Spectrum sharing deployment of femtocells brings interferences which dramatically degrade network performance. Hence, interference control is a crucial challenge for femtocell networks. In this paper, we propose a power optimization approach for 5G femtocell networks consisting of macrocell and underlying femtocells to manage the interference. Firstly, we formulate the problem based on a non-cooperative game to analyze the competition among the users to access shared spectrum. We then design a pricing mechanism in the utility function to guarantee quality of service (QoS) requirements of macro users. The mechanism lets the macro users experience lower interference and achieve the minimum required data rate. As a result, QoS requirements of both macro and femto users are fulfilled in a non-cooperative manner. We also design a minimax decision rule to optimize the worst-case performance and find an optimal transmission power for each user. By adjusting the optimal power for each user, the maximum aggregate interference is minimized, and the network throughput is maximized. Finally, we develop an iterative learning- based algorithm to implement the proposed scheme and achieve the game equilibrium. Theoretical analysis and simulation results verifies the effectiveness of the proposed mechanism in terms of throughput maximization, QoS assurance and interference mitigation.

毫微微蜂窝的频谱共享部署带来了显着降低网络性能的干扰。因此，干扰控制是毫微微蜂窝网络的关键挑战。在本文中，我们为由宏蜂窝和底层毫微微蜂窝组成的 5G 毫微微蜂窝网络提出了一种功率优化方法来管理干扰。首先，我们基于非合作博弈制定问题来分析用户之间访问共享频谱的竞争。然后，我们在效用函数中设计了一种定价机制，以保证宏用户的服务质量 (QoS) 要求。该机制让宏用户体验到较低的干扰并实现所需的最低数据速率。结果，以非合作方式满足宏用户和毫微微用户的QoS要求。我们还设计了一个 minimax 决策规则来优化最坏情况下的性能并为每个用户找到最佳传输功率。通过为每个用户调整最优功率，最大聚合干扰最小化，网络吞吐量最大化。最后，我们开发了一种基于迭代学习的算法来实现所提出的方案并实现博弈均衡。理论分析和仿真结果验证了所提出机制在吞吐量最大化、QoS保证和干扰抑制方面的有效性。我们开发了一种基于迭代学习的算法来实现所提出的方案并实现博弈均衡。理论分析和仿真结果验证了所提出机制在吞吐量最大化、QoS保证和干扰抑制方面的有效性。我们开发了一种基于迭代学习的算法来实现所提出的方案并实现博弈均衡。理论分析和仿真结果验证了所提出机制在吞吐量最大化、QoS保证和干扰抑制方面的有效性。