Global Conference on Mathematics and Applications (GCMA)

This final-year research project explores the application of graph theory and probabilistic graphical models to simulate attack propagation in telecommunication networks. The theoretical framework covers Bayesian and Markov networks to model uncertainty, as well as dynamic graphs to capture temporal evolution. A practical case study is conducted on Algérie Télécom’s FTTH (Fiber to The Home) network in the Metidji area. By modelling the GPON (Gigabit Passive Optical Network) architecture as a weighted directed graph, the Max-Flow, Min-Cut theorem is applied to evaluate network robustness and performance. The findings pinpoint critical failure points (such as the OLT-SRO link) and identify bottlenecks. Finally, the study recommends physical maintenance strategies and logical security enhancements to mitigate the impact of cyberattacks and technical failures.
KeyWords. Graph theory, Bayesian networks, Dynamic graphs, Max-flow, Min-cut, FTTH, Network security, Algérie Télécom.