ResearchPod Summary
Supply chain management (SCM) education often relies on simulations to teach complex concepts like inventory oscillation and disruption propagation. However, traditional simulation tools frequently present data in abstract, static tables that fail to capture real-world complexity. This disconnect between numerical data and operational reality often leads to student disengagement and hinders the development of intuitive decision-making skills. The authors conducted a formative study with students and instructors to identify these barriers, noting a lack of contextual feedback, limited support for counterfactual exploration, and opaque agent behavior.
To address these challenges, the researchers developed SupplyNet, a gamified visual simulation environment. The system is built on a contextual, graph-based multi-agent framework powered by Large Language Models (LLMs). Unlike traditional rule-based simulations, SupplyNet allows agents to exhibit human-like reasoning and adapt to varying business contexts. The system provides a manipulable decision space through three primary components:
A user study involving university students demonstrated that SupplyNet significantly improves learning engagement and perceived understanding compared to traditional baseline simulations. Participants reported that the ability to visualize causal propagation and compare parallel decision paths made the learning process more active and reflective. The findings suggest that integrating LLM-driven multi-agent simulations with interactive visualization tools can transform SCM education from passive observation into a dynamic, exploratory experience.
AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.