摘要： 为提高网络的利用率，解决DT-MSN中网络部署问题，运用子图的扩展理论，给出一种判定是否存在和全网同构的样本网络方法。在存在样本网络的前提下，用最小生成树原理提出一种基于节点间最大连接次数的节点分簇算法。该算法可以构建一种样本网络，在投递的数据流上和全网同构。构建一个样本网络场景，并和全网作实验对比分析。结果表明，构建样本网络模型可以代替全网做性能的预评估，能为部署大型网络场景提供可靠的性能评估报告，具有较好的实际应用价值。

关键词: 容延迟移动传感器网络, 最小生成树, 子图扩展, 样本网络场景, 分簇算法, 性能预评估

Abstract: To improve the utilization of Delay Tolerate Mobile Sensor Network(DT-MSN) and to deal with the networks deployment issues, this paper proposes a decision condition about whether the global network has a sample network of isomorphic by applying the sub-graph expansion under a specific scenario, and then proposes a algorithm about clustering the sensors that are connected with maximum times by using minimum spanning tree on condition of existing sample networks. The method can build an isomorphic network between the sample network and global network at data flow. Simulation is used to prove this scenario existing the isomorphic networks, and establishes a sample network. Result shows the sample network scenario can replace the global network to assess network’s performance. So it can offer reliable evaluation reports before setting a large scenario, and have an important practical application.

Key words: Delay Tolerate Mobile Sensor Network(DT-MSN), minimum spanning tree, sub-graph expansion, sample network scenario, clustering algorithm, performance pre-evaluation

中图分类号: 

• TP393