Adaptive AI for QoS Management in 6G Networks: Leveraging DRL for Optimal Performance

Department of Electrical Engineering Technology, College of Applied Industrial Technology (CAIT), Jazan University, Baysh, Saudi Arabia.

^* Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.

Published online: 26 June 2026

Abstract

The research investigated the application of Adaptive AI for Quality of Service (QoS) Management in 6G Wireless Networks that are affected by both α-η-μ fading and Heavy Co-channel Interference (CCI); this was done through the use of Deep Reinforcement Learning (DRL) methods where an agent learns optimal ways to solve problems such as Power Control, Channel-Switching and Modulation Schemes via interacting within a simulated environment. Simulations based on the use of MATLAB showed that the DRL agent found a convergence point and outperformed non-adaptive static approaches in terms of four metrics: throughput, latency, packet-delivery-ratio, and fairness In addition to utilizing adaptive AI as a means to improve system performance, the DRL-based agent also was able to utilize less power than its static counterparts, and provided good quality of service (QoS) for various forms of traffic (e.g., eMBB, URLLC, mMTC). Additionally, heatmap and ablation studies were completed to show that every design choice from the research conducted throughout this work have been contributing to successfully utilizing adaptive AI as an aid in next-generation wireless systems to be developed through 6G.