Leveraging machine learning and optimization models for enhanced seaport efficiency

Abstract This study provides an overview of the application of predictive and prescriptive analytics in seaport operations and explore the potential of integrating predictive outputs into prescriptive analytics to advance research in this field. A systematic review of 124 papers was performed to identify and classify key topics based on application areas, types of applications, and employed techniques. Our findings show a growing interest in developing either predictive or prescriptive analytics models to improve seaport operational efficiency. However, there is limited research combining predictive outputs with prescriptive analytics for data-driven decision-making. Additionally, the hybridization of machine learning and operations research techniques remains underexplored. One promising area is applying machine learning models, such as reinforcement learning, to solve optimization problems. Predictive maintenance and data-enabled operational control measures for port equipment and facilities are also highlighted as interesting future research areas.