电力信息与通信技术2024，Vol.22Issue(5)：63-69,7.DOI:10.16543/j.2095-641x.electric.power.ict.2024.05.10
基于图注意力网络的无线信道功率资源优化分配
Optimized Wireless Channel Power Resource Allocation Based on Graph Attention Network
摘要
Abstract
In order to better optimize the transmission power of nodes in ad hoc networks and further improve the overall network throughput,this paper proposes a transmission power allocation algorithm based on graph neural network(GNN)theory.This algorithm takes the"unfolded weighted minimum mean square error"(UWMMSE)iterative algorithm as the overall framework,introduces the"graph attention network"model in the iterative structure,and trains specific parameters through unsupervised learning mechanism,while maintaining good optimization performance and accelerating algorithm convergence.The simulation results show that the power optimization allocation algorithm proposed in the paper can significantly reduce computational complexity while achieving better performance than similar algorithms.关键词
图注意力网络/展开加权最小均方误差/功率分配/图神经网络/5GKey words
graph attention network/unfolded weighted minimum mean square error/power allocation/graph neural network/5G分类
信息技术与安全科学引用本文复制引用
周想凌,胡晨,罗弦,吕苏,罗先南..基于图注意力网络的无线信道功率资源优化分配[J].电力信息与通信技术,2024,22(5):63-69,7.基金项目
国家电网有限公司总部科技项目资助"融合5G短切片、4G短复用的电力无线核心骨干专网关键技术研究与应用"(5108-202218280A-2-415-XG). （5108-202218280A-2-415-XG）
