eduzhai > Applied Sciences > Engineering >

Deep Reinforcement Learning for Joint Beamwidth and Power Optimization in mmWave Systems

  • Save

... pages left unread,continue reading

Document pages: 5 pages

Abstract: This paper studies the joint beamwidth and transmit power optimizationproblem in millimeter wave communication systems. A deep reinforcement learningbased approach is proposed. Specifically, a customized deep Q network istrained offline, which is able to make real-time decisions when deployedonline. Simulation results show that the proposed approach significantlyoutperforms conventional approaches in terms of both performance andcomplexity. Besides, strong generalization ability to different systemparameters is also demonstrated, which further enhances the practicality of theproposed approach.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×