User Grouping-Based RSMA for Active RIS with Finite Blocklength: Spectral and Energy Efficiency Tradeoff

Rate-splitting multiple access (RSMA) has emerged as a promising multiple access technique for next-generation wireless systems, offering flexible inter-user interference management and enhanced transmission robustness. However, its achievable common stream rate is often constrained by the user with the worst channel condition. The integration of RSMA and active reconfigurable intelligent surfaces (RIS) has been considered as a powerful strategy to mitigate this inherent limit by reshaping the propagation environment and enhancing channel conditions. Most existing works adopt a scheme where all users employ successive interference cancellation (SIC) to decode the common stream under the infinite blocklength regime, which causes the common stream rate to be constrained by the user with the poorest channel quality and the infinite blocklength assumption leads to unquantified performance degradation in practical communication systems. In this paper, we consider the resource efficiency (RE) as the system metric, which balances the energy efficiency (EE) and spectral efficiency (SE), while accounting for the impact on RSMA with finite blocklength. Specifically, we jointly optimize multiple design variables, including base station (BS) precoding, rate allocation, RIS configurations and user grouping. We transform the problem by a two-layer iteration and further decompose it into three interrelated subproblems and develop an optimal algorithm for precoding and rate allocation based on the successive convex approximation (SCA) method. The penalty dual decomposition (PDD) method is employed to optimize the discrete RIS configuration, while the simulated annealing (SA) method is applied for user grouping. Simulation results validate the effectiveness of the proposed method, demonstrating notable improvements over existing benchmarks.