Nature-Inspired Evaluation of Data Classes for Trust Management in MANETs

In mobile ad hoc networks cooperation in packet forwarding is usually based upon the mixture of direct and indirect reciprocity. Nodes evaluate the level of cooperation of other network participants using a trust system. Its performance depends on classes of data used for description of forwarding behaviour of nodes. This paper studies the influence of two particular classes of data referred to as personal and general. A dynamic Nature-inspired model based on replicator equation is used to study the influence of these classes on the performance of the nodes. Computational experiments demonstrate that the distinction between personal and general data is not necessary if node´s priority is the forwarding efficiency or battery conservation. However, if network user wants to maximise his throughput or to create the best possible protection against selfish nodes, than prioritising personal data improves his performance.