Ms. Kaoutar Boussaoud, software developer and researcher at SCC, successfully defended her PhD dissertation on June 12, 2025, at the National Institute of Posts and Telecommunications (INPT) in Rabat, Morocco.
Her dissertation, conducted within the SEED research team of the STRS Laboratory (CEDOC2TI), presents a comprehensive data-driven approach to enhancing Software-Defined Networking (SDN) using machine learning and multi-agent reinforcement learning (MARL). Ms. Boussaoud developed advanced techniques for congestion detection and traffic optimization, including supervised flow classification, federated learning for distributed detection, and a Dueling DQN-based multi-agent framework that enables both proactive congestion control and adaptive bandwidth allocation. Her work further explores dual-agent architectures that integrate failure recovery strategies, increasing the resilience and robustness of SDN infrastructures under dynamic and large-scale network conditions.
At KIT, Kaoutar Boussaoud is part of the BMBF project DAPHONA, where she develops web services and designs algorithms for scientific data in the field of nanophotonics. Her work bridges innovative technologies and applied data science for scientific discovery.
SCC congratulates Kaoutar Boussaoud on the successful completion of her doctorate and wishes her all the best for her future career.
Dissertation of Kaoutar Boussaoud: Software-Defined Networks – Enhancing Traffic Management using Machine Learning