RIS Optimization and Scaling Laws in Multi-Operator Systems: Is Quadratic Scaling Achievable?

2026-05-02

AI Summary

This work addresses the challenge in multi-operator coexistence scenarios where conventional reconfigurable intelligent surfaces (RISs) unintentionally reflect non-serving operators’ signals, causing interference and exacerbating channel fluctuations. The authors propose a practical RIS design framework that simultaneously maximizes the received power for the serving operator and enforces a constant equivalent reflected channel for non-serving operators, deriving a closed-form optimal solution. The study reveals, for the first time, that quadratic scaling of received power with the number of RIS elements is achievable only when employing beyond-diagonal RIS (BD-RIS) architectures with sufficient interconnectivity. Specifically, the group size $G_s$ in group-connected BD-RIS must satisfy $G_s \geq L$, where $L$ denotes the number of operators; otherwise, only linear gain is attainable. Theoretical analysis and simulations demonstrate that, in a two-operator system, a BD-RIS with $G_s=2$ achieves a 13 dB gain over conventional RIS at 128 elements.

Abstract

This paper studies multi-operator wireless communication systems aided by general reconfigurable intelligent surface (RIS), including both conventional single-connected RIS and beyond-diagonal RIS (BD-RIS). Specifically, we consider a system where multiple operators coexist in the same area over different frequency bands, each with a single-antenna base station, while one operator serves its single-antenna user with the aid of an RIS. In such a system, the RIS may unintentionally reflect signals from the non-serving operators, leading to inter-operator interference and rapid fluctuations of their effective channels. To address this issue, we propose a practical RIS design framework that maximizes the received signal power of the serving operator while enforcing fixed RIS-reflected channels of the non-serving operators. We derive closed-form solutions to the resulting optimization problem, based on a novel technique to deal with the coupled unitary and linear equality constraints. We further give scaling law analysis of the received signal power. For a two-operator system, the received signal power scales quadratically with the number of RIS elements for group-connected BD-RIS with group size Gs>=2, whereas for conventional single-connected RIS it scales only linearly. More generally, for an L-operator system with L-1 non-serving operators, the scaling-law transition occurs at Gs=L, where quadratic scaling is achieved when Gs>=L, and linear scaling otherwise. These results demonstrate that, in a multi-operator system, quadratic scaling is achievable only with BD-RIS architectures having enough interconnections. Simulation results validate the analysis and show the significant gain of BD-RIS over conventional RIS in multi-operator systems. In particular, group-connected BD-RIS with Gs=2 achieves a 13dB gain over conventional RIS in a two-operator system with a 128-element RIS.

Problem

• multi-operator systems
• reconfigurable intelligent surface (RIS)
• inter-operator interference
• channel fluctuations
• beyond-diagonal RIS

Innovation

• Reconfigurable Intelligent Surface (RIS)
• Beyond-Diagonal RIS (BD-RIS)
• Multi-Operator Interference
• Quadratic Scaling Law
• Group-Connected Architecture