Adaptive fuzzy control for buffer loading regulation of network node

Telecommunication Systems and Computer Networks

The goal of this study is in developing a method for controlling the transmission of information packet traffic. The control is based on buffer loading regulation of network node. The fuzzy controller is used for regulation of buffer loading. The operating principles of the controller combine fuzzy logic theory methods and fuzzy neural networks theory. The controller relies on a system of fuzzy logic to calculate the probabilities of marking a packet with a buffer overload flag. After this the packet is sent to the destination node. After receiving a packet with a buffer overload flag, the destination node sends the command to reduce the traffic intensity to sending node. When the controller starts to work incorrectly, the adaptation algorithm is used. The adaptation algorithm is based on using the fuzzy neural network that processes information about the control result of information packet traffic transmission. After this the fuzzy neural network generates commands for adjusting the controller. We have verified the effectiveness of the study’s results, finding a reduced probability of loss of information packets and an increase in the level of using channels in the data network.