Adaptive Learning System forFlow Control

摘要

The computer network flow controlproblem has been a recurring theme for long time. The main objective has always been to use network resources efficiently where to prevent the data packet loss, which requires retransmission of the lost information. This article presents the theoretical background necessary to understand the flow control, the existing research in this area and a theoretical background of the intelligent systems and the neural networks. The MF-ARTMAP neural network is introduced, which fulfill the conditions of adaptive learning systems. Experiment for usage of the MF-ARTMAP ANN as adaptive learning system for congestion control is proposed.