Munawar, MuteenMuteenMunawarLee, KyungchunKyungchunLee2026-04-202026-04-2020251536-1276https://imec-publications.be/handle/20.500.12860/59124This paper addresses a multi-user multiple-input multiple-output (MIMO) wireless network enhanced by a dual-polarized intelligent reflecting surface (DP-IRS). The system encompasses multiple DP antennas at both transmitters and receivers, each equipped with a single RF chain featuring independent phase shifters. Our objective is to minimize the total transmit power at the transmitter while satisfying per data stream target spectral efficiency constraints for all users. To achieve this, we introduce a multi-step alternating optimization algorithm. Additionally, we propose a novel scheme to reduce the computational complexity of beamforming design in IRS-assisted multi-user wireless networks. Specifically, by leveraging users’ channel information, we classify them into two distinct groups, simplifying problem formulations and significantly decreasing computational costs with minimal impact on performance. Numerical results illustrate the superiority of our proposed algorithms over various benchmark schemes. In a comparison between DP-IRS and Simple IRS, DP-IRS requires 41.7% less transmit power for 50 reflecting elements and eight users, while targeting a spectral efficiency of 4.6 bps/Hz per data stream for each user. Additionally, the proposed algorithm, based on the user grouping scheme, yields a 75.5% reduction in computational cost compared to the algorithm without user grouping, with negligible performance degradation.engJournal article10.1109/TWC.2025.3568507WOS:001591677200043INTELLIGENTOPTIMIZATIONMETASURFACECONVERSIONSURFACES