Energy Efficiency Optimization in RIS-Aided Wireless Networks: Active Versus Nearly-Passive RIS with Global Reflection Constraints

This work addresses the topic of energy efficiency (EE) maximization in wireless networks aided by a reconfigurable intelligent surface (RIS), considering both active and nearly-passive RISs. Moreover, unlike available literature, the analysis considers that the RIS reflection constraints are not enforced separately on each reflecting element, but the RIS is characterized by a single reflection constraint that is jointly enforced on all the reflecting elements. This allows for a wider set of feasible RIS reflection matrices, which leads to better performance, which is theoretically analyzed. Then, two provably convergent energy efficiency maximization algorithms with polynomial complexity are developed, which optimize the reflection coefficients of the RIS, the mobile users’ transmit powers, and the linear filters at the base station. Numerical results are illustrated to quantify the performance of the proposed methods and identify the operating points in which active or nearly-passive RISs should be preferred from the energy efficiency standpoint.