Statistical-learning control of multiple-delay systems with applications to ATM networks

Congestion control in the ABR class of ATM network presents interesting cha llenges due to the presence of multiple uncertain delays. Recently, probabi listic methods and statistical learning theory have been shown to provide a pproximate solutions to challenging control problems. In this paper, using some recent results by the authors, an efficient statistical algorithm is u sed to design a robust, fixed-structure, controller for a high-speed commun ication network with multiple uncertain propagation delays.