Max-Min Fairness in Stacked Intelligent Metasurface-Aided Rate Splitting Networks

TL;DR

This work tackles fair resource allocation in downlink multiuser MISO by uniting RSMA with stacked intelligent metasurfaces (SIM) to perform wave-domain beamforming. The authors formulate a max-min rate problem (P1) that jointly optimizes base-station power coefficients and SIM phase shifts under unit-modulus constraints and a decodability constraint, solving it via alternating optimization using SCA for power and Riemannian conjugate gradient (RCG) on the complex circle manifold for the SIM. Results show that SIM-RSMA achieves min-rate performance close to SIM-SDMA and outperforms SIM-NOMA, with notable gains from increasing aperture $M$ and layers $L$, though the fully analog regime inherently biases toward SDMA under fairness objectives. These findings highlight the potential of wave-domain beamforming to enable fair, efficient multiuser operation with reduced digital precoding, while motivating future hybrid digital-analog SIM-RSMA designs to fully exploit RSMA under fairness-driven objectives.

Abstract

This paper investigates a downlink multiuser multiple-input single-output system that integrates rate-splitting multiple access (RSMA) with a stacked intelligent metasurface (SIM) to enable wave-domain beamforming. Unlike conventional digital beamforming, the proposed system leverages the programmable phase shifts of the SIM to perform beamforming entirely in the wave domain. In contrast to existing literature, this work introduces a fairness-centric SIM-RSMA design that shifts the emphasis from maximizing sum-rate to ensuring fair allocation of resources. In particular, we formulate a max-min rate optimization problem that jointly optimizes transmit power coefficients at the base station and SIM phase shifts. Given the non-convex nature of this problem, we develop an alternating optimization framework, where the power allocation is optimized through successive convex approximation and SIM beamforming is optimized using the Riemannian conjugate gradient method. Simulation results indicate that combining SIM with RSMA yields superior max-min performance compared to its integration with space division multiple access or non-orthogonal multiple access.

Figures (8)

• Figure 1: Illustration of the proposed SIM-assisted RSMA MISO system.
• Figure 2: Simulation setup of the proposed SIM-assisted RSMA system.
• Figure 3: The impact of the transmit power $P_{\max}$ on the minimum rate.
• Figure 4: The impact of the number of users $K$ on Jain's fairness index.
• Figure 5: The impact of the number of users $K$ on the minimum rate.