Learning controlled forwarding strategy improving probabilistic routing in DTNs

Delay tolerant networking (DTN) is able to provide communication services in areas where there is no guarantee that a fully connected path between source and destination exists at any time, and seek to address the technical routing issues in wireless networks. Traditional routing protocols are unable to deliver messages between hosts in such conditions. PRoPHET protocol is among the most widely used routing protocols for such networks. The choice of message forwarding strategies can have a major impact on system performance. In this paper, we propose a controlled forwarding strategy for probabilistic routing protocols that can optimize different performance metrics. Based on some assumptions, we have set new rules for message transmission to neighboring nodes. We show that PRoPHET protocol along with our new forwarding strategy can result in much performance improvement in terms of delivered and relayed messages. Learning automata is used to give the optimal threshold value.