动态连通网络时间演化特性及传输机制研究

Many mobile ad hoc networks, such as Delay/Disruption Tolerant Networks and Opportunistic Networks, have the essential feature of partial and time-evolving connectivity. In these dynamically connected networks (DCNs), the assumption that end-to-end communication routes exist between nodes at any instant of time does not hold, which makes the connectivity properties varied dramatically and the traditional mobile ad hoc routing and transmission mechanisms unsuitable as a result. By experimentally examining the topological features of DCN and modelling nodes mobility, the main aim of the project is to theoretically analyze the time evolving process and network connectivity of DCN to establish designing principles for transmission mechanisms of routing and forwarding, caching management, and data interchange. .Following the research routine of topology characterizing, network modelling, behavior analyzing, and transmission designing, the project focuses on the key problem of how the time evolving properties of DCN affect the performance of transmission mechanisms, appealing to the methods of statistical analysis, graph-based calculus, stochastic process, and optical decision. Specifically, the topological elements which affect the connectivity of DCN are to be identified, the network evolving laws in terms of time-scale are to be revealed, which will provide the guidelines for the designing of routing protocols, caching management strategies, and data interchange methods in DCNs. The expected research results can be potentially applied in Wireless Sensor Networks, Vehicular ad hoc Networks, and Mobility-assisted Social Networks, promoting the development and application of dynamically connected networks.

动态连通并随时间演化，是容迟/容断网络、机会网络、移动自组网的共性特征。在这类网络中，节点之间并不总是存在实时连通的端到端路径，针对传统网络的连通性质分析和路由传输机制不再适用。本项目拟采用实证分析与理论研究、仿真实验与原型实现并举的方法，对动态连通网络的时间演化特性及传输机制进行研究，重点从该类网络的拓扑特征分析与节点移动建模入手，对其时间演化过程及网络连通性进行分析和度量，进而设计实现与之相适应的路由转发机制、缓存管理策略与数据交换方法。拟沿着"研究拓扑特征、建立网络模型、分析网络性质、设计传输机制"的路线，利用统计分析、图论推演、随机过程、最优决策等理论与方法，围绕动态连通网络的时间演化特性对数据传输有何影响这一核心问题开展研究，有望在发现新的网络结构要素、理解网络演化内在规律、设计优化传输机制等方面有所突破。相关研究成果可广泛应用于传感网、车载网、移动社交网等多种网络应用环境。

本项目面向非实时连通的动态网络，对其时间演化机理、网络连通性质及其影响要素等进行建模与分析，并围绕时间演化特性对网络数据传输有何影响这一核心问题开展研究。项目的主要工作包括：.（1）利用CRAWDAD第三方公开数据集、上海出租车数据集和依托单位校园网数据集，对动态连通网络进行了充分的实证性分析，提出了考虑节点社区特性、自身偏好等因素的节点移动模型，分析了节点移动对动态网络结构性质的影响。.（2）从时间演化的视角对动态网络的连通性进行了深入的研究，在时变图基础上提出一种关键行程演化图理论模型，刻画连续时间上节点间关键行程的演化，考察节点间的时间距离、可达中心性等连通特性，进而定义了网络密度、期望速率等衡量网络传输性能的连通性度量。.（3）面向移动社交网、传感网、车载网、无线Mesh网等不同的应用场景，提出了多种涵盖功率控制、拓扑管理、信道接入、路由决策等内容的传输控制机制，设计了分别基于连通预测、社区划分及运动轨迹分析的路由转发算法和与之相适应的缓存管理与数据交互机制。 .（4）设计了CjegSim等仿真工具并开放源码，开展了大量的仿真实验；基于Android平台开发了适用于动态连通网络的移动内容分发原型系统，并在国家863计划未来网络研究项目中得到初步应用。