基于鲁棒非合作博弈的飞蜂窝网络资源分配与干扰管理

项目名称： 基于鲁棒非合作博弈的飞蜂窝网络资源分配与干扰管理

项目编号： No.61473247

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 刘志新

作者单位： 燕山大学

项目金额： 83万元

中文摘要： 飞蜂窝网络是提高频谱复用效率，解决室内无线信号有效覆盖的关键技术。在共享信道工作模式下，家庭用户及宏用户之间固有的干扰关系使得频谱分配和干扰管理问题尤为突出。而信道增益、用户接入、信息获取等因素的不确定性直接影响用户服务质量。在复杂动态环境下，如何以分布式算法实现不同类型用户利益及网络资源利用的最大化是本项目研究的出发点。首先分析无线网络环境不确定因素的参数化描述，利用概率统计与分析方法刻画影响信道增益、传输速率的主要因素，建立概率约束优化模型。其次考虑网络用户自私行为及环境不确定性，构造综合频谱分配与干扰管理的非合作鲁棒博弈问题，以竞争、议价等机制诱导用户趋于平衡，实现整体资源的优化配置。进一步研究网络用户多入多出、多主多从情景下的鲁棒约束功率控制方法，针对动态接入的突发干扰，建立接纳控制策略。最后考虑算法实现过程中的边界约束和异步时延，基于饱和控制稳定性分析研究迭代算法的一致收敛性。

中文关键词： 网络化系统；飞蜂窝网络；资源管理；鲁棒优化；博弈论

英文摘要： Femtocell network is one of key techniques to improve the efficiency of frequency and solve the problem of coverage indoor. The frequency allocation and reference management are increasing important when the femtocell and macrocell coexist and share the same channels. The interference among users can not be avoided in this scenario. Besides, the uncertainties of network environment, channel state and access of users will affect the quality of service. We focus on how to develop the distributed algorithm which can optimize the utilization of different users and maximize the whole resources of network. First, the uncertain factors of wireless network which are related to channel gain and transmitting rate are described by probability theory and statistic method. And the optimization model with probability constraints will be formulated. Second, taking into consideration of selfish of users and uncertainties of network environment, the non-cooperative game is used to model the problem of frequency allocation and interference management among different kinds of users. To realize the optimal configuration of the whole resources, the competition and bargaining are introduced to induce the users to tend to the equilibrium. Next, the networks with multi macrocell users and multi femtocell users will be considered, and the strategies of power control with robust constraints are to develop. For the burst of interference caused by dynamic accessing, admission control strategies based on supermodular games will be developed. Finally, considering the effect of boundary constraint of the game formulation, the uniform convergence of the distributed iteration algorithm will be addressed based on the saturation control theory.

英文关键词： Networking System;Femtocell Networks;Resource Management;Robust Optimization;Game Theory