Intent Driven Near-Field STAR-RIS Beamforming via Hybrid Quantum-Classical Optimization for MIMO-NOMA IoT Downlink Networks

Intent-Based Networking (IBN) necessitates autonomous radio infrastructures capable of translating high-level service intents into self-optimizing configurations with minimal human intervention. Inspired by this vision, we examine the intention to maximize downlink spectral efficiency subject to quality of service (QoS) and power constraints in a near-field simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) assisted MIMO-NOMA system. The Fresnel-zone geometry induces element-wise path-loss dispersion and quadratic phase curvature, making the joint optimization of discrete phase shifts θ_x,y ∈ Θdis, continuous focusing gains g_x,y and energy-splitting ratios (β_t,x,y, β_r,x,y), together with BS precoders {W_k}, a high-dimensional mixed-integer non-convex program. We propose Quantum approximate-and-gradient optimizer for near field (QAQG-NF), an IBN-compliant hybrid quantum-classical solver that alternates (i) block-coordinate quantum approximate optimization (QAOA) for the combinatorial phase codebook and (ii) quantum natural-gradient descent (QNGD) for the continuous variables. For a 2×6 m (N=192) STAR-RIS at 3.5 GHz, QAQG-NF attains 7.46 bit/s/Hz at P_TX=38 dBm, i.e., 5.1× the sum-rate of a near-field reflecting-only projected-gradient baseline and 2.8× that of a far-field projected-gradient ascent design while converging in < 40 outer iterations.