ABSTRACT Integrating intelligent reflective surfaces (IRSs) with millimeter‐wave (mmWave) massive MIMO systems is a promising strategy for enhancing the performance of next‐generation wireless networks. However, practical limitations complicate system throughput optimization, such as the restricted allocation of dedicated radio frequency (RF) chains per antenna and the passive nature of IRS phase shift elements. This paper demonstrates that the equivalent channel gain can be maximized by appropriately designing the IRS component phases to optimize spectral efficiency, resulting in challenging non‐convex optimization problems. To address this, we employ two approaches for maximizing channel gain. The first approach is an algorithm based on Eigenvalue Decomposition (EVD). Given the computational intensity required to calculate eigenvalues and eigenvectors, we propose an alternative algorithm based on gradient projection methods (GPM), which offers significantly reduced computational complexity despite a slight performance reduction compared to EVD. Additionally, we introduce a Phase Coherence Method (PCM) framework for designing hybrid precoder and combiner matrices in both the transmitter and receiver, addressing the challenge of uni‐modular matrices. Comprehensive simulations reveal that the proposed algorithms achieve near‐optimal performance, surpass alternative algorithms, and effectively advance the integration of IRS with mmWave massive MIMO systems.
Momen‐Tayefeh et al. (Wed,) studied this question.