Performance evaluation of OMP, SVD, deep unfolding, and genetic algorithm-based hybrid beamforming for IRS-assisted mmwave MU-OFDM-mMIMO in B5G networks

Abstract

This paper examines the performance analysis of the hybrid beamforming algorithms, which are Orthogonal Matching Pursuit (OMP), Singular Value Decomposition (SVD), Deep Unfolding, and Genetic Algorithms (GA), in IRS-assisted MU-OFDM massive mMIMO systems operating in mmWave channels for B5G networks. This paper evaluates these algorithms across a comprehensive set of performance metrics, including Spectral Efficiency (SE), Energy Efficiency (EE), Bit Error Rate (BER), and Outage Probability, with an IRS configuration with 2-bit quantized phase shifts meticulously optimized to maximize the system sum rate. The research uses an advanced simulation framework to examine the algorithms’ performance under different Signal-to-Noise Ratio (SNR) values, ranging from -10 dB to 30 dB, that reflect typical mmWave channel environments. The simulation results show that OMP performs better in achieving high SE and minimizing BER, especially at high SNR values, and GA is better at optimizing EE, which makes it suitable for energy-constrained scenarios. The inclusion of IRS technology significantly enhances the overall system reliability and efficiency, with a notable reduction in outage probability, which validates its potential as a key component in B5G network design. These results provide valuable insights into the practical implementation and optimization of hybrid beamforming strategies, which will guide the development of robust and efficient next-generation wireless communication systems.