![](/images/graphics-bg.png)
Dynamic load balancing in ad hoc wireless networks using fuzzy neural networks
Other Title(s)
موازنة الحمل الديناميكي في الشبكات اللاسلكية الخاصة باستخدام تقنية الشبكات العصبية المضببة
Dissertant
Thesis advisor
Comitee Members
Marhun, Ali F.
al-Saffar, Ala A.
Abbas, Salim A.
University
University of Basrah
Faculty
Science College
Department
Department of Computer Science
University Country
Iraq
Degree
Master
Degree Date
2013
English Abstract
Load balancing is a primary issue in Ad hoc wireless networks.
When the network has a better load balancing, the multiple routes are simultaneously enabled and data packets are distributed over them.
In ad hoc wireless network many traffic types reach continuously with different rates and in different times.
These rates may be in two cases (low Measured in Kbps and High Measured in Mbps), where each traffic rate needs to link (or node) capabilities coordinate with it, for keeping the throughput of network, when the imbalanced load of traffic among links (or nodes) reduces the throughput of the ad hoc wireless networks.
There is necessary to use a technique to provide load balancing between the links (or nodes) of network to avoid exceeding the arrival traffic rates on the capacity of links (or nodes).
Therefore, load balancing is very necessary.
This thesis presents six suggested methods for solving load balancing problem in ad hoc network.
The proposed six methods are classified in three categories according to what their metrics are dealing with:- Category 1:- Dynamic load balancing of links Category 2:- Dynamic load balancing of nodes Category 3:- Dynamic load balancing of links and nodes These methods indicated above are developed to work under Fuzzy Neural Networks (FNNs) environment.
For achieving dynamic load balancing that is able to deal with the continuous changes in the network.
In the first category, two methods are suggested.
Two fuzzy neural networks are designed to deal with load balancing for links.
One of these FNNs has two inputs (Load of link, Traffic size) and one output (Probability distributed of load the link), but the other has three inputs (Bandwidth of link, Link data size and Traffic size) and two outputs (Load of link and Probability distributed of load the link) .
While in the second category, three methods are suggested to achieve dynamic load balancing on nodes in the network.
The first one deals with the load on the source node and the load on the neighboring nodes to decide the probability distributed the load of neighboring nodes.
While in the second method find the load of the node based on three measurements which are the queue size, throughput of the node and arrival rate of traffic to the node.
At last method contains fuzzy neural system which uses two fuzzy neural networks at same time.
The first fuzzy neural network based on two measurements which are queue length and permitted queue size to result a queue state which is consider one of the inputs in the second fuzzy neural network in addition to throughput of node to result the load of this node.
At last in the third category, only one method is presented to deal with load balancing for each of links and nodes in the network which has two inputs ( Load of Sender node and Load of link) and one output ( Load for the sender or receiver node).
The work in this thesis is applied on three ad hoc wireless networks (AN1, AN2, and AN3).
It is the testing under two different programing environments.
The first, is programming under C++ environment and the second, is the discrete-event simulation system OMNET++.
The results of testing prove the high performance of the six approaches.
The core aim of this thesis is improving the throughput and reducing the traffic load of the network.
Through that, the load balancing for links or nodes in ad hoc wireless network is determined.
Main Subjects
Information Technology and Computer Science
No. of Pages
71
Table of Contents
Table of contents.
Abstract.
Abstract in Arabic.
Chapter One : Introduction.
Chapter Two : Preliminaries.
Chapter Three : Load balancing in ad hoc wireless network using fuzzy neural networks.
Chapter Four : Conclusion and future work.
References.
American Psychological Association (APA)
al-Musawi, Zaynab Sad Karam. (2013). Dynamic load balancing in ad hoc wireless networks using fuzzy neural networks. (Master's theses Theses and Dissertations Master). University of Basrah, Iraq
https://search.emarefa.net/detail/BIM-744412
Modern Language Association (MLA)
al-Musawi, Zaynab Sad Karam. Dynamic load balancing in ad hoc wireless networks using fuzzy neural networks. (Master's theses Theses and Dissertations Master). University of Basrah. (2013).
https://search.emarefa.net/detail/BIM-744412
American Medical Association (AMA)
al-Musawi, Zaynab Sad Karam. (2013). Dynamic load balancing in ad hoc wireless networks using fuzzy neural networks. (Master's theses Theses and Dissertations Master). University of Basrah, Iraq
https://search.emarefa.net/detail/BIM-744412
Language
English
Data Type
Arab Theses
Record ID
BIM-744412