Circle announces the
development of an advanced Prediction System for optimising operations at Ro-Ro marine
terminals. This innovation represents a significant step towards more efficient and
automated management of port logistics activities, reducing trailer dwell times and improving
operational planning.
The process is developed in several steps, starting with a descriptive analysis to examine data
distributions. Next, pre-processing is carried out, which includes the handling of missing or
inconsistent values to ensure the accuracy and quality of the information. Next, advanced
predictive models such as K-Nearest Neighbours (KNN) and Deep Neural Networks (DNN) are
trained. Finally, the testing and consolidation phases take place to validate and optimise the
performance of the models.
The system integrates a prediction process in order to optimise logistics operations over the
required time horizon. Through the processing of key data based on historical and real-time
data, such as the direction and loading status of trailers, the presence of dangerous goods,
the type of refrigeration, the origin, destination and ETA of the ship, Circle is able to estimate
the expected dwell time within the terminal, providing strategic information for the planning
of operations