Feasibility-Screened SINR Matching and Actual-Power CCP for Sum-Rate Maximization in Downlink PD-NOMA Networks
Year: 2026
Type: preprint
Abstract: The throughput advantage of power-domain non-orthogonal multiple access (PD-NOMA) depends on two coupled decisions: selecting decodable user pairs and allocating powers without violating quality-of-service (QoS), successive-interference-cancellation (SIC), ordering, and total-power constraints. This paper proposes a two-stage framework that first constructs feasibility-screened signal-to-interference-plus-noise ratio (SINR) preferences and then optimizes the actual strong- and weak-user powers by difference-of-concave (DC) programming and the convex-concave procedure (CCP). The actual-power parameterization converts the QoS and power-priority requirements into affine constraints and avoids the bilinear coupling created by independently optimizing total subchannel power and a power-split factor. A Hessian argument proves that the fixed-pair sum rate has the form F(p) − G(p), with F and G concave. Linearizing G gives a tight concave lower-bound surrogate; accepted CCP iterates are therefore non-decreasing in the exact objective under the stated solution conditions, while global optimality is not claimed. A MATLAB R2016b implementation is evaluated using 500 paired Monte Carlo realizations per point. At 100 candidate users, SINR-CCP improves mean sum rate by 2.80% over Gain-CCP, isolating the matching contribution, and by 14.41% over Gain-Heuristic for the complete framework. At 40 dBm, the corresponding gains are 3.95% and 24.02%. The paired 95% confidence intervals remain positive at the reported endpoints, and the numerical audit records monotone accepted-objective traces with negligible affine-constraint residuals. The results support a selective, interference-aware PD-NOMA gain rather than universal superiority over orthogonal access.
Source: Research Square
Institutions: Multimedia University, Telekom Malaysia Berhad (Malaysia)
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
Funder: Telekom Malaysia Berhad
Award: G7778711678
OpenAlex ID: W7225869312
