A deep learning approach for port congestion estimation and prediction

ABSTRACT

This study proposes high-frequency container port congestion measures based on Automatic Identification System (AIS) data. Vessel movement information of 3,957 container ships from March 2017 to April 2017 is included. The world top 20 container ports’ berth and anchorage areas are identified through Density Based Spatial Clustering of Applications with Noise (DBSCAN) and convex hull methods, and their hourly port congestion statuses are depicted in terms of the traffic volume and turnaround time. The constructed congestion measures overcome the disadvantages of the traditionally used port or industry data, which is heterogenous, behind the time and not easy to obtain publicly. The higher frequency (hourly) of the proposed measures can effectively monitor any slight change in port performance. A Long Short-Term Memory (LSTM) neural network model is then proposed for congestion prediction using constructed congestion measures. Point prediction and sequence prediction are both performed. We innovatively introduce congestion propagation effects into the prediction model as input features. Using Shanghai, Singapore and Ningbo ports as case studies, results show that the inclusion of congestion propagation effect can improve the prediction performance especially for sequence prediction. This study provides significant implications and decision support for container shipping market participants.

KEYWORDS:

• Port congestion measures
• port congestion prediction
• AIS data
• LSTM
• congestion propagation effect

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Though originally, we have collected information of 6320 containerships with MMSI from Lloyd’s list, we can only match 3957 containerships with its AIS data. The reasons for this low number of matches are four folds: 1. Some ships are demolished; 2. Some multi-purpose containerships turn to other markets, such as dry bulk; 3. Some containerships do not provide service during the observation period; 4. New ships may be delivered after 2017.

2. A higher speed limit would result in less accurate berth area recognition.

3. The turnaround time in the berth area represents load-discharge time, while in the anchorage area, it represents waiting time.

4. We also switch the decision sequence of three hyperparameters, e.g., first decide the number of epoch and batch size then choose the number of neurons, the results all support the current selection.

Additional information

Funding

This work was supported by National Science Foundation of China [Project No. 72001123] and Tsinghua University Initiative Scientific Research Program.