Self-Directed Spectrum Allocation Framework for Integrated TN-NTN 6G Networks

2026-07-20

AI Summary

This work addresses the challenge of jointly optimizing interference management, user fairness, and system throughput in dynamic spectrum allocation for 6G integrated terrestrial–non-terrestrial networks. To this end, the authors propose a Q-learning-based adaptive channel allocation scheme that formulates spectrum assignment as a Markov decision process. A multi-objective reward function is designed to simultaneously enhance throughput, ensure fairness, and suppress interference, while an ε-greedy strategy facilitates efficient exploration during learning. Experimental results demonstrate that the proposed method achieves an average throughput of 28.5 Mbps, a Jain’s fairness index of 0.75, and a 26.3% reduction in interference compared to random allocation, with an average reward of 37.5. These outcomes underscore the approach’s superior adaptability to dynamic traffic conditions and its effectiveness in balancing competing performance objectives.

Abstract

This paper proposes a self-adaptive channel assignment framework based on Q-learning, where agents learn optimal policies by observing network load, interference conditions, and temporal traffic dynamics within a Markov decision process (MDP). A multi-objective reward function is designed to jointly optimize system throughput, user fairness, and interference mitigation, while an ε-greedy strategy is employed to facilitate effective exploration. Simulation results demonstrate stable convergence, achieving an average reward of 37.5 and an average throughput of 28.5 Mbps. Moreover, the proposed approach achieves a Jain's fairness index of 0.75 and reduces interference by 26.3% compared to random allocation by adaptively responding to dynamic traffic patterns.

Problem

• spectrum allocation
• integrated TN-NTN
• 6G networks
• dynamic traffic
• interference mitigation

Innovation

• Q-learning
• self-adaptive spectrum allocation
• integrated TN-NTN
• multi-objective reward
• Markov decision process