基于用户社群和跨域协作的多目标导向无线资源管理研究

The rapid development and popularity of mobile social networking continue to fuel the explosive demand growth of wireless services. With myriad of wireless broadband network applications permeating all aspects of our daily lives, such growth presents serious challenges to the already scarce bandwidth resources and heavy cellular network load. Smart user terminals, forming social networks based on wireless connections, are capable of supporting content sharing and distributed content storage or caching. Wireless social interactions, in conjunction with advanced wireless technologies including device-to-device (D2D) direct link and multi-hop cooperative communications, present viable new paradigms for wireless content delivery while achieving low energy, high reliability, low delay, and high bandwidth-efficiency without additional load on cellular base stations. This project focuses to develop new analysis and effective tools based on social network awareness and cooperation for multi-layer resource allocation and management with respect to multiple objective functions. The special characteristics of this project stem from its integrated exploitation of traditional physical communication models, together with social interaction and social networking relationship among various wireless terminals, as well as content distribution and popularity, to overcome design challenges of heterogeneous network optimization for improving reliability and spectrum efficiency. Our methodology exploits multi-layer information from physical environment, user social interactive behaviors, application content analysis and tagging, and device constraints to achieve performance improvement in optimization of bandwidth efficiency and power control. We utilize mathematical tools such as convex optimization, statistics, and graph theory to overcome technical challenges encountered when solving multi-objective optimization problems and designing practical low complexity algorithms. Following our technical thread, we shall investigate the problem of user clustering and group resource allocation driven by multi-layer network information to form effective service-oriented social groups. Leveraging user social interaction, we further study and develop key technical solutions to overcome the critical challenges in achieving optimized performance trade-off in terms of bandwidth efficiency, high information secrecy, and low delay for multi-media content delivery. The results and the outcomes of this research project shall provide a theoretical framework and practical guidelines for wireless resource management and optimization in the ongoing evolution toward 5G wireless communications.

移动社交网的迅猛发展深度激发了移动用户通信需求，对紧缺频谱资源及高负荷基站管理提出新挑战。以社群存在的智能终端用户支持内容共享和分布式缓存，结合包含设备间直传和多跳协作的蜂窝D2D通信技术，为用户间实现低能耗、高可靠、低延迟、快速高效的传输且不增加基站负荷提供可能。本项目重点研究蜂窝异构网中基于用户社群和跨域协作的多目标导向无线资源管理，特点在于联合传统物理环境，依据用户间社群关系与交互量化模型，结合内容分布和用户需求等多域资源信息，利用图论和统计论等对频谱高效、能量高效、内容获取高效及传输安全等多目标问题建模，并通过凸优化等数学工具进行优化，围绕该技术路线研究多域信息驱动下业务导向的用户社群形成，进而研究基于用户社群的能量受限下面向频谱效率最大、安全性能最优、内容获取时间最短等差异化目标的性能极限和折中问题，致力于提出一套多目标导向的无线资源管理优化理论与方法，为向5G无缝演进提供支持。

本项目针对当前无线通信网络中通信能耗高企、频谱资源匮乏、无线流量激增、以及重复资源索取造成的资源浪费等问题，立足于社群感知异构无线通信网，研究面向移动社交的用户社群与无线物理环境的特征与联系，提出了一系列的融合物理资源域、社群交互域、内容资源域的跨域资源调度管理理论和方法。项目达到了预期目标，取得了良好的效果。主要成果包括：.1. 业务导向的多目标用户社群动态形成。本项目研究了社交网络中用户社群交互关系的量化表征方法，并面向系统平均获取资源时间最小化、系统吞吐量最大化、能量效率最大化等多重目标，设计了设备间直传、多跳协作等场景下业务导向的多目标用户社群形成方法，从而实现无线资源管理效率、内容资源缓存效率的提高。.2. 面向谱效和能效的用户社群无线资源管理。本项目针对移动用户的物理环境状态以及社群交互关系对建立通信链路的可靠性和鲁棒性的影响，设计了无线链路接纳控制和过滤机制，同时以系统吞吐量、能量效率最大化为目标，以用户公平性、资源高效利用等为原则，研究了用户与用户之间、用户与资源之间的高效匹配方法，实现资源管理和控制的优化。.3. 保障安全传输的用户社群无线资源管理。本项目挖掘社群用户之间的协作关系，在设备间直传、中继协作通信等场景中研究了新型协作安全传输机制，在保障系统安全性的同时，进一步以提升能效和谱效为目标，设计了相应的无线资源优化管控方案，最终实现系统性能的整体改善。.4. 内容获取高效的基于内容缓存的用户社群无线资源管理。本项目针对不同内容资源流行度的特性差别进行建模，联合基站及用户等通信节点，设计了面向异构的内容资源缓存机制，同时基于用户社群的交互关系，研究了面向资源获取可靠性的协作优化缓存，并进一步针对文件流行度的变化趋势，设计了用户主动更新缓存机制，最终实现用户对所需内容的高效获取。