Technical gazette, Vol. 32 No. 3, 2025.
Original scientific paper
https://doi.org/10.17559/TV-20241015002061
Genetic Algorithm Based Approach for Optimization of Fair Power Allocation in Pd-NOMA Systems
Seda Kirtay
orcid.org/0000-0003-2415-9131
; Marmara University, Maltepe, Istanbul, Turkiye
*
Veysel Gokhan Bocekci
; Marmara University, Maltepe, Istanbul, Turkiye
Kazim Yildiz
; Marmara University, Maltepe, Istanbul, Turkiye
* Corresponding author.
Abstract
Power Domain Non-Orthogonal Multiple Access (PD-NOMA) systems provide a viable solution for improving user fairness and spectral efficiency in wireless communication networks. However, to maximize system performance while fulfilling numerous users' high Quality of Service (QoS) needs, innovative fair power allocation algorithms are required. This paper provides a genetic algorithm (GA)-based approach for adaptively optimizing power distribution for three users at varied distances and channel conditions in PD-NOMA systems. The proposed technique dynamically modifies the Power Allocation based on the users' distances from the Base Station (BS), channel fading, and modulation schemes, by offering fair power allocation therefore reducing the Bit Error Rate (BER). Unlike traditional approaches such as water-filling and fixed power allocation, GA successfully explores wide solution spaces, reducing the possibility of local optima and guaranteeing equitable power distribution. The simulation results were compared to the conventional approach as uniform, revealing considerable performance gains with BER reductions of nearly zero up to 100 dBm for the furthest, mid-range, and nearest users, respectively. GA converges within 4000 iterations and maintains the fitness value at 0.7. These findings show GA's capability to deliver dynamic solutions that adapt to changing network circumstances, as well as its ability to improve fairness and efficiency in PD-NOMA systems.
Keywords
fair power allocation; genetic algorithm; NOMA; non-orthogonal multiple access; power domain
Hrčak ID:
330560
URI
Publication date:
1.5.2025.
Visits: 632 *