Journal of Advanced Transportation

Research Article

Machine Learning for Promoting Environmental Sustainability in Ports

Table 1

Problems and ML techniques for promoting environmental sustainability in maritime port logistics.

Articles

Areas

Research scope

Environmental problems

Machine learning

Data collection methods

Case studies

Overall port

Seaside

Yard

Landside

Emission

Water pollution

Noise pollution

Energy saving

Renewable energy

Solid waste

Input

Techniques (tools)

Output

[24]

●

Port performance evaluation

●

Port assets, berth quantity, and geographical location

Net profit, cargo throughput, and NOx emissions

Secondary data sources

17 Chinese ports

[28]

●

Port performance evaluation

●

Facility, vessel and other pollution incidents

KDE

A smoothed graph for the distribution of pollution incidents probability density

Secondary data sources

10 American ports

[49]

●

Truck scheduling

●

Historical data, truck arrival time, administrative waiting start and end time; intermediate waiting start and end time; node-specific forecasting parameters, e.g., dispatching modes and storage policies; and external forecasting parameters, weather and traffic information

NN (BP)

Waiting time, arrival rates that translates into a reduction of traffic congestion and air pollution

Secondary data sources

An empty container depot in Northern Germany

[50]

●

Truck scheduling

●

Number of clusters and the archive containing n solutions

K-means

Cluster centroids

Secondary data sources

Port of Hamburg (Germany)

[31]

●

Port performance evaluation

●

Berth length, the number of cranes, terminal area for the efficiency estimation. City gross domestic product, variance inflation factors, and emissions control regulations

TEUs handled and the impact of emissions control regulations

Secondary data sources

48 ports in Europe

[46]

●

Yard crane demand

●

The average of the previous day load, the average of the previous week load, the same hour load for previous day, and the previous hour load

NN (BP) and SVM

RTG crane demand of one hour

Primary data sources

Port of Felixstowe in the UK

[40]

●

Berthing

●

Maximum continuous rate (MCR) measured by megawatt, shaft speed

Emissions (NO_x, SO_x, CO₂, and CO)

Primary data sources

Two ocean-going vessels in Australia

[48]

●

Antiswing crane

●

Trolley position, trolley speed, loading angle, and angular velocity

NN (ANFIS)

The driving force of the trolley

Secondary data sources

—

[29]

●

Port performance evaluation

●

Number of quay crane, acres, berth and depth, undesirable output (CO₂), and desirable outputs (calls, throughput, and deadweight tonnage)

SOM

Clusters of decision-making units (DMUs)

Secondary data sources

20 American container ports

[41]

●

Berthing

●

Ship identification, position, speed, course, heading and navigational status, and timestamp

Emissions (NO_x, SO₂, PM_2.5, VOC, CO, NH₃, CO, N₂O, and CH₄)

Primary data sources

Ports of Newcastle, Jackson, Botany, and Kembla in Australia

[44]

●

Berthing

●

Observed video

PCA

The background with low-rank property and the foreground with sparse property

Primary data sources

Unknown

[37]

●

Port performance evaluation

●

RTGC number, block number, handling container specification, stevedoring full or empty category, handling volume for a task, and the number of clusters

K-means

Resource allocation for container terminals

Secondary data sources

A container terminal on the east coast of China

[42]

●

Berthing

●

The net tonnage, deadweight tonnage, actual handling volume, and efficiency of facilities

GBoost, RF, NN (BP), PR, and KNN

Energy consumption

Secondary data sources

Jingtang port (China)

[25]

●

Port performance evaluation

●

CO₂ emission driver factors of the city where the port is located are gross domestic product, total resident population, the number of port berth, total imports, total exports, the first industrial value, the secondary industrial value, the primary industrial value, gross industrial production, fixed assets investment in the tertiary industry, per capita income, railway freight volume, highway freight volume, and waterway freight volume

Spatial clustering

Clusters of similar ports in terms of environmental sustainability (LISA cluster maps of PCD carbon emissions)

Secondary data sources

30 Chinese container ports

[32]

●

Port performance evaluation

●

Number of berths, the length of the terminal, the number of staff, and the total fixed assets

Cargo throughput, NO_X emissions, SO_X emissions, and solid waste containers

Secondary data sources

18 Chinese ports

[27]

●

Port performance evaluation

●

Highly correlated input variables

PCA

Determinant factors of the survey are lean management, green operational practices, green behavior (green participation and green compliance), and green climate

Survey

Kaohsiung container port (Taiwan)

Determinant factors of the survey are lean management, green operational practices, green behavior (green participation, green compliance), and green climate

Green performance (financial and nonfinancial)

[45]

●

Noise of moving ships in port areas

●

Draught, speed, and ship-to-microphone distance

Sound emitted

Primary data sources

Industrial port of Livorno (Italy)

[51]

●

Truck scheduling

●

Container features are cycle, type, weight, special (e.g., hazard shipping), agreement (between stakeholders), vessel departure time, distance (of two containers in the yard), customs clearance, dwell time, and final destination

Hierarchical clustering

Container groups

Secondary data sources, survey

Port of Altamira (Mexico) and Port of Genoa (Italy)

[30]

●

Port performance evaluation

●

Energy consumption and number of employees

Hierarchical clustering, PR

Total gross weight of goods, air pollutant emissions, and the rank of ports in terms of eco-efficiency

Secondary data sources

24 European container ports

[36]

●

Indoor air quality prediction (RORO)

●

CO concentration and load (number of cars)

NN (BP)

The reference flow rate of the ventilation system

Secondary data sources

A liner between Egypt and Saudi Arabia ports

[38]

●

Berthing

●

ETA features (date, time, and weekday) and ship features (ship type and length)

SVM

Arrival time of vessels

Secondary data sources

—

[33]

●

Air quality prediction

●

Fine particulate mass and fine particulate composition

Air quality

Primary data sources

Long Beach (US)

[12]

●

AGV

●

Scheduled arrival, departure, and load/unload start time, planned berthing place, planned position of front and rear of the ship, and number of containers to load and unload

NN (BP)

Availability of AGVs

Primary data sources

Hamburg container terminal (Germany)

[52]

●

Container truck emissions

●

Highly correlated data of traffic and particle number concentrations (PNC)

PCA

Principal components (container truck volume, other vehicles volume, and PNC data)

Secondary data sources

Waigaoqiao port (China)

[34]

●

Air quality prediction

●

Type of pollutant, the operating mode, and gross tonnage of ships

Emissions (SO₂, NO_x, CO₂, VOC, PM, and CO)

Secondary data sources

Ports of Ambarlı, Izmir, Mersin, and Kocaeli (Turkey)

[47]

●

RTG crane

●

Energy consumption of hoist, gantry, and trolley

General energy consumption of RTG

Secondary data sources

Casablanca port (Morocco)

[35]

●

Air quality prediction

●

Meteorological data, air quality data, and shipping activity data

RNN and LSTM

Emissions (PM_2.5, PM₁₀, SO₂, O₃, NO₂, CO)

Secondary data sources

Busan port (Korea)

[43]

●

Berthing

●

Hourly data of energy (electricity) prices and load demands

LSTM, NN (BP), Elman, RBF

Day-ahead prices of energy

Secondary data sources

A navigation route in Australia

[26]

●

Port performance evaluation

●

Highly correlated input variables

PCA

Air quality, rate of treatment of wastewater, standard-reaching rate of nearshore water, green coverage rate in developed areas, and expenditure on energy-saving investments per capita

Secondary data sources, survey

15 Chinese seaports

Air quality, rate of treatment of wastewater, standard-reaching rate of nearshore water, green coverage rate in developed areas, and expenditure on energy-saving investments per capita

Hierarchical clustering

The rank of ports based on environmental sustainability features