面向分层异构网络的资源分配：一种稳健分层博弈学习方案

Resource Allocation for Heterogeneous Wireless Networks: A Robust Layered Game Learning Solutions

该文研究了信道状态不确定条件下分层异构微蜂窝网络中的无线资源分配优化问题。首先引入信道不确定模型描述无线信道的随机动态性,并将该问题建模为考虑信道不确定度的双层鲁棒斯坦伯格博弈;然后给出了该博弈的均衡点分析;最后提出了一种分布式改进型分层Q学习方案以实现宏基站和微基站的均衡策略搜索。理论分析和仿真表明,所提出的分层博弈模型可以有效抑制由于信道状态不确定引起的收益下降。所采用的学习方案较传统Q学习方案收敛速度明显加快,更加适用于短时快变的通信环境。

This paper investigates a resource allocation scheme in heterogeneous wireless small cell networks with imperfect Channel State Information （CSI）. In this work, the math expression for the stochastic dynamic uncertainty in CSI is proposed for model analysis and the robust Stackelberg game model with various interference power constraints is established firstly. Then, the Stackelberg game Equilibrium （SE） is obtained and analyzed. Lastly, an improved hierarchical Q-learning algorithm is also given to search the Stackelberg equilibrium strategies of ma^ro-cell base station and small-cell base station. Both theoretical analysis and simulation results verify the proposed scheme can effectively restrain declining revenue due to incomplete CSI and the proposed algorithms can improves the convergence rate, especially applicable to the fast varying communication environment.