Please use this identifier to cite or link to this item:
https://dspace.upt.ro/xmlui/handle/123456789/7257
Title: | ANN Based Intelligent Congestion Controller for High Speed Computer Networks [articol] |
Authors: | Shajahan, B. Alavandar, Srinivasan |
Subjects: | Congestion control Artificial neural network Quality of service Packet loss Queue length |
Issue Date: | 2018 |
Publisher: | Timișoara : Editura Politehnica |
Citation: | Shajahan,B.; Alavandar,Srinivasan. ANN Based Intelligent Congestion Controller for High Speed Computer Networks. Timişoara: Editura Politehnica, 2018. |
Series/Report no.: | Journal of Electrical Engineering;Vol 18 No 3 |
Abstract: | This paper presents a controller design based on artificial neural networks for the avoidance of congestion in computer networks. The quality of service of networks is not guaranteed by conventional controller, due to the presence of non-linearity and uncertain characteristics. The controller designed with ANN doesn’t require higher model accuracy. The ANN- controller is designed with multilayer neural network mainly to deal with the outgoing data packets, dropped data packets on the bottleneck with constant link capacity, variable link capacity, and variable link delay and with the specified number of transmitted interests, bottleneck band width, and bottleneck delay for a particular interval from various consumers. Around 1 to 6000 interest packets per second have been given and the analytical results for this controller are presented and they are compared with various training algorithms. The results show that the controller designed with artificial neural network is robust, provides faster response and tackles new parameters very well particularly TCP (Transmission Control Protocol) session and round trip time. |
URI: | https://dspace.upt.ro/xmlui/handle/123456789/7257 |
ISSN: | 1582-4594 |
Appears in Collections: | Articole științifice/Scientific articles |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
BUPT_ART_Shajahan_f.pdf | 916.99 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.