无线泛在环境下网络资源与业务匹配理论研究

In wireless ubiquitous environment, with the emergence of new applications and increase of application requirements, network resources become increasingly scarce. And, it is more and more difficult for network resources to satisfy application requirements. On the other hand, there is a serious imbalance for application requirements in time, space or heterogeneous networks. Therefore, it is very challenging and valuable to explore the feasible way that requirement distribution match resource distribution (i.e. limited resources can satisfy ubiquitous requirements). So, this project propose a novel multi-domain restructuring idea of network resources to solve the above problem. The main contents of this project are as follows. (I) Prediction theory of wireless application requirements: concept of user basic cluster will be constructed, the corresponding behavior law as well as behavior motive will be researched, and then user oriented hierarchical prediction method will be developed in this part. (II) Multi-domain restructuring theory of wireless network resources: single-dimension restructuring strategies in time domain, space domain or network domain are designed respectively, complete multi-domain restructuring space will be constructed subsequently, and computation methods of wireless transmission capacity will be researched deeply in this part. (III) Optimal matching theory and its implementation: matching efficiency metrics will be formulated, bilevel programming based dynamic game optimization model will be researched carefully, the corresponding fast heuristic algorithm will be designed, and logic implementation architecture for matching theory will be developed finally.

未来无线泛在环境下，随着新型业务的不断涌现以及业务需求的迅猛增长，网络资源变得日益紧缺，资源无法承载业务的问题亦日趋严重。另一方面，业务量在时间、空间和异构网络上均呈现出非常不均衡的分布规律。因此，探索业务分布匹配于资源分布的可行途径，以有限的资源满足泛在的业务需求，将极具挑战而富有价值。为此，本课题创新性地提出了网络资源的多域重构思想，主要内容包括：无线业务的预测理论，构建用户基本群集的概念，挖掘相应的业务行为规律及其产生动因，进而开发面向用户的业务层次化预测方法；无线资源的多域重构理论，分别设计基于时域、空域和网络域的单维重构策略，进而构建完备的多域重构立体空间，并探索相应的无线传输容量计算方法；资源与业务最优匹配理论及其实现方案，建立匹配效率指标的数学表达式，着重研究基于双层规划的动态博弈资源优化模型，进而设计相应的启发式快速求解算法，以此为基础开发整体理论的逻辑实现架构。

本项目面向未来无线泛在环境下资源难以承载业务的科学问题，综合运用多种理论和技术，深入探索相应的根本解决途径。为此，本项目从无线业务预测、无线资源重构和资源与业务的最优匹配等三个层面展开深入研究。首先，分别基于Apriori理论和Markov模型，提出了若干用户业务行为的分析与预测算法，构建了无线业务的预测理论，从而为网络资源的精准分配提供了必要的环境信息。其次，从核心算法、关键技术和支撑模块等方面研究网络资源的多域重构，提出了一系列新颖有效的算法，构建了无线资源的重构理论，从而为网络资源的精准分配提供了可行的实现途径。再次，从路由优化、能量优化、协同优化和资源匹配等方面研究无线资源的最优分配，提出了一系列新颖有效的算法，构建了资源与业务的最优匹配理论，从而实现了“从根本上解决资源与业务的匹配问题”的预期目标。此外，基于终端协同思想，设计并开发出三套功能完备且性能优异的产品系统。本项目已经较好地完成了各项预定的研究目标和考核要求，为推动“泛在的无线接入实现泛在的服务需求”走向现实做出了应有的贡献。