Machine learning models integrated with strategic theories for predicting delivery time in supply chains

The delivery time of goods in the oil and gas supply chain is a critical enabler for achieving overall supply chain excellence and business success. However, the uncertainty of the delivery times of goods is a significant issue, making it critical for businesses to develop effective logistics and shipping strategies. Consequently, this paper proposes machine learning models to predict delivery times of the items in oil and gas supply chain and enhance its responsiveness. The proposed approach integrates predictive analytics with the Resource-Based View (RBV) and Dynamic Capabilities View (DCV) to align operational forecasting with strategic objectives. Key factors influencing delivery time are identified through a combination of literature review and expert input. Several machine learning models are trained and tested using real-world oil and gas supply chain data. The results indicate that transportation mode, item complexity, and supplier location are the most influential predictors of delivery time. Among the evaluated models, the ensemble approach demonstrates the best performance, achieving prediction accuracy exceeding 85% and exhibiting strong generalization capability. The proposed models equip supply chain managers with actionable decision-support tools to enhance scheduling, reduce uncertainty, and improve inventory planning and logistics coordination for better disruption response. From a strategic perspective, integrating machine learning with RBV and DCV strengthens organizational responsiveness and supports sustained competitive advantage in dynamic supply chain environments.