A neuro-adaptive TCP-like protocol with cost constraints

In this paper we design a neuro-adaptive congestion control system, via regulating the packet round trip time for each user application, without violating a maximum cost per packet value, prescribed by the user. Reliable desired round trip values are provided via a neural net estimator, which consequently are utilized for the design of a neural network rate controller. To improve the achievable application sending time, we introduce the notion of communication channels, whose number is kept constant throughout transmission, that operate in parallel (virtual sources). We assume the nonexistence of uncontrollable traffic flows (e.g., UDP) as well as negligible propagation delays. Simulation studies illustrate the approach