Forecasting container throughput with long short-term memory networks

Abstract

Purpose Better forecasting always leads to better management and planning of the operations. The container throughput data are complex and often have multiple seasonality. This makes it difficult to forecast accurately. The purpose of this paper is to forecast container throughput using deep learning methods and benchmark its performance over other traditional time-series methods. Design/methodology/approach In this study, long short-term memory (LSTM) networks are implemented to forecast container throughput. The container throughput data of the Port of Singapore are used for empirical analysis. The forecasting performance of the LSTM model is compared with seven different time-series forecasting methods, namely, autoregressive integrated moving average (ARIMA), simple exponential smoothing, Holt–Winter’s, error-trend-seasonality, trigonometric regressors (TBATS), neural network (NN) and ARIMA + NN. The relative error matrix is used to analyze the performance of the different models with respect to bias, accuracy and uncertainty. Findings The results showed that LSTM outperformed all other benchmark methods. From a statistical perspective, the Diebold–Mariano test is also conducted to further substantiate better forecasting performance of LSTM over other counterpart methods. Originality/value The proposed study is a contribution to the literature on the container throughput forecasting and adds value to the supply chain theory of forecasting. Second, this study explained the architecture of the deep-learning-based LSTM method and discussed in detail the steps to implement it.

Forecasting and planning during a pandemic: COVID-19 growth rates, supply chain disruptions, and governmental decisions

“…We used the death and recovery rates as the independent variables for the multivariate forecasting methods. We calculated the Mean Absolute Scaled Error (MASE) and the Symmetric Mean Absolute Percentage Error (SMAPE) for each iteration ( Makridakis et al, 2020 ; Shankar, Ilavarasan, Punia & Singh, 2019 ). We calculated the relative errors by dividing with the corresponding error from the naïve method ( Punia, Nikolopoulos, Singh, Madaan & Litsiou, 2020 ).…”

Deep learning-based container throughput forecasting: a triple bottom line approach

“…Because of its potential to extract temporal characteristics of the time series, it is used widely for univariate time series analysis (Bandara et al, 2020). It is observed to exhibit better prediction accuracy for univariate container throughput forecasting when compared with traditional time series methods like ARIMA, Holt-Winter's, ETS, TBATS and popularly used machine learning methods like neural network (NN) and hybridized ARIMA-NN (Shankar et al, 2019). Though there is no other literature available on CT forecasting using LSTM, but the LSTM is popularly used in otherdomainslikeenergy (Wangetal.,2019a,b;Zhouetal.,2019),finance (Altanetal.,2019;Caoetal., 2019;Wu et al, 2019), transportation (Tian et al, 2018;Zhao et al, 2019) and weather (Qing and Niu, 2018;Salman et al, 2018;Yu et al, 2019).…”

“…3.4.3 Proposed ARIMAX-LSTM network. Complex deep learning techniques like LSTM are observed to be very promising for sequential learning problems (Punia et al, 2020a;Shankar et al, 2019). For multivariate CT forecasting, the vector correction models are observed to be widely used either as a stand-alone technique or in combination with averaging models like ARIMA.…”

“…The managerial decisions, like scheduling, inventory optimization, etc., are largely dependent on forecasts (Harrison and Fichtinger, 2013). Moreover, the container and shipping industry is a billion-dollar industry that is believed to experience huge cost benefits with little improvement in port management (Shankar et al, 2019). As a result, the forecaster is keen to develop a forecasting method with improved accuracy.…”

“…Albeit, these methods have been accepted and implemented widely in the recent literature, but their implementation in operations management, particularly container supply chain, is still open for exploration. The deep learning-based long short-term memory network (LSTM) method is used to forecast CT for the Port of Singapore (Shankar et al, 2019), but the influence of additional factors on CT forecasting is open for investigation. This study is an attempt to leverage the potential of deep learning to develop a decision model for CT prediction.…”