The research explores network congestion management within LANs, showcasing a multi-tier system model leveraging SDN, AQM, and AI/ML paradigms for predictive traffic flow control. It critically examines literature on traffic management, highlights gaps like scalability and integration challenges, and introduces a novel intelligent multi-layer approach utilizing advanced predictors like GNNs. Methodologies integrate closed-loop control principles and optimized TCP protocols. Concluding with...