Date of Award

2002

Document Type

Master Thesis

Degree Name

Master of Engineering (Research)

Department

Electronic Engineering

First Advisor

Dr. Dirk Pesch

Abstract

Due to the growth of the subscriber base and the introduction of new services, mobile communications networks’ demand for wireless resources is increasing. This has resulted in an increasing need to change the frequency allocation of a network in order to adapt the cellular network to new frequency demands.

Currently, a planning tool generates a new frequency plan, which is. then deployed to cell sites. The deployment should ideally be carried out with minimum disruption to network operation, which makes this deployment process a critical undertaking as any errors during or after the deployment can leave the network with poor or no service.

Today, two methods are used to carry out this deployment. Binary download, which requires to shut down parts of the cellular system in order to download the frequency database, and sequencing deployment, where only the cell being modified will be shut down. To shut down a large part of the network is becoming more and more undesirable as operators require 99.999% up- time and users expect service availability around the clock, in particular for emergency related services. On the other hand, sequential deployment reduces this disruption but deploys the frequency plan not in a structured fashion but rather randomly.

This work presents two different cell sequencing schemes, a sequential scheme and a parallel scheme, for on-line frequency deployment which solve the deployment problem for minimum disruption, in conjunction with a protocol that will control the transition between plans. Genetic Algorithms have been used as a stochastic optimisation technique to obtain the optimum cell sequence for deployment. The effectiveness of the proposed schemes is demonstrated through extensive computer simulation and results are compared with existing techniques.

Access Level

info:eu-repo/semantics/openAccess

Share

COinS