A novel learning based solution for efficient data transport in heterogeneous wireless networks

There has been a spectacular growth in the use of wirelessnetworks in recent times and consequently, adapting TCP to the wirelessnetworks is a hot topic of current research. However, most of theexisting works proposed for this problem have been designed for specificwireless networks, or they necessitate changes at either the receiver, atthe intermediate nodes, or at both, because of which their deploymentbecome very difficult. Therefore, here we propose a TCP variant whichworks over both multi-hop ad hoc wireless networks as well as singlehopwireless networks. We use a learning based method to dynamicallychange the congestion window size according to the network conditions.Our protocol does not rely on any explicit feedback from the networkand requires only sender-side modifications. Through extensive simulationswe show that our protocol achieves the performance improvement,in terms of goodput, packet loss, and fairness to the competing flows.