Abstract
This study introduces a novel approach for energy-efficient resource allocation in millimeter-wave networks, assisted by multiple intelligent reflecting surfaces (IRS). The proposed framework optimizes the dynamic ON/OFF control and phase shifts of IRS elements, along with beamforming (BF) at access points (AP), under practical constraints. Unlike existing methods, our model enhances energy efficiency (EE) by optimizing a fixed number of ON IRS elements. We present innovative algorithms, including a modified nested fractional programming (NFP) for BF and a simulated annealing (SA)-type algorithm for phase shift and element selection. Our results demonstrate a 6.5-fold improvement in EE under a realistic scenario compared to benchmark, highlighting the effectiveness of our approach as a crucial strategy for future 6G networks.