Yanjia Li, Kelcy Kexin Han, Tianrui Hu, Yi-Fan Cao, Huamin Qu, Sicheng Song (2026) — Hong Kong University of Science and Technology, City University of Hong Kong. 📄 Full text (arXiv)
SupplyNet is a gamified visual simulation system that uses a contextual graph-based llm multi-agent framework to model interdependent supply chain dynamics. Designed for professional-training in supply chain management (SCM), it replaces traditional abstract simulations with a manipulable decision space combining an interactive network view, a branching timeline for "what-if" exploration, and a task-oriented analysis console.
Key Features
LLM-driven agents model realistic supplier, manufacturer, distributor, and retailer behaviors, responding adaptively to learner decisions. This moves beyond scripted simulation scenarios to generate emergent, context-sensitive dynamics.
Visual exploratory learning is supported through three integrated components: an interactive network view showing real-time system state, a branching timeline enabling counterfactual comparison, and a task-oriented console for structured performance breakdowns. Together these support causal tracing and comparative reasoning.
User study results suggest SupplyNet increases engagement and supports perceived understanding of supply chain dynamics, demonstrating the potential of pairing contextual multi-agent simulation with visualization for active-learning in operational domains.
Implications for AI in Education
SupplyNet represents a novel application of intelligent-tutoring principles beyond traditional academic subjects into professional education. The system's use of generative-ai agents to create adaptive, responsive simulation environments points toward broader applications in stem-education and professional training where complex systems understanding is required.
Related Pages
- intelligent-tutoring — Core concept page on AI tutoring systems
- llm — Large language models in education
- active-learning — Active learning pedagogical approaches
- professional-training — AI applications in professional training
- generative-ai — Generative AI in educational contexts
- simulation-based-learning — Simulation approaches in education
- visualization — Data visualization for learning
- stem-education — STEM education with AI support