Project Details
Abstract Arabic
missing
Abstract English
An Automated Network Redesign Tool (ANRT), which consists of Genetic Algorithms (GA)
in its core, is introduced in this research. The presented ANRT is a software program that
was written in C++ and used to input any given network structure with its corresponding link
weights, and will process an output network which relatively has an overall minimum routing
costs within its nodes. In this study, we introduce the “Modified Scale-Free network model”,
which is designed based on both the CISCO 3-layered network structure and the scale-free
network model. The modified scale-free network has been tested and compared to other network
structures (random, small-world and scale-free). It has resulted in significant enhancement in
terms of overall minimum routing costs after applying the redesign process. It showed that the
modified scale-free network is the best choice when designing a network from scratch and will
result in minimum routing costs in comparison to other network structures. On the other hand,
a redesign process is normally applied on an existing network, therefore, the proposed ANRT
is based on the concept of manipulating the existing network structure and links in order to
result in overall minimum routing costs while maintaining the existing structure. In this study,
three network examples (random, small-world and scale-free) have been tested using the ANRT
and the results showed an overall minimum routing costs reduction by 16.56 %, 36.83 %, and
19.28% respectively. The random networks have resulted in smaller improvement due to their
structure, and the scale-free network has become a modified scale-free network after applying
the ANRT. The proposed ANRT has proven to be efficient and accurate in terms of re-designing
any given network while utilizing existing structure and resulting in better overall routing costs.
Status | Finished |
---|---|
Effective start/end date | 1/05/08 → 22/12/11 |
Fingerprint
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.