Spectrum Pooling-Based Optimal Internetwork Spectrum Sharing for Cognitive Radio Systems

Spectrum pooling, which allows the secondary networks to utilize the available spectrum bands from different licensed networks, is one of the most promising technologies for dynamic spectrum sharing in cognitive radio systems. Most previous work on spectrum pooling concentrates on the system architecture and the design of flexible access algorithms and schemes. In this paper, a distributed scheme for optimal internetwork spectrum sharing among multiple cognitive radio systems is proposed. Besides, the spectrum access price and spectrum efficiency are considered as the design criteria in the proposed scheme. The spectrum sharing problem is formulated as a restless bandits system, which dramatically reduces the computational complexity by simply allocating the new available band to the secondary network with the lowest index. Furthermore, extensive simulation results illustrate the significant performance improvement of the proposed scheme improves compared to the existing scheme.