Spectral Efficiency of Centralized and Decentralized Cooperative Networks with Relay Selection

In decode-and-forward cooperative networks, the best relay node is often selected among many available nodes to provide the diversity gain. Decentralized selection mechanisms such as timer-based relay selection are commonly used to achieve sub-optimal performance with low overhead. In this paper, we focus on the spectral efficiency of both centralized and decentralized relay-selection schemes. The impact of overhead, including channel estimation and feedback signals, is discussed. We show that centralized relay-selection achieve a higher spectral efficiency than decentralized schemes, only if the number of relays is small or the channel is static. Although decentralized schemes significantly reduce the amount of overhead, they also suffer from unavoidable performance loss. We also provide criteria for determining which scheme should be applied for given system environments.