NEW TRAINING PATTERN SELECTION METHOD FOR ATM CALL ADMISSION NEURAL CONTROL

A new training pattern selection method for adaptive neural control us ing backpropagation learning is presented. When applying this method t o an asynchronous transfer mode (ATM) call admission control, some adv antages are observed: A very small training pattern table is sufficien t, the learning is independent of observed data, and the controller is easily adaptable to traffic changes. The proposed neural control mode l is analysed by computer simulations in heterogeneous traffic environ ments and the results show its effectiveness compared with other metho ds.