| 基于深度强化学习的超密集网络中多用户上行功率控制方法 |
| |
| 引用本文： | 毛 晋,,熊 轲,,位 宁,,张 煜,张锐晨,.基于深度强化学习的超密集网络中多用户上行功率控制方法[J].南京师范大学学报,2022,0(1):016-23. |
| |
| 作者姓名： | 毛 晋 熊 轲 位 宁 张 煜 张锐晨 |
| |
| 作者单位： | (1.北京交通大学计算机与信息技术学院,北京 100044)(2.交通数据分析与挖掘北京市重点实验室,北京 100044)(3.中兴通讯股份有限公司,广东 深圳 518057)(4.移动网络和移动多媒体技术国家重点实验室,广东 深圳 518055)(5.国网能源研究院有限公司,北京 102209) |
| |
| 摘 要： | 针对超密集网络中由于用户数量多、相互距离近,通信过程中彼此之间干扰大,导致频谱利用率不高的问题,建立了通过优化控制发射功率同时提升系统信息容量和满足服务质量的用户数量的优化问题. 由于该问题非凸且功率控制为离散变量,将其建模为马尔科夫决策过程. 在此基础上,提出了基于深度强化学习的功率控制算法,并设计了相应的动作空间、状态空间及奖励函数. 仿真结果表明,所提算法与最大发射功率策略和随机发射功率策略相比,分别提高了至少15.9%的信息容量和至少10.7%的用户服务质量满足率. 与不考虑用户服务质量满足率提升的算法相比,所提算法通过适当降低信息容量,提升了用户服务质量满足率.
|
| 关 键 词： | 超密集网络 功率控制 信息容量 服务质量 深度强化学习 |
Power Control in Ultra Dense Network:A DeepReinforcement Learning Based Method |
| |
| Mao Jin,' target="_blank" rel="external">,Xiong Ke,' target="_blank" rel="external">,Wei Ning,' target="_blank" rel="external">,Zhang Yu,Zhang Ruichen,' target="_blank" rel="external">.Power Control in Ultra Dense Network:A DeepReinforcement Learning Based Method[J].Journal of Nanjing Nor Univ: Eng and Technol,2022,0(1):016-23. |
| |
| Authors: | Mao Jin " target="_blank">' target="_blank" rel="external"> Xiong Ke " target="_blank">' target="_blank" rel="external"> Wei Ning " target="_blank">' target="_blank" rel="external"> Zhang Yu Zhang Ruichen " target="_blank">' target="_blank" rel="external"> |
| |
| Affiliation: | (1.School of Computer and Information Technology,Beijing Jiaotong University,Beijing 100044,China)(2.Beijing Key Laboratory of Traffic Data Analysis and Mining,Beijing 100044,China)(3.ZTE Corporation,Shenzhen 518057,China)(4.State Key Laboratory of Mobile Network and Mobile Multimedia Technology,Shenzhen 518055,China)(5.State Grid Energy Research Institute Co.,Ltd.,Beijing 102209,China) |
| |
| Abstract: | For ultra-dense networks,in view of the problem of low spectrum utilization due to excessive users and large interference,an optimization problem is formulated to increase the system information capacity and satisfy the number of users with the quality of service(QoS)by optimizing the transmission power. Since the problem is non convex and the power control is a discrete variable,it is modeled as a Markov decision policy process. To this end,a power control algorithm based on deep reinforcement learning is proposed,and the corresponding action space,state space and reward function are designed. Simulation results show that compared with the maximum transmit power strategy and random transmit power strategy,the proposed algorithm improves the information capacity by at least 15.9% and the satisfaction of users’ QoS by at least 10.7%. Moreover,compared with the algorithm without considering the improvement of user’s QoS,the proposed algorithm improves the user’s QoS by appropriately reducing the information capacity. |
| |
| Keywords: | ultra-dense networks power control information capacity QoS deep reinforcement learning |
|
| 点击此处可从《南京师范大学学报》浏览原始摘要信息 |
|
点击此处可从《南京师范大学学报》下载全文 |
|
