Leonel Hernandez Collante, Hugo Hernandez Palma, Aji Prasetya Wibawa, Mario Orozco Bohorquez, Mario Antonio Vergara Sotelo
This project addresses a critical challenge in university telecommunications: managing and preventing network congestion. With the exponential increase in digital service usage and connectivity demands, the University Institution of Barranquilla (IUB) recognizes the necessity of proactively addressing potential disruptions that could affect the quality and continuity of educational and administrative services. To tackle this issue, we propose the development of an integrated artificial intelligence and web visualization system (AI-WEB) designed to predict and mitigate network congestion. The research follows a mixed-method approach, combining quantitative analysis of network traffic data with qualitative assessments of user experience and system performance. Central to our methodology is implementing a predictive model based on Long Short-Term Memory (LSTM) recurrent neural networks, specifically chosen for their ability to capture and analyze temporal dependencies in sequential data. This model processes historical network traffic patterns to forecast potential congestion scenarios, enabling proactive management strategies. Complementing the predictive model, an adaptive and interactive web-based interface has been developed to provide real-time visualizations of network conditions, predictive congestion alerts, and actionable insights for network administrators. This dual approach—leveraging AI-driven forecasting and an intuitive decision-support system—enhances the efficiency of network resource allocation and improves overall user experience. Future work will focus on refining the LSTM model through continuous training with real-time data, integrating additional machine learning techniques to improve prediction accuracy, and expanding the AI-WEB system's applicability to broader telecommunication environments. By establishing a scalable and replicable framework, this project aims to contribute to developing intelligent network management solutions that foster a more efficient and resilient digital infrastructure in educational institutions. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
Faculty of Engineering, University Institution of Barranquilla IUB, Barranquilla, Colombia; Department of Industrial Engineering, Corporación Universitaria Latinoamericana CUL, Barranquilla, Colombia; Faculty of Engineering, Universitas Negeri Malang, Malang, Indonesia; Department of Computer Science and Electronics, Universidad de La Costa CUC, Barranquilla, Colombia; Department of Social and Human Sciences, Universidad Simón Bolívar, Barranquilla, Colombia
