Date of Award

1999

Degree Type

Thesis

Degree Name

M.Sc.

Department

Computer Science

First Advisor

Bandyopadhyay, S.

Keywords

Computer Science.

Rights

CC BY-NC-ND 4.0

Abstract

The design process of the ring routed wavelength optical network encompasses the search for a layout order of nodes such that the total overall traffic is minimized in terms of distant communication and traffic frequencies. The optimal solution is a member of a massive domain set reaching sizes in the order of n!, where n is the number of nodes in the network; typically n = 100. A brute force, linear search algorithm can be implemented, but when executed in search for an optimal solution, the algorithm is a computational challenge as it becomes time consuming and renders itself unfeasible in realistic design time criteria. In researching a suitable search algorithm with the flexibility to adapt to changing network traffic parameters and render a near optimal solution in reasonable design time constraints, the genetic algorithm presents a candidate solution to be tested. The goal of this thesis is to implement a custom genetic algorithm and examine its potential in reaching near optimal solutions in the optical design application framework. The course of work was divided into three major development stages. In the first stage, a simple object oriented GA model was developed (SGA). Then the model was customized to the ring routed network design application, known as the Simple Optical Genetic Algorithm (SOGA) and finally the revised algorithm is implemented and termed the Optical Genetic Algorithm (OGA). In smaller networks (up to 12 nodes) the GA is compared to a brute force linear algorithm to test its performance. For larger networks, the GA was compared to a random search algorithm to test its effectiveness. In both cases, the GA has shown to surpass the other algorithms in generating a pool of near optimal solutions in reasonable time constraints. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1999 .K63. Source: Masters Abstracts International, Volume: 40-03, page: 0724. Adviser: Subic Bandgopadhyay. Thesis (M.Sc.)--University of Windsor (Canada), 1999.

Share

COinS