Reflection Resource Management for Intelligent Reflecting Surface Aided Wireless Networks

• Published in: IEEE transactions on communications Vol. 69; no. 10; pp. 6971 - 6986
• Main Authors: Gao, Yulan, Yong, Chao, Xiong, Zehui, Zhao, Jun, Xiao, Yue, Niyato, Dusit
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.10.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; alternating direction and method of multipliers (ADMM); Approximation; Approximation algorithms; Array signal processing; Beamforming; Computational geometry; Convexity; group sparsity; Intelligent reflecting surface (IRS); Mathematical programming; Modular structures; Modules; passive beamforming; Phase shifters; Power demand; Quality of service; Reconfigurable intelligent surfaces; reflection resource management; Resource management; transmit power allocation; Wireless networks
• DOI: 10.1109/TCOMM.2021.3093312
• ISSN: 0090-6778

In this paper, the adoption of an intelligent reflecting surface (IRS) for multiple user pairs in two-hop networks is investigated. Different from the existing studies on IRS that mainly focused on tuning the reflection coefficients of all elements, we consider the implementation of true reflection resource management (RRM) through the identification of the best triggered module subset. More precisely, the implementation of true RRM builds on the premise of our proposed modular IRS structure consisting of multiple independent and controllable modules. In the context of modular IRS structure, we investigate the signal-to-interference-plus-noise ratio (SINR)-based max-min problem subject to per source terminals (STs) power budgets and module size constraint, via joint triggered module subset identification, transmit power allocation, and the corresponding passive beamforming. Whereas this problem is NP-hard due to the module size constraint, which can be addressed by the convex sparsity-inducing approximation to the hard module size constraint using mixed <inline-formula> <tex-math notation="LaTeX">\ell _{1,F}\text {-norm} </tex-math></inline-formula>, where it yields a suitable semidefinite relaxation. Using techniques from separable convex programming, we provide a two-block alternating direction method of multipliers (ADMM) algorithm for the approximated problem. Numerical simulations are used to validate the analysis and assess the performance of the proposed algorithm as a function of the system parameters. Further energy efficiency (EE) performance comparison demonstrates the necessity and meaningfulness of the introduced modular IRS structure. Specifically, for a given network setting, there is an optimal value of the number of triggered modules for system, when the EE is considered.