Autonomous spectrum sharing by well-designed games

From static spectrum allocation to nowadays more liberated policies such as spectrum refarming or even opportunistically exploiting the so-called spectrum holes or "white spaces," we have witnessed many changes all over the world regarding how spectrum is being allocated. The message is clear, and that spectrum allocation needs to be more dynamic and adaptive to the environment and applications. However, dynamic spectrum sharing and access is complicated in many ways. Firstly, it requires global knowledge of channel states for all communication links in the entire network and, secondly, the required large-scale optimization would be computationally prohibitive to achieve, not to mention that channel states vary over time as well. Importantly, there is a strong desire that such dynamic spectrum sharing be realized by a large number of uncoordinated mobile radios in a distributed and autonomous fashion. This is the focus of this chapter which discusses game-theoretic methods for self-optimization of cognitive mobile radios in spectrum sharing. The chapter will begin by reviewing the not-so-flexible spectrum management in cellular networks and then covering the topics of using forward-looking games in spectrum allocation. A major result is that autonomous spectrum sharing leading to spectral-efficient solutions is shown possible by well-designed games requiring only local channel knowledge at individual mobile radios and such interactive self-optimization can also be employed under the hierarchical spectrum sharing model in which primary spectrum owners are present and need to be protected.