Enhancing Supply Chain Decision-Making through Machine Learning and Mathematical Modelling Approaches

Abstract

Supply chain optimization is important in improving operational efficiency, lowering costs, and optimizing decision-making processes. This study investigates the combination of machine learning and mathematical modelling in supply chain management through Python programming. Some data-driven strategies, such as demand forecasting, supplier performance evaluation, defect rate prediction, and inventory optimization, were applied to improve supply chain efficiency.

Primary Python libraries including Pandas, NumPy, and Matplotlib were employed to process data and visualizations. Random Forest Regression model was utilized to make demand predictions with moderate success. K-means clustering was used to measure the performance of suppliers through defect rates and lead times and show inefficiency in specific clusters. A Random Forest Classifier was used to forecast defect rates, though only moderately so, i.e., it would need to be tuned. Linear Programming was applied to maximize the level of stock so that cost would be minimized without running out of stock.

The results highlight the importance of mathematical modeling and machine learning in supply chain decision-making. Improving predictive models and the use of real-time data in dynamic supply chain management is the future research that must be conducted.