Argumentation based negotiation in cognitive radio networks

Cognitive radio networks allow for a more dynamic allocation of network resources (such as the radio spectrum) compared to conventionally engineered networks. We consider a scenario where each node acts as an autonomous agent, maintaining a knowledge base of the network conditions and pursuing its own communication goals. The agents use negotiation to reach agreements about the allocation of network resources and the configuration of forwarding paths. The contribution of this paper is a novel, argumentation based negotiation framework. Instead of a simple exchange of offers, the agents also offer arguments in favor of their proposals and critiques of received offers. Arguments can contain relevant new information about the state of the network, allowing the agents to update their knowledge base. The system is governed by a collection of rules and policies, which can be invoked in the support of requests or, in some cases, to justify their denial. To illustrate the proposed scheme, we designed a negotiation protocol and argumentation language for a communication model involving discrete service levels. We verify the flow of negotiation through a simulation study.