Large Language Model-Driven Demand Forecasting and Inventory Optimization for University Physical Education Resource Supply Chain

This study proposes an intelligent management system for university physical education resource supply chains based on Large Language Models (LLMs), aiming to address the problems of inaccurate demand forecasting and inefficient inventory management in traditional physical education resource allocation. By constructing a deep learning framework incorporating LLMs and combining multi-dimensional information including historical data, seasonal factors, course schedules, and student preferences, precise demand forecasting for sports equipment, facilities, and teaching resources is achieved. The research employs a pre-trained language model based on the Transformer architecture, combined with time series analysis and reinforcement learning algorithms, to develop dynamic inventory optimization strategies. Experimental results demonstrate that compared to traditional methods, this system improves demand forecasting accuracy by 23.7%, increases inventory turnover rate by 31.2%, and achieves a resource utilization rate of 89.6%. This research provides a novel solution for intelligent management of university physical education resources, offering significant theoretical value and practical implications.OPEN ACCESS Received: 24/08/2025 Accepted: 28/10/2025 Published: 03/02/2026