A multi-stage dynamic spectrum sharing framework in cognitive radio networks

Dynamic spectrum access becomes a promising approach to increase the efficiency of spectrum usage. Serveral game theoretic model based dynamic spectrum sharing scheme were proposed to harmonize the system global utility and individual payoffs. Since DSAN users with limited rationality and selfishness adapt their behaviors to environmental dynamics gradually rather than quickly response to the change of environment as in traditional game, we focus on evolutionary features of the game process. We proposed a multistage dynamic spectrum sharing framework in which two stages are iterated. In Stage I, the centralized spectrum allocation is processed with the goal of getting interference free allocation with maximum system utilization. Stage II covers evolutionary spectrum trading and the evolutionary game based solution is extended to support global information and re-division of groups. Simulation shows the framework works well with dynamics of adaptation of users in DSANs, while achieving equilibrium for system utilities.