Thesis Abstracts


An Application of Multivariate Analysis to Time of Day Routing in Telecommunications Networks

Isabelle Smith | December 2000

The work presented in this thesis is a component of a larger project that is currently under way at the Center for Operations Excellence. This project was initiated in May of 1999. The main goal of this project is to help TELUS reduce their yearly investments in the non-toll network by optimizing the current use of their telephone network. We chose to accomplish this by determining a new set of routing rules that would make the most efficient use of their current network. Three main approaches were selected to obtain this new set of routing rules. The first one was the development of a simulation that could serve as a test-bed to compare different sets of routing rules and to determine the optimal set. The second approach was the use of a linear program that would generate routing rules that minimize network utilization. Finally, we have approached the challenge of optimizing the network utilization by using multivariate analysis to help develop routing rules for time of day routing. This last approach is described in this thesis.

The objective of this thesis was to develop a methodology, using multivariate analysis, that would help us find if there were possible alternate routes that could be added to the current ones, which would take advantage of excess capacity in certain regions in the network.

This methodology was developed in SAS and included the use of Principal Components Analysis, Clustering Analysis and the development of an algorithm that would search for adjacent arcs that had sufficient available capacity to serve as alternate routes during certain periods of the day.

It was found that there are alternate routes that can be used during certain times of the day and that these routes can be found by using the methodology described in this thesis. This result is of great value to telephone companies as this means that there is a way for them to use existing capacity more efficiently and therefore avoid or delay investments into adding capacity to the telephone network.


 > Return to Thesis Abstracts