Dynamic spectrum decision in multi-channel cognitive radio networks with heterogeneous services
We study a dynamic channel selection framework for cognitive radio networks (CRNs) which support both delay sensitive and best effort services. Unlike existing works in the literature, we consider the effect of heterogeneous radio frequency characteristics and heterogeneous primary user activities on channel selection in multi-channel CRNs. Optimal spectrum decision policies are obtained to achieve minimum delay using dynamic programming techniques, such as Markov decision process (MDP) and reinforcement learning, under different assumptions. To address the computational complexity issue in the MDP solutions, a myopic scheme is proposed based on the estimated packet sojourn time.