Congestion control of ATM networks using multilayer neural network approach: multiple source/single switch scenario

This paper proposes an adaptive control methodology using neural networks (NNs) for the available bit rate service class in an ATM network. Adaptive methodology is developed to control traffic where sources adjust their transmission rates in response to the feedback information from the network switches. Specifically, the ATM traffic at a given switch is modeled as a nonlinear discrete-time system, and a two-layer NN controller is designed to prevent congestion. Tuning methods are provided for the NN to estimate the unknown traffic. Mathematical analysis is given to demonstrate the stability of the closed-loop system so that a desired quality of service (QoS) can be guaranteed. No learning phase is required for the NN and initialization of the network weights is straightforward. Simulation results are provided to justify the theoretical conclusions for multiple source/single buffer scenario with and without feedback delays present.