Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/13170
Title: Application of neural networks in communication network routing
Authors: Soo, Ah Bok.
Keywords: DRNTU::Engineering::Computer science and engineering::Computer systems organization::Computer-communication networks
DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Issue Date: 1998
Abstract: This project evaluates the application of neural networks for routing in communication networks. Concepts related to routing techniques are first discussed and an overview of neural networks with regard to the model and design is briefly presented. A sample communication network is then used in which each message node is modeled with a neural network. The neural networks are trained and tested with partially known routing information. Routing within the communication network is simulated under fixed and variable traffic conditions, in static and adaptive modes. The results are analyzed for routing accuracy and performance of the neural networks under various traffic conditions. A comparison of the results with those of static routing and those published by other authors is also made. Based on the analysis, it is found that routing with neural networks generally performs not as good as static routing under fixed traffic condition but performs better than static routing under variable traffic condition. The results, however, do not clearly show that routing with neural networks in adaptive mode will improve the routing accuracy significantly as compared to that in static mode.
URI: http://hdl.handle.net/10356/13170
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
SooAhBok1998.pdf
  Restricted Access
Main report10.14 MBAdobe PDFView/Open

Page view(s) 50

227
Updated on Jan 15, 2021

Download(s)

9
Updated on Jan 15, 2021

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.