Spectrum Trading in Cognitive Radio Network: An Agent-based Model

In this paper, we propose an agent-based spectrum trading mechanism, where secondary users (SUs) can access the spectrum through the Secondary Network Operator (SNO). Buying spectrum from primary users(PUs) and reselling them to secondary users(SU), SNO plays an agent-like role in the spectrum trading process. The most significant challenge for SNO is how to make most profitable strategy. We address this challenge for both noncompeting and competing Secondary Network Operators, using theories and techniques from Newsvendor Problem and Supermodular Game. We prove that optimal strategies can be achieved in one or multiple trading periods by distributed algorithms. To our best knowledge, this model is the first solution to agent-based spectrum trading, and SNO’s optimal strategy is also addressed for the first time.