Energy Efficiency Optimization in Reconfigurable Intelligent Surfaces-Assisted Downlink User-Centric Networks

Recently, user-centric network (UCN) is regarded as an effective method to improve the capacity for all users by providing communication service from a cluster of access points (APs). However, due to the densely deployed APs, the high hardware costs and power consumption become the bottleneck of UCN. To address this issue, we introduce a low-cost and high-gain reconfigurable intelligent surface (RIS) into downlink UCN, where passive RISs are deployed to assist the downlink transmission from APs to users. To achieve the energy efficient communication in proposed RIS-assisted downlink UCN, a global energy efficiency optimization problem is formulated with respect to user association, phase shift and downlink power allocation. Due to the non-convexity of the formulated problem, we first deal with the user association problem through linear conic relaxation method. Then the phase shift as well as power allocation are optimized alternately, where phase shift is optimized through fractional programming method and power allocation is tackled by constructing a surrogate function. Simulation results demonstrate that our proposed algorithm could bring about 375% gain on energy efficiency comparing to the “No RIS” benchmark algorithm, which also corresponds to about 165% spectral efficiency gain.