Self-Interested Nodes Modeling in Mobile Wireless Ad Hoc Networks

The Increasing adoption of the techniques of mobile wireless ad hoc network in the field of civilian applications arouses many urgent problems. As we known nodes to consist a MANET may belong to different users, thus each node aims to pursue the best quality of service for its owner, i.e., they are self-interested nodes trying to achieve the maximum utility. Researchers have shown that selfish nodes in a MANET can greatly decrease the performance of the network, even to the point where no messages can be sent at all. By the similarities between MANET and multi-agent system (MAS), to research some problems in MANET from the view of MAS is an impressive idea. In this paper nodes in MANET are modeled as intelligent agents using an agent programming language named Conceptual Agent Notation (CAN). We don't use CAN directly but extend its semantics with a set of core operations which capture the essence of ad hoc network, thus provide some solid theoretic underpinning for researching MANET from the view of MAS.