【摘要】 Ad Hoc网络中由于节点移动、计算限制等因素导致其拓扑结构多变,需要频繁重建路由,直接在原始拓扑上执行数据转发算法显得极其困难,影响分组转发可靠性和效率。利用社会网络和行为规律的知识,在构建连通子图的基础上,通过分析、预测节点联系的时间、联系时间间隔和空间位置等方法,构建了一个联系时空的预测转发模型。通过仿真实验对该模型进行验证,结果表明:该模型能有效地预测并建立转发路由,在分组转发成功率、端到端平均时延以及平均路由开销等性能上较现有模型有较大的提升。更多还原

【Abstract】 In ad hoc networks,the routing table needs frequent reconstruction due to quick-changed topology caused by node’s movements and computation limits. The reliability and efficiency of packet forwarding will be affected since it is quite difficult to execute data forwarding algorithm directly on the original topology. With the knowledge of social networks and behaviour rules,we first construct a connectivity subgraph,and then build up a prediction forwarding model of contact spacetime by analysing and forecasting the node’s contact time,contact time interval and spatial location as well. Verification of the model is done through stimulation experiment,results show that this model can effectively forecast and establish forwarding route,and has greater improvement,compared with the existing model,in packet forwarding success rate,end-to-end performance on the average delay and average routing overhead.更多还原