Distributed resource allocation in cognitive radio systems based on social foraging swarms

The goal of this paper is to propose a distributed resource allocation strategy for the access of opportunistic users in cognitive networks. The opportunistic, or secondary, users are modeled as a set of agents looking for the most appropriate slots to use, in the time-frequency domain, with the goal of avoiding conflicts among themselves and to yield possibly no interference to the primary users. The problem is how to coordinate the access from the secondary users in a totally decentralized fashion, requiring a minimal coordination among them. We propose a solution to this problem based on a social forage swarming model, where the search for the most appropriate slots is modeled as the motion of a swarm of agents in the resource domain (typically the time-frequency plane), looking for "forage", representing a function inversely proportional to the interference level. The swarm tends to move in the time-frequency region where there is less interference, while satisfying one basic requirement: minimize the spread in the resource domain while avoiding collisions between the allocations of different users. These properties are enforced through the decentralized minimization of a potential function that incorporates a long-range attraction between the resources occupied by secondary users (to minimize spreading) and a short-range repulsion (to avoid conflicts).