基于严格势博弈的干扰对齐  

Interference Alignment Based on Exact Potential Game

作　　者：章扬[1,2] 周正[1] 石磊[3] 李斌[1] 李德建[1] 

机构地区：[1]泛网无线通信教育部重点实验室(北京邮电大学),北京100876 [2]中国石油大学(华东)计算机与通信工程学院,山东青岛266580 [3]中国电子科技集团公司第五十四研究所,石家庄050081

出　　处：《北京邮电大学学报》2013年第2期50-54,共5页Journal of Beijing University of Posts and Telecommunications

基　　金：国家自然科学基金项目(60972079;61271180;60902046);中央高校基本科研业务费专项资金项目(2012RC0103);韩国知识经济部仁荷大学ITRC基金项目(NIPA-2011-C1090-1111-0007)

摘　　要：在经典多输入多输出(MIMO)干扰信道干扰对齐算法中,单纯最大化用户期望信号势必对其他用户形成干扰,从而限制整体信道容量的提升.从多对干扰链路构成一个博弈群体的角度出发,提出了一种基于严格势博弈的干扰对齐算法;构建了基于博弈论的干扰对齐算法数学模型,通过构造有界势函数证明该博弈为有界严格势博弈,能通过有限次迭代收敛到ε纳什均衡点;讨论了代价因子对算法性能的影响.实验结果表明,相对于经典的min-INL和max-SINR 2种干扰对齐算法,新算法具备更优的性能,可显著提高复杂MIMO干扰信道中的多用户信道容量.Maximizing the user＇s expected signal may cause interferences to other users in current inter- ference alignment algorithms for multiple-input multiple-output （MIMO） interference channels. It re- duces the total channel capacities. Considering the fact that multiple interfering links consist of a game group, a new interference alignment algorithm is presented based on the exact potential game theory. The algorithm designs a mathematic model for interference alignment algorithm based on game theory, and proves that the proposed game is a hounded exact potential game by constructing a bounded potential function which converges to a ε-Nash equilibrium through finite interactions, and also the impact of cost factor on the performance of algorithms is discussed. Simulation shows that the new algorithm is with bet- ter performance than the existing minimum-interference （min-INL） and maximum-signal interference noise ratio （max-SINR） interference alignment algorithms and can significantly improve the multi-user network＇s capacity.

关 键 词：多输入多输出 干扰对齐 严格势博弈 ε纳什均衡点 弱有限改进特性 

分 类 号：TN911.7[电子电信—通信与信息系统]