Multilayer Feed-forward Neural Network learning based Dynamic Chinese restaurant model for dynamic spectrum access in cognitive radio networks

As an effective approach to improve spectrum efficiency, cognitive radio network make it possible for secondary users (SU) to share the spectrum with primary users (PU), on the condition that the primary users have preemptive priority. In this paper, we applied the Dynamic Chinese restaurant game, which ideally modeled the spectrum sensing and access in cognitive network. We propose the use of Multilayer Feed-forward Neural Networks (MFNN) as an effective method for users to learn the network state, which can be regarded as how a customer learn the table state in the restaurant. Moreover, in order to select an optimal table for a secondary user, subsequent secondary users' sequential decisions are considered. The effectiveness and efficiency of the proposed scheme is verified in the simulation.