RIS-Aided Massive MIMO Performance-Complexity Trade-off Optimization

This paper explores the performance-complexity trade-off for reconfigurable intelligent surface (RIS)-aided massive MIMO (mMIMO) systems, focusing on maximizing the sum–spectral efficiency (SE) under zero-forcing (ZF) with instantaneous channel state information (CSI). The study employs two distinct optimization approaches: manifold-based optimization and a metaheuristic evolutionary genetic algorithm (GA). The paper analyzes the effectiveness of these methods in optimizing the system sum–SE via optimizing the passive beamforming in a RIS-aided mMIMO system under real-world passive RIS constraints. Key objectives include standardizing system models, validating the manifold-based approach and results, and evaluating the performance of both strategies. The paper highlights the pros and cons of each method and analyzes their performance for different scenarios, including various user configurations and system parameters. Numerical simulations are conducted to showcase the performance of both methods in terms of sum–SE and computational complexity. The study concludes by summarizing the key findings and highlighting the importance of optimizing RIS-aided mMIMO systems for enhanced communication performance and efficiency.