Optimizing Decision-Making in Supply Chain Management Using Machine Learning and Mathematical Modeling Techniques

Abstract

Supply Chain Management (SCM) is advancing through the incorporation of machine learning and mathematical modeling methods, improving decision-making effectiveness. The aim of this study is to improve the supply chain functions through four algorithms: Long Short Term Memory (LSTM), Support Vector Machine (SVM) Genetic Algorithm (GA), and Reinforcement Learning (RL). A result of the forecasting experimental data shows that LSTM achieved 94.3% accuracy forecast which exceeds SVM’s 87.8%. GA provides 23% improvement in savings with respect to conventional optimization techniques and RL brings a 15% increase in delivery efficiency through real-time decision making. Comparison of these techniques with existing methods reveals the greater efficiency of hybrid models that combine information from AI with more traditional means. In addition, the research evaluates how blockchain and digital twin technologies improve transparency, and security in supply chains. Despite these issues, the results show that supply chain efficiency can be markedly improved using AI powered decision making. As future studies, combined optimization frameworks and supply chain management ways that focus on sustainability should be investigated to have robust and environmentally friendly supply chains. By arguing that smart supply chain decision needs to be based on empirically validated theoretical assumptions, this research sheds important light on how scholars and industry experts can try to improve supply chain performance through smart decision making.