Neural network based cell scheduling in ATM node

This letter deals with the queueing discipline for the superposition of arrival processes in the asynchronous transfer mode (ATM) case, particularly for the bursty traffic. A competitive algorithm, based on neural networks, for the cell scheduling is suggested. Computer simulations demonstrate that the new algorithm may be very useful in the bursty traffic case: the cell loss rate is at least one order of magnitude lower than in case of the classical round-robin discipline.