Far East Journal of Experimental and Theoretical Artificial Intelligence
Volume 3, Issue 1, Pages 21 - 38
(February 2009)
|
|
OPTIMIZING DESIGN OF COMPUTER NETWORKS USING MULTI-OBJECTIVES LINEAR PROGRAMMING AND ARTIFICIAL INTELLIGENCE TECHNIQUES
Taghrid Abdel Rahman Imam (Egypt)
|
Abstract:
This paper introduces a package based Multi-objectives Linear Programming (MOLP) and Artificial Intelligence (AI) techniques to optimally design computer networks under given performance attributes. The objectives of the MOLP/AI package are to minimize delay (cost or distance), and to maximize the traffic flow and the network reliability. The developed software package is oriented towards network optimization. It has a very well developed graphical user interface. It is based on several fixed types of famous network design algorithms (shortest path, max-flow, min cost flow, network simplex etc.). It is directed mainly to solve problems on graphs and set of optimization problems on networks (e.g., shortest path, max flow).
Tabular representation is realized by adjacency matrix where only one parameter may be specified, which is insufficient for network optimization. For illustration, the package has been implemented on a message switching public data network. |
| Keywords and phrases: multi-objectives linear programming, reliability optimization, modeling networks, shortest path and artificial intelligence (AI). |
| Communicated by Shun-Feng Su |
| Number of Downloads: 145 | Number of Views: 585 |
|
