Pocket switched Network (PSN) is an emerging research magnet that has been attracting researchers. This paper provides an introduction to PSN and covers some of its basic features and challenges. Continuous connectivi...Pocket switched Network (PSN) is an emerging research magnet that has been attracting researchers. This paper provides an introduction to PSN and covers some of its basic features and challenges. Continuous connectivity is difficult in infrastructure independent communication. In this paper, we will see how PSN provides an effective alternative. We also explain some routing protocols that can be incorporated for effective data forwarding. This paper also discusses possible applications and areas where PSN can be effectively used and some ideas that may foster future research in PSN routing.展开更多
为了探索便携交换网络的演化规律,研究其网络行为预测中的链路预测问题。便携交换网络具有节点移动性、节点间间歇性连接、高延迟等特点,其链路预测面临的挑战是节点相遇的机会性和拓扑的时变性,获得其高质量链路预测的关键是如何较全...为了探索便携交换网络的演化规律,研究其网络行为预测中的链路预测问题。便携交换网络具有节点移动性、节点间间歇性连接、高延迟等特点,其链路预测面临的挑战是节点相遇的机会性和拓扑的时变性,获得其高质量链路预测的关键是如何较全面地获取节点的属性。作者提出基于学习自动机和萤火虫算法的链路预测方法(link prediction approach for pocket switched network based on firefly algorithm,FA-LP)。采用学习自动机对节点进行自适应聚类,完成网络的社区划分;定义社区属性影响系数和移动行为影响系数,构建反映便携交换网络社区属性、节点移动性和节点间间歇性连接的相似性指标;将该指标与CN、RA、AA等指标融合,得到便携交换网络的相似性指标向量;借助差分整合移动平均自回归模型的时间序列分析能力,提取相似性指标向量序列的演化规律;采用萤火虫算法优化所构建的二分类器,预测节点对下一时刻的连接状态。INFOCOM2006和MIT两个真实数据下的实验结果表明,与受限玻尔兹曼机、弱评估器等方法相比,FA-LP具有更高的准确率和更好的稳定性。展开更多
文摘Pocket switched Network (PSN) is an emerging research magnet that has been attracting researchers. This paper provides an introduction to PSN and covers some of its basic features and challenges. Continuous connectivity is difficult in infrastructure independent communication. In this paper, we will see how PSN provides an effective alternative. We also explain some routing protocols that can be incorporated for effective data forwarding. This paper also discusses possible applications and areas where PSN can be effectively used and some ideas that may foster future research in PSN routing.
文摘为了探索便携交换网络的演化规律,研究其网络行为预测中的链路预测问题。便携交换网络具有节点移动性、节点间间歇性连接、高延迟等特点,其链路预测面临的挑战是节点相遇的机会性和拓扑的时变性,获得其高质量链路预测的关键是如何较全面地获取节点的属性。作者提出基于学习自动机和萤火虫算法的链路预测方法(link prediction approach for pocket switched network based on firefly algorithm,FA-LP)。采用学习自动机对节点进行自适应聚类,完成网络的社区划分;定义社区属性影响系数和移动行为影响系数,构建反映便携交换网络社区属性、节点移动性和节点间间歇性连接的相似性指标;将该指标与CN、RA、AA等指标融合,得到便携交换网络的相似性指标向量;借助差分整合移动平均自回归模型的时间序列分析能力,提取相似性指标向量序列的演化规律;采用萤火虫算法优化所构建的二分类器,预测节点对下一时刻的连接状态。INFOCOM2006和MIT两个真实数据下的实验结果表明,与受限玻尔兹曼机、弱评估器等方法相比,FA-LP具有更高的准确率和更好的稳定性。
