Author(s): Baviera Miguel; Aguilar Jose; Molines Jorge

Linked Author(s): Jorge Molines Llodrá

Keywords: Port Planning; Container Throughput Forecasting; Artificial Neural Network; Deep Learning; Time Series Forecasting

Abstract: Accurate forecasting of port container throughput is crucial for an efficient and optimal port planning and design. Any error on the throughput prediction during the planning stage could lead into port congestion or an infrastructure underusage. This could result into economic losses as well as an environmental impact. This research explores the application of Artificial Neural Networks (ANN), Long Short-Term Memory (LSTM) and ARIMA on four different ports of the Spanish port system. Total and incremental monthly throughput are estimated. Incremental monthly throughput forecasting provides better results since the non-stationarity component is reduced.

DOI:

Year: 2024