Dynamic Chinese Restaurant Game in cognitive radio networks

In a cognitive radio network with mobility, secondary users can arrive at and leave the primary users´ licensed networks at any time. After arrival, secondary users are confronted with channel access under the uncertain primary channel state. On one hand, they have to estimate the channel state, i.e., the primary users´ activities, through performing spectrum sensing and learning from other secondary users´ sensing results. On the other hand, they need to predict subsequent secondary users´ access decisions to avoid competition when accessing the ”spectrum hole”. In this paper, we propose a Dynamic Chinese Restaurant Game to study such a learning and decision making problem in cognitive radio networks. We introduce a Bayesian learning based method for secondary users to learn the channel state and propose a Multi-dimensional Markov Decision Process based approach for secondary users to make optimal channel access decisions. Finally, we conduct simulations to verify the effectiveness and efficiency of the proposed scheme.