Generated Core Network Planning and Characterization
The study of optical core networks often relies on a small number of topologies available in the state of the art. However, many relevant scenarios, including less studied countries or regions, do not have publicly available reference topologies. This limits the ability to perform justified studies, especially when large datasets are needed for machine learning or artificial intelligence applications. A possible solution is to use artificially generated topologies, but this requires a rigorous validation process to determine whether they can serve as credible substitutes for reference networks.
Our group has a topology-generation methodology that allows the creation of realistic network topologies from country-related information. The aim of this thesis is to create different sets of topologies by changing the parameters before generation and to analyze how these changes affect the different layers of the networks. This characterization will make it possible to create more tailored topologies that use different parameter sets based on the selected area.
