LLMs for Supply Chain Management

2025-05-24

AI Summary

Existing large language models (LLMs) exhibit limited performance on supply chain management (SCM) tasks, lack rigorous modeling of strategic interactions among stakeholders, and fail to update classical SCM theories—such as the bullwhip effect—in light of modern dynamics. Method: We propose the first SCM-specific retrieval-augmented generation (RAG) framework, integrating domain-adapted LLMs, multi-agent game-theoretic modeling, and knowledge graph–enhanced reasoning. Our approach achieves deep coupling between RAG and SCM domain knowledge, enabling the first systematic LLM-based analysis of both horizontal and vertical supply chain games. Contribution/Results: The model passes standardized SCM benchmark exams and successfully replicates the Beer Game. RAG improves task accuracy significantly. Game-theoretic analysis not only reproduces established findings but also uncovers three novel classes of cooperation–competition dynamics, thereby extending the theoretical boundaries of the bullwhip effect and advancing SCM theory through AI-driven insight generation.

Technology Category

• Game Theory and Economic Paradigms: Cooperative Game Theory
• Multiagent Systems: Mechanism Design
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

The development of large language models (LLMs) has provided new tools for research in supply chain management (SCM). In this paper, we introduce a retrieval-augmented generation (RAG) framework that dynamically integrates external knowledge into the inference process, and develop a domain-specialized SCM LLM, which demonstrates expert-level competence by passing standardized SCM examinations and beer game tests. We further employ the use of LLMs to conduct horizontal and vertical supply chain games, in order to analyze competition and cooperation within supply chains. Our experiments show that RAG significantly improves performance on SCM tasks. Moreover, game-theoretic analysis reveals that the LLM can reproduce insights from the classical SCM literature, while also uncovering novel behaviors and offering fresh perspectives on phenomena such as the bullwhip effect. This paper opens the door for exploring cooperation and competition for complex supply chain network through the lens of LLMs.

Problem

• Develop RAG framework for dynamic knowledge integration in SCM
• Create domain-specialized LLM for expert-level SCM tasks
• Analyze supply chain competition/cooperation via LLM-driven game theory

Innovation

• RAG framework integrates external knowledge dynamically
• Domain-specialized SCM LLM achieves expert-level performance
• LLMs analyze supply chain competition and cooperation