Container Terminal Workload Modeling Using Machine Learning Techniques

Abstract

Container terminals are complex facilities which serve to maritime transportation. Effectivity and productivity of the terminals are very crucial for international transportation. Often, terminal managers deal with logistic problems with high costs. Reducing operation cost and time are the key elements for a smooth and secure process. To minimize the port stay time of the vessels, each part of the operation should be optimized. In order to serve container vessels efficiently, the port management should analyze the cargo handling process considering the ship and container characteristics. In this scope, the purpose of this study to model quay crane handling time in traditional container terminals based on port operations data using machine learning techniques. To conduct operational efficiency analysis, we have analyzed container terminal operations using alternative regression models based on operations data of more than 400.000 handling movements in a traditional terminal in Turkey. Results suggest that the efficiency of the terminal can be increased in planned rush hours such as the periods before the mealtime and the vessels’ operation completion.