Abstract:
This thesis develops and analyses robustness measures for telecommunications networks. It defines a broad measure of robustness that meets criteria applicable to network managers. A detailed survey of aspects of this measure already researched is presented. Specific numerical and/or analytical procedures are developed to provide information on the duration of system failure which is an important aspect of robustness. The two principal techniques used in this task are stochastic methods and simulation methods. An importance sampling approach is developed to produce better performance when applying simulation methods to network problems. Examples of the implementation of these procedures on network structures are given.
