Energy Efficiency Optimization Based on Deep Reinforcement Learning in D2D Communication Network
ZHONG Xing
LI Jun
LI Zhengquan
ZHU Minghao
SHEN Guoli
ZHANG Xixi
Abstract：In order to solve the problems of spectrum resource shortage,serious inter-link interference and high energy con-sumption of mobile communication networks,for the Device to Device(D2D)network,a deep reinforcement learning(RL)algo-rithm based on a distributed framework is considered,which uses double Q and dueling architecture to solve the problems of exces-sive state space and Q value overestimation.Simultaneous wireless information and power transfer(SWIPT)are used to effectively compensate for system energy consumption.Under the non-linear constraints of the minimum required throughput of cellular users and the power splitting ratio of D2D user,based on the changing location and channel state information(CSI),synchronously allo-cating resource blocks(RB),D2D transmitted power and power splitting ratio to achieve energy efficiency(EE)optimization.The simulation results show that under the premise of ensuring sum rate of cellular users,the proposed scheme is better than the base-line algorithm in terms of the total EE of the D2D links,and is more robust in communication networks with large-scale mobile us-ers.
Keywords：SWIPTEERB allocationpower splitting ratioDRL
Publication Date：2023-11-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：9( 2475-2482,2489 )