注：本代码是从Jun Fang教授的主页下载 **DQN-power-control** is the code for applying deep reinforcement learning for spectrum sharing in cognitive radios, available at http://www.junfang-uestc.net/codes/DQN-power-control.rar. The code has been tested on Ubuntu 14.04 + tensorflow 1.0.0 + Python 2.7. Reference: Xingjian Li, Jun Fang, Wen Cheng, Huiping Duan, Zhi Chen, and Hongbin Li, "Intelligent Power Control for Spectrum Sharing in Cognitive Radios: A Deep Reinforcement Learning Approach," was accepted by IEEE Access, May 2018. Written by: Wen Cheng Email: JunFang@uestc.edu.cn Get Start: cd ./code python main.py Optional parameters (main.py): noise - The estimation error of RSS num_sensor - The number of sensors policy - Power update policy of PU, suppose to be 1 or 2