Multi-armed bandits with dependent arms for Cooperative Spectrum Sharing

摘要: Cooperative Spectrum Sharing (CSS) is an appealing approach for primary users (PUs) to share spectrum with secondary (SUs) because it increases the transmission range or rate of PUs. Most previous works are focused on developing complex algorithms which may not be fast enough real-time variations such as in channel availability. Instead, we develop a learning mechanism PU enable CSS strongly incomplete information scenario low computational overhead. We model discover SU interact and what offer make combination Multi-Armed Bandit (MAB) Markov Decision Process (MDP). By means Monte-Carlo simulations show that, despite its overhead, our proposed converges optimal solution significantly outperforms ϵ-greedy heuristic. This algorithm can extended include more sophisticated features while maintaining desirable properties speed convergence.