With the development of radio communication technology, spectrum resource has been more widely employed, and the scarceness of wireless spectrum hampers the continual development of wireless communication. Using the licensed spectrum in "secondary utilization" way, cognitive radio can solve availably the problem of lacking wireless spectrum. This new technology arises as a hot topic in wireless communication in recent years.This paper mainly researches dynamic spectrum allocation algorithm, which is one of cognitive radio key technologies. At the beginning of this paper, the network architecture and the system platform are analyzed,and the four models of the dynamic spectrum allocation algorithm are introduced. Secondly, it mainly studies the existing spectrum allocation algorithm, which based on graph coloring theory. Aiming at the disadvantage of overmuch allocation time in intrinsic allocation algorithm, an improved spectrum allocation algorithm is given which adopts the method of grouping. Simulation shows that the improved algorithm can reduce the time of spectrum allocating, at the same time achieve the similar performances. Finally, aiming at the disadvantage of that the intrinsic algorithm doesn't take the different needs from different users into account in the process of distribution, another improved algorithm is given, which considers user's demand. The simulation shows that the improved algorithm could increase the fairness of the allocation and the utilization of spectrum comparing with the existing algorithm. |