AI-Assisted Hybrid GA–PSO Channel Allocation Under 3GPP TR 38.901 UMa for Efficient 5G Radio Resource Management

The increased traffic and heterogeneity in the 5G/B5G network require efficient radio resource management (RRM). However, the existing methods, such as GA and PSO, have poor convergence speed and require proper initial conditions. Additionally, learning-based methods have high computational complexity. Hence, in this paper, a novel AI-assisted Hybrid Channel Allocation (AI–HCA) framework is proposed by using a support vector regression (SVR)-based predictive initialization method and GA-PSO optimization. Simulation results using the 3GPP UMa channel model show that the proposed method has a 28% reduction in call blocking probability (CBP), a 15-25% enhancement in spectral efficiency (SE), and a 18-25% enhancement in energy efficiency (EE) with a faster convergence speed (35-40 iterations) than the existing methods. Additionally, the proposed method has been validated using the ANOVA test (p < 0.05) to confirm the significant improvements. Hence, the proposed AI-assisted Hybrid Channel Allocation framework has the potential to provide a low-complexity, robust, and scalable solution for intelligent radio resource management in 5G/B5G networks.