A scalable graph neural network for multi-antenna beamforming MU-MISO systems

A Bipartite Graph Neural Network Approach for Scalable Beamforming Optimization

This paper develops a bipartite graph neural network (bgnn) framework, a scalable deep learning solution designed for multi-antenna beamforming optimization of multi-user multiple-input single-output (mu-miso) downlink systems.This approach partitions the beamforming optimization procedure into multiple suboperations dedicated to individual antenna vertices and user vertices.Separated vertex operations lead to scalable beamforming calculations that are invariant to the system size.The vertex operations are realized by a group of deep neural networks (dnn) modules that collectively form the bgnn architecture.Identical dnns are reused at all antennas and users so that the resultant learning structure becomes flexible to the network size.As a result, the trained bgnn can be universally applied to arbitrary multi-antenna downlink systems.Numerical results validate the advantages of the bgnn framework over conventional methods.