基于深度强化学习算法的蜂窝网络功率控制  

Power Control of Cellular Networks Based on Deep Reinforcement Learning Algorithm

作　　者：陈玲玲[1] 何伟 冯琦 CHEN Lingling;HE Wei;FENG Qi(Jilin Institute of Chemical Technology,Jilin Jilin 132022,China)

机构地区：[1]吉林化工学院,吉林吉林132022

出　　处：《信息与电脑》2024年第3期74-76,123,共4页Information & Computer

摘　　要：为了更加有效地利用频谱资源,提高频谱效率,在蜂窝网络中提出了频谱效率最大化问题,建立了最大发射功率约束下的功率控制模型。同时,为了解决信道状态信息不完美和迭代时间过长的问题,采用深度Q学习(DeepQ-Network,DQN)算法进行功率控制。另外,使用Python仿真,将基于DQN的算法与加权最小均方误差(Weighted Minimum Mean Square Error,WMMSE)算法和分数规划(Fractional Programming,FP)算法相对比,结果表明提出的基于DQN的算法在归一化性能和累计分布概率上接近最优,可以得到更高的频谱效率。In order to enhance the utilization of spectrum resources and improve spectrum efficiency,the problem of spectral efficiency maximization is proposed in cellular networks,and a power control model under the constraint of maximum transmitting power is established.At the same time,in order to solve the problem of imperfect channel state information and too long iteration time,Deep Q-Network(DQN)algorithm is used for power control.In addition,In addition,by using Python simulation,the DQN algorithm is compared with the Weighted Minimum Mean Square Error(WMMSE)algorithm and the Fractional Programming(FP)algorithm.The results show that the proposed DQN algorithm achieves performance close to the optimal solution in terms of normalized performance and cumulative distribution probability,and can obtain higher spectral efficiency.

关 键 词：蜂窝网络 深度强化学习 功率控制 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]