The advancements in digital technology and the resultant economies of scale offered by high capacity fibre transmission systems have enabled network operators to provide more cost-effective telecommunications services to their increasing customer base. However, with the deployment of such systems, the emphasis on service quality and network reliability has increased considerably. To address this and many other network management issues, the SDH digital transmission standard was established. While supporting enhanced network management functionality, it was only through the subsequent development of highly intelligent Network Elements (NEs) that the inherent cost benefits of SDH have been realised. Each NE is characterised by its own topology and associated restoration schemes. As the economic viability of these survivable topologies are dependent on varying network conditions, it has been shown that the most cost-effective networks will be based on a combination of these topologies. The complexity of these highly integrated networks and the immense capital investment involved suggests that much consideration should be given to the design and optimisation of such networks. This research work concerns an investigation into this particular area of network optimisation. It involves the development of an optimisation design package that facilitates the generation and analysis of survivable SDH networks through the implementation o f highly efficient optimisation techniques. Existing optimisation techniques are utilised to this end, and enhanced upon to address aspects of the SDH design model that were previously not taken into consideration. This results in the development of innovative and highly efficient algorithmic procedures for solving this particular design problem, and "NetOpt" - a prototype design tool for implementing such procedures. Based on a set of generated test case networks, a detailed analysis and evaluation o f these procedures and the resultant network solutions is presented. In conclusion, it is shown that CAD-based tools are essential for the generation and analysis of survivable SDH networks.