Efficient Selection of Decoding Order for Rate Maximization in Uplink RSMA

In uplink rate-splitting multiple access (RSMA), system performance critically depends on the decoding order at the base station, as it directly governs residual interference and achievable sum rate. However, jointly optimizing decoding order with power allocation is a highly non-convex and combinatorial problem, making exhaustive search impractical. In particular, the discrete nature of the decoding order variable further complicates the optimization process, as it prevents the direct application of conventional convex optimization techniques. This letter introduces an alternating optimization framework that efficiently optimizes the selection of decoding order. To handle the integer-valued decoding order variables, we reformulate the problem using a binary surrogate variable that captures user interference relationships, with additional constraints to ensure feasibility. The resulting problem is then convexified using the fractional programming method. Simulation results demonstrate that the proposed scheme achieves near-optimal sum-rate performance, closely approaching exhaustive search results for both small- and large-scale system configurations, while incurring significantly lower computational complexity. © 2026 IEEE.