🤖 AI Summary
To address the challenges of joint channel and power allocation, lack of global information, and high coordination overhead in multi-user dynamic spectrum sharing for cognitive radio networks, this paper proposes a distributed resource allocation algorithm based on game-theoretic learning. The method integrates non-cooperative game modeling with online reinforcement learning, enabling rapid convergence to Nash equilibrium without a central controller and adaptive optimization of spectrum access policies. Compared with conventional approaches, the proposed algorithm achieves significant performance gains: an average 18.7% increase in total system capacity, a 22.3% improvement in spectrum utilization, and a 35.1% reduction in channel collision rate. Moreover, it features low signaling overhead, strong robustness to dynamic environments, and excellent scalability—providing an efficient and practical autonomous coordination solution for distributed cognitive networks.