Dynamic Resource Allocation in Cognitive Radio Two-Way Relay Networks Using Sequential Auctions

In this paper, we consider a cognitive relay network, in which multiple pairs of secondary users (SUs) share the unlicensed spectrum by the overlay mode, and for each SU pair, the two ends exchange data via a common relay node that senses the spectrum and allocates the unused bandwidth. Here, we formulate the spectrum allocation problem as a multi-object auction and introduce the mechanisms of sequential first price auction (SFPA) and sequential second price auction (SSPA), where for each mechanism, the optimal bidding strategy is analyzed and the corresponding algorithm is given. Simulation results show that both the proposed algorithms cost only 25% similar to 35% feedback overhead, but achieve about 95% total transmission rate (with SSPA having a slightly better performance), compared with the optimal algorithm.