Artificial Intelligence Driven Optimization of Global Supply Chain Performance

Abstract

Global supply chain systems are becoming increasingly complex due to rising customer demand, transportation uncertainty, inventory fluctuations, and operational costs. Traditional supply chain management approaches often struggle to ensure efficiency, responsiveness, and cost minimization in a dynamic global environment. This study aims to explore how artificial intelligence-driven techniques can optimize global supply chain performance through data-based decision-making and predictive analysis. The main objective of the research is to evaluate the effectiveness of machine learning models in improving supply chain classification and performance optimization. For this study, a dataset containing 5,000 rows and 38 variables related to supply chain operations was used. The target variable was derived from total cost and classified into three categories: High, Low, and Medium. Data preprocessing techniques, including missing value handling, categorical encoding, feature scaling, and train-validation-test splitting (70%-15%-15%), were applied. Several machine learning models, including MLP Classifier, Support Vector Machine (SVM), and Gradient Boosting Classifier, were implemented and evaluated using accuracy, precision, recall, F1-score, ROC curve, precision-recall curve, and confusion matrix. The experimental results indicate that the Gradient Boosting Classifier achieved the best performance, with 99.69% training accuracy, 98.40% validation accuracy, and 98.40% test accuracy. It also achieved strong classification metrics across all classes, demonstrating its ability to effectively classify supply chain cost levels. The findings suggest that AI-based models can significantly improve supply chain monitoring, prediction, and optimization. Therefore, the study concludes that artificial intelligence is a highly effective approach for enhancing global supply chain performance and supporting smarter operational decision-making.