Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/38990
Title: The application of artificial neural network for optimization of MP-CSMA/CD protocol
Authors: Jiao, Zhi Hua
Keywords: DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
Issue Date: 1997
Abstract: In heavy traffic, the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) protocol suffers from numerous packets collisions resulting in a degradation of perfor-mance. A modified p—persistent CSMA/CD protocol(MP-CSMA/CD) has been proposed earlier which aims to maximize throughput performance. In this project, an artificial Neu-ral Network(NN) is utilized to optimize the MP-CSMA/CD protocol. The effects of neural network configurations and training parameters including learning rate, momentum and hidden neurons on neural network training are investigated. The simulation results show that the general throughput performance of neural network controlled MP-CSMA/CD local area network is better than that of CSMA/CD. In addition, the performance of the MP-CSMA/CD(NN) protocol under different load distributions (Even or Uneven load) is investigated. Some distribution functions are used to distribute the traffic along the bus to simulate actual traffic in the LAN. To ascertain the feasible implementation of this protocol, the effects of packet propagation delay are examined. Packet propagation delays may result in a drift in the probability of transmission due to the difference in sampled throughputs at different stations. Our simulations show that the trained neural network is insensitive to this noise in the sampled throughput and is able to steer the probability p in even or uneven load.
URI: http://hdl.handle.net/10356/38990
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
JiaoZhiHua1997.pdf
  Restricted Access
Main report10.92 MBAdobe PDFView/Open

Page view(s)

227
Updated on Dec 1, 2020

Download(s)

2
Updated on Dec 1, 2020

Google ScholarTM

Check

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