AI-Enhanced Supply Chain Management in Agriculture

. This study introduces a comprehensive AI framework aimed at enhancing agricultural supply chain management via the combined utilization of four machine learning models. The system utilizes a Long Short-Term Memory (LSTM) network for accurate demand forecasting, a Reinforcement Learning (RL) model for dynamic logistics optimization, a Convolutional Neural Network (CNN) for automated quality control through computer vision, and an XGBoost-based ensemble for real-time decision support. The components are integrated into a singular, user-friendly web application developed with Flask and underpinned by a MongoDB database. The system effectively resolves significant inefficiencies, including demand-supply discrepancies, logistical delays, and post-harvest losses. Experimental results indicate substantial performance enhancements, with the LSTM model attaining a R² score of 0.89, the RL model realizing an 18.7% reduction in costs, the CNN model achieving 97.3% accuracy in defect detection, and the decision support system delivering actionable recommendations with 92.5% accuracy. This study demonstrates that a comprehensive, AI-driven