面向D2D通信的社交网络内容分发关键技术研究

With the rapid development of mobile communication technology, the mobile service is showing explosive growth. Networks have undergone a rapid shift from end-to-end communication to emerging content delivery of social networking services. D2D（ communication is a key technology in the fifth mobile communication system, D2D communication based content distribution of social network service is then deemed as the most promising research area for the next generation of broadband mobile communications. Despite of all these advantages, the related research remains in its initial stage, while lots of challenges still reside in the fundamental theory and key technologies that ask for further investigated...This project takes the spatial the spatial and temporal characteristics of social network, user behavior evolution characteristics, transmission quality of service as the research object, and uses graph optimization, multi-objective optimization, game theory as research tool, studies the new theories and new methods of content distribution of social network service via D2D communication. The main research works are as follows. 1) studying the neighborhood nodes dynamic networking theory; 2) analyzing the basic characteristics of the social network services, studying user behavioral cognitive caching technology; 3) studying cooperation incentive strategy of the content distribution and offloading strategy by joint optimization of business characteristics and node characteristics. The research results of the project are expected to lay the theoretical foundation for the study of the content distribution of social network service, and promote the application in the future.

随着移动通信技术的快速发展，移动业务呈现出爆炸性增长的态势。网络的主要承载业务已从传统的互联互通业务向新兴的社交网络业务内容分发转变。利用5G移动通信关键技术之一的D2D通信特性进行社交网络内容分发服务是支撑未来新一代移动通信最具潜力的研究方向。目前处在起步阶段，存在着具有挑战性的基础理论和关键技术问题，有待深入研究。本项目以无线通信网络中节点时空特性、社交网络用户行为演化特征、传输服务质量等作为研究对象，采用图优化、多目标优化、博弈论等方法，研究D2D通信机制下的社交网络内容分发的新理论和新方法。主要内容包括：1）研究移动邻域节点动态组网方法和理论；2）分析社交网络业务的基本特性，研究用户业务行为认知的无线边缘缓存技术；3)研究基于业务特性和节点特性联合优化的协作激励内容分发策略和业务分流策略。项目的展开将有望推动未来社交网络内容分发服务的应用，并为其提供基础理论指导和关键技术支撑。

近年来，移动互联网的业务日益丰富，移动终端数据传输量也随之增加，传统的蜂窝网络面临着频谱资源匮乏，基站负荷大等挑战，极大限制了蜂窝网络的发展。D2D通信技术作为5G的关键技术之一应运而生，用户之间可以通过D2D通信技术直接传输内容，能够减轻基站的负荷、降低能耗、提高网络的传输速度等。本项目将社交网络引入到D2D通信中，将两者有机的结合，分析了基于社交网络的D2D用户配对策略、社交感知缓存策略、资源分配等主要内容。重点包括：1）构造基于移动辅助设备的文件分段D2D缓存模型，加入移动辅助设备帮助D2D用户文件缓存，以此同时，将文件分段或分块进一步提高缓存命中率；2）将物理域与社交域相结合，不仅考虑用户间的物理距离，也考虑用户间的社交关系，同时在计算用户的社交属性时采用真实的数据集，通过深度学习构建推荐系统，获取用户的偏好；3）为了迎合绿色通信，我们采用图论、博弈论、深度学习、凸优化、贪婪算法、强化学习等算法理论为工具设计了最大化能源效率、最小化成本、最大化通信速率、最大化吞吐量等函数。项目所研究内容以理论创新为基础，注重成功的实际应用，具有重要的理论和应用价值。