This PhD thesis considers some of the more emerging problems in network modelling, namely the design of survivable hierarchical networks, <i>Traffic Engineering </i>(TE)<i> </i>and generally traffic management in survivable multi-service networks with <i>Quality of Service </i>(QoS) prerequisites and the planning of wireless access networks. So, in the context of the research work presented in this thesis:- Novel survivable hierarchical network design, wireless access network planning and traffic management techniques were developed. These techniques involve optimisation methods based on <i>Linear Programming </i>(LP) and <i>Integer Linear Programming </i>(ILP), as well as heuristic methods based on <i>graph theory </i>and <i>computational intelligence (genetic optimisation </i>and <i>simulated annealing). </i> A unified framework for off-line TE, on-line/dynamic routing and path restoration (facility restoration) that can be used in survivable multi-service QoS networks was also developed. Existing traffic management techniques were improved so that to support advanced QoS and survivability characteristics. At first, the objectives of this project are presented followed by a brief analysis of the problems encountered in the network design process. Next, the new methods for designing survivable hierarchical networks are analytically described followed by the developed wireless access network design techniques. After that, the novel traffic management methods and the aforementioned framework, developed in the context of this thesis are presented. Test results are provided together with most of the developed methods. The test results actually indicate that the developed methods can efficiently solve small, medium or even large problems, all developed methods are computationally tractable and the performance of the developed heuristic method is very close to this of the corresponding LP and ILP optimisation methods. The new heuristic methods are solved in a fraction of the time (less than 30%) that the equivalent optimisation methods are solved. Note that the specially developed design and simulation software tool <i>NetLab </i>was used in order to test and evaluate the new design and traffic management methods. Finally, a summary of the work carried out and the results achieved is presented followed by the conclusions and suggestions for further work.
|Publisher||University of Portsmouth|
|Source Sets||Ethos UK|
|Type||Electronic Thesis or Dissertation|
Page generated in 0.0113 seconds