社会网络(opportunistic social networks)能够利用节点移动创造的相遇机会,在缺乏持续端到端连接的网络中,为用户提供稳定的消息分发途径,但在消息分发效率以及用户体验方面存在不足.为提高消息分发系统的性能、改善网络用户体验,提出一种基于节点兴趣匹配的机会社会网络分发机制.通过引入混合结构的机会社会网络分发系统解决网络拓扑信息获取不全与节点计算能力不足的问题;从节点行为规律与兴趣爱好2方面对网络进行分析,并提出一种用于复杂关系数据分析的联合聚类方法;针对用户需求,设计消息属性与节点兴趣匹配优先的消息分发策略.仿真结果表明,该机制能够在投递率、投递时延、缓存占用率等方面提升网络性能,且具有较高的分发效率、覆盖率与兴趣匹配度.

关键词: 机会网络, 消息分发, 联合聚类, 兴趣匹配, 拥塞控制

Abstract: Opportunistic social networks can provide message dissemination service through encounters created by nodes’ movement in intermittently connected or completely disconnected wireless networks, but there are still some shortcomings in message dissemination efficiency and user experience. In order to promote the performance of dissemination system and improve the user experience, an opportunistic social network message dissemination mechanism based on node interest matching is proposed. A kind of opportunistic social networks with hybrid architecture is introduced to avoid problems such as incomplete information of network topology and low computing ability of nodes. Considering the fact that node behaviors and interests can both affect the performance of dissemination, the analytical method designed takes both of them as reference factors, and an improved co-clustering algorithm for data with complex relationship is proposed. In addition, the interest matching rate for nodes is used to show what percentage of buffer is occupied by the useful messages to user, and a novel message dissemination strategy is proposed to raise the rate. Simulation results show that the proposed scheme can both improve the capability of opportunistic social networks in terms of delivery rate, latency and buffer usage, and promote the performance of message dissemination systems in terms of efficiency, coverage rate and interest matching rate.

Key words: opportunistic network, message dissemination, co-clustering, interest matching, congestion control