Two novel learning algorithms to solve the spectrum sharing problem in cognitive radio networks

To improve the spectral efficiency in cognitive radio networks, it is essential for cognitive radio users to be equipped with intelligent learning capability. Many different learning methods have been applied in different kinds of cognitive radio network models. This study presents two novel learning algorithms that can be applied to cognitive radio network models based on IEEE802.22. One is a no-regret learning method and the other is a reinforcement learning algorithm. The experimental results show that both methods can be effectively applied in cognitive radio networks. Moreover, the reinforcement learning out performs the no-regret learning method.