Batch Gradient Descent-based Optimization of WMMSE for Rate Splitting Strategy

Abstract

Rate Splitting (RS) multiple access is a general and robust multiple access framework for downlink multi-antenna communication systems, which splits each user's message into common and private parts, and superposes the common message and the private message for transmission. To find the transmit beamforming design, the Alternate Optimization Weighted Minimize MSE (AO-WMMSE) algorithm is often used as one of the approaches to maximize WSR due to its low computational complexity and good convergence. However, AO-WMMSE is inefficient in large-scale calculations and when the number of users and antennas is large. We propose a Batch Gradient Descent-based Weighted Minimum MSE (BGD-WMMSE) method rapidly optimize the linear precoder for large-scale data computing. Simulations show that BGD-WMMSE can achieve the same size of maximum achievable rate region as AO-WMMSE in one-layer RS deployment.