Dynamic Adaptation in an Intelligent Multi-Tutoring System: A Multi-Agent Approach
Zakaria Rida, Hadhoum Boukachour, El Yadari Mourad, Mustapha Machkour · IEEE Access · 2025
Online learning provides new educational opportunities but continues to face high dropout rates compared to traditional classroom settings. To address this issue, we propose IMTS (Intelligent Multi-Tutoring System), an intelligent tutoring system that an intelligent tutoring system that integrates both multi-tutoring and multidisciplinary capabilities. IMTS synergistically integrates automated tutoring (intelligent agents) and human tutoring (teachers and peers) to deliver personalized support to learners. The proposed architecture is based on a multi-agent system (MAS) that dynamically adapts pedagogical interventions according to the learners’ profiles and needs. By analyzing learning progress in real time, the system continuously adjusts its recommendations and guides the student toward the most appropriate support (educational resources, peer interaction, or teacher intervention). IMTS was developed using the JADE platform and integrated into a customized Learning Management System (LMS). An experiment conducted with students demonstrated improved academic performance, increased engagement, and reduced dropout rates. These findings suggest that IMTS represents a significant advancement in online education by enabling more interactive, responsive, and personalized tutoring.