Utilization of a deep learning-based fuel consumption model in choosing a liner shipping route for container ships in Asia

Abstract Designating the ideal shipping route can spare expenses, enlarge profits and improve the competitiveness of shipping companies. Liner shipping route choice is mainly contingent on fuel cost, which always contributes the major proportion of the ship's operating cost. Although many studies on this topic have been carried out, none are based on the fuel consumption forecast model designed by the advanced machine learning method. This paper provides a platform idea for selecting the optimal operating route for container ships to minimize fuel cost by using an asymmetric traveling salesman problem (ATSP) algorithm solution, in which the fuel consumption model for the route is estimated based on the deep-machine learning method. Five input variables are given in the model including average velocity, sailing time, ship's capacity, wind speed, and wind direction. The mean absolute percentage error (MAPE) of the model is 5.89%, indicating that the predictive result obtains a very high accuracy, close to 95%. The optimal model is thus applied in combination with ATSP to address the optimal solution for a certain route.

[1]  Chuanlei Yang,et al.  Investigation of ANN and SVM based on limited samples for performance and emissions prediction of a CRDI-assisted marine diesel engine , 2017 .

[2]  G. Nagarajan,et al.  Artificial neural network approach to predict the engine performance of fish oil biodiesel with diethyl ether using back propagation algorithm , 2016 .

[3]  Ole Winther,et al.  Statistical modelling for ship propulsion efficiency , 2012 .

[4]  Shan Sung Liew,et al.  Bounded activation functions for enhanced training stability of deep neural networks on visual pattern recognition problems , 2016, Neurocomputing.

[5]  J. Woo,et al.  The effects of slow steaming on the environmental performance in liner shipping , 2014 .

[6]  Hui Zhao,et al.  Study on China-EU Container Shipping Network in the Context of Northern Sea Route , 2015 .

[7]  Wang Hui,et al.  Comparison of several intelligent algorithms for solving TSP problem in industrial engineering , 2012 .

[8]  Ahmet Teke,et al.  The optimized artificial neural network model with Levenberg–Marquardt algorithm for global solar radiation estimation in Eastern Mediterranean Region of Turkey , 2016 .

[9]  Yongtu Liang,et al.  A voyage with minimal fuel consumption for cruise ships , 2019, Journal of Cleaner Production.

[10]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[11]  Deok-Hwan Kim,et al.  Solving local minima problem with large number of hidden nodes on two-layered feed-forward artificial neural networks , 2008, Neurocomputing.

[12]  Qiang Meng,et al.  Robust bunker management for liner shipping networks , 2015, Eur. J. Oper. Res..

[13]  Harilaos N. Psaraftis,et al.  Speed Optimization vs Speed Reduction: the Choice between Speed Limits and a Bunker Levy , 2019, Sustainability.

[14]  Yeong-Dae Kim,et al.  Heuristic algorithms for a multi-period multi-stop transportation planning problem , 2002, J. Oper. Res. Soc..

[15]  J. Thill,et al.  Shipping route choice across geographies: Coastal vs. landlocked countries , 2016 .

[16]  Michael R. Lyu,et al.  A hybrid particle swarm optimization-back-propagation algorithm for feedforward neural network training , 2007, Appl. Math. Comput..

[17]  Young-Tae Chang,et al.  Publisher’s NoteAssessing noxious gases of vessel operations in a potential Emission Control Area , 2014 .

[18]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[19]  Rasmus Rasmussen,et al.  TSP in Spreadsheets – a Guided Tour , 2011 .

[20]  Linying Chen,et al.  Provision of Emission Control Area and the impact on shipping route choice and ship emissions , 2017 .

[21]  Rob J Hyndman,et al.  Another look at measures of forecast accuracy , 2006 .

[22]  Georgios K. D. Saharidis,et al.  The berth allocation problem: Optimizing vessel arrival time , 2009 .

[23]  Maja Škurić,et al.  Ship emissions and their externalities in cruise ports , 2015, Transportation Research Part D: Transport and Environment.

[24]  Osman Turan,et al.  An artificial neural network based decision support system for energy efficient ship operations , 2016, Comput. Oper. Res..

[25]  Mehdi Khashei,et al.  An artificial neural network (p, d, q) model for timeseries forecasting , 2010, Expert Syst. Appl..

[26]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[27]  Adam Sobey,et al.  Physics-based shaft power prediction for large merchant ships using neural networks , 2018, Ocean Engineering.

[28]  Mingyang Zhang,et al.  Data-driven ship energy efficiency analysis and optimization model for route planning in ice-covered Arctic waters , 2019, Ocean Engineering.

[29]  James J. Corbett,et al.  The effectiveness and costs of speed reductions on emissions from international shipping , 2009 .

[30]  Shuaian Wang.  Formulating cargo inventory costs for liner shipping network design , 2017 .

[31]  A. Maragkogianni,et al.  Evaluating the social cost of cruise ships air emissions in major ports of Greece , 2015 .

[32]  Ülkü Alver Şahin,et al.  Estimating of shipping emissions in the Samsun Port from 2010 to 2015 , 2018, Atmospheric Pollution Research.

[33]  R. A. Zemlin,et al.  Integer Programming Formulation of Traveling Salesman Problems , 1960, JACM.

[34]  Wen-Xu Wang,et al.  Detecting hidden nodes in complex networks from time series. , 2012, Physical review. E, Statistical, nonlinear, and soft matter physics.

[35]  F. E. Grubbs.  Procedures for Detecting Outlying Observations in Samples , 1969 .

[36]  Dung-Ying Lin,et al.  Ship routing and freight assignment problem for liner shipping: Application to the Northern Sea Route planning problem , 2018 .