Backward predictive congestion control notification in ATM networks using neural network prediction
Watit Benjapolakul, Aimaschana Niruntasukrat, P. Na nagara · 2002
This paper presents neural network predictor used for traffic control in the Backward Predictive Congestion Control Notification (BPCN) scheme in Asynchronous Transfer Mode (ATM) network. The purpose of this study is to compare with the performance of traffic controllability using a Recursive Least Square (RLS) predictor. According to the results, the loss ratio of the traffic is reduced to 18% and the transmission delay is reduced to 88% compared with the cases of the RLS predictors.