Turbo-FQCA: Budget Constrained RIS Phase Optimization With Robust Throughput and Minimal Overhead

Reconfigurable Intelligent Surfaces (RIS) promise large array gains but raise a critical challenge: configuring hundreds of discrete phase shifters in real time under stringent latency and control overhead constraints. This paper proposes Turbo-FQCA, a quantization-aware, coordinate-ascent solver designed for single 5G-NR slot feasibility. Turbo-FQCA integrates a rank-1 proxy objective, Gauss–Southwell element ordering, warm-start initialization, and compact delta payload signaling to deliver near-optimal spectral efficiency within tight update and timing budgets. The proposed method achieves millisecond-level runtime through rank-1 updates and early stopping, while bit-efficient payloads scale only logarithmically with RIS size. Extensive simulations confirm three key outcomes: i) Turbo-FQCA consistently reaches ε–stationarity within a dozen flips, with nearly one-third of runs requiring no updates from warm start; ii) the achieved spectral efficiency (≥ 120 bps/Hz) remains robust across number of RIS elements, quantization levels and update toggles; and iii) control payloads remain in the range of a few bytes per slot, making signaling overhead negligible. Accordingly, Turbo-FQCA is not positioned as a classically optimal solver, but as a system-aware RIS control strategy that explicitly balances spectral efficiency, latency, and control signaling overhead under practical 5G-NR constraints.