Multi-Agent Learning Path Planning via LLMs

TL;DR

This paper tackles personalized learning path planning by integrating large language models within a multi-agent framework. MALPP leverages three specialized agents—Learner Analytics, Path Planning, and Reflection—guided by Cognitive Load Theory and Zone of Proximal Development to generate interpretable, cognitively aligned paths. Empirical results on the MOOCCubeX dataset show MALPP outperforms baselines in knowledge sequence consistency and cognitive-load alignment, with ablations underscoring the importance of CLT, ZPD, and reflective evaluation. The work advances trustworthy, explainable AI in education and demonstrates a scalable, learner-centered approach to adaptive instruction powered by LLMs.

Abstract

The integration of large language models (LLMs) into intelligent tutoring systems offers transformative potential for personalized learning in higher education. However, most existing learning path planning approaches lack transparency, adaptability, and learner-centered explainability. To address these challenges, this study proposes a novel Multi-Agent Learning Path Planning (MALPP) framework that leverages a role- and rule-based collaboration mechanism among intelligent agents, each powered by LLMs. The framework includes three task-specific agents: a learner analytics agent, a path planning agent, and a reflection agent. These agents collaborate via structured prompts and predefined rules to analyze learning profiles, generate tailored learning paths, and iteratively refine them with interpretable feedback. Grounded in Cognitive Load Theory and Zone of Proximal Development, the system ensures that recommended paths are cognitively aligned and pedagogically meaningful. Experiments conducted on the MOOCCubeX dataset using seven LLMs show that MALPP significantly outperforms baseline models in path quality, knowledge sequence consistency, and cognitive load alignment. Ablation studies further validate the effectiveness of the collaborative mechanism and theoretical constraints. This research contributes to the development of trustworthy, explainable AI in education and demonstrates a scalable approach to learner-centered adaptive instruction powered by LLMs.

Figures (9)

• Figure 1: Multi-agent-based explainable learning path planning framework
• Figure 2: Multi-agent collaboration mechanism for learning path planning
• Figure 3: Prompt design of the learner analytics agent
• Figure 4: Prompt design of the path planning agent
• Figure 5: Prompt design of the reflection agent