JOIV : Int. J. Inform.

Visualization, 8(1) - March 2024 1-17

INTERNATIONAL
JOURNAL ON
INFORMATICS
VISUALIZATION

INTERNATIONAL JOURNAL
ON INFORMATICS VISUALIZATION
journal homepage : www.joiv.org/index.php/joiv

Artificial Intelligence and Machine Learning for Green Shipping:

Navigating towards Sustainable Maritime Practices
Hoang Phuong Nguyen a, Cao Thao Uyen Nguyen b, Thi Men Tran b, Quoc Hai Dang c,
Nguyen Dang Khoa Pham d,*
a
Academy of Politics Region II, Ho Chi Minh City, 700000, Viet Nam
b
Faculty of International Trade, College of Foreign Economic Relation, Ho Chi Minh city, 700000, Vietnam
c
Institute of Postgraduate Education, Ho Chi Minh City University of Transport, Ho Chi Minh City, 700000, Viet Nam
d
PATET Research Group, Ho Chi Minh City University of Transport, Ho Chi Minh City, 700000, Viet Nam
Corresponding authors: *[email protected]

Abstract— This paper aims to investigate the role that artificial intelligence (AI) plays in promoting sustainability in the marine industry.
The report demonstrates the potential of AI-driven technology to improve vessel operations, decrease emissions, and promote
environmental stewardship. This potential is shown by detailed examination of existing trends, problems, and possibilities. Several vital
studies highlight the significance of policy interventions that encourage the use of artificial intelligence. These interventions include
financial incentives, legal frameworks, and programs to increase capability. Throughout this work, the importance of the role that
artificial intelligence plays in driving efficiency, safety, and sustainability is emphasized. This work also highlights the urgent need for
action to address climate change and environmental degradation in the marine sector. The marine industry can lessen its carbon
footprint, decrease pollution, and improve ecosystem health if it shifts to various alternative fuels, renewable energy sources, and
technologies powered by artificial intelligence. At the end of this work, an appeal is made to policymakers, industry stakeholders, and
technology providers, urging them to prioritize investments in artificial intelligence research and development and to create
collaboration to speed up the transition to a marine sector that is more sustainable and resilient.

Keywords— Green shipping; sustainability; renewable energy; artificial intelligence; machine learning.

Manuscript received 15 Oct. 2023; revised 11 Dec. 2023; accepted 9 Feb. 2024. Date of publication 31 Mar. 2024.
International Journal on Informatics Visualization is licensed under a Creative Commons Attribution-Share Alike 4.0 International License.

energy-efficient equipment that minimizes emissions is

I. INTRODUCTION imperative. Ports are involved with various emissions
Maritime transportation is a crucial player in global originating from many sources [14], [15]. The emissions
transportation [1]; it helps drive the growth of the global encompass marine vessels, vehicles, and cargo-handling
economy due to its ability to reduce transportation costs and systems, and they significantly impact the ecosystem [16]–
facilitate speedier intermodal operations across other modes [18]. Furthermore, it is crucial to acknowledge that activities
of transportation [2]–[5], which ports serve as crucial links for such as container drayage and interterminal transit
the transportation of products between the shoreline and the significantly contribute to the emissions generated by ports
country's interior [6], [7]. In addition, the issue of [19], [20]. Consequently, the need to research environmental
environmental sustainability has become a significant focus improvements at maritime ports, which serve as crucial
for many marine industries relating to port activities, shipping, transportation hubs, has increased. This is because these
and shipbuilding due to the problems posed by climate change upgrades enhance several environmental sustainability areas
and the increasing needs of the logistics and transportation that governments and corporations are striving to address [21],
industries [8]–[10]. The augmentation of marine [22].
transportation has partially contributed to the growth of urban The marine industry, which plays a significant part in
economies; nevertheless, it has also resulted in resource international commerce and transportation, is undergoing
depletion and environmental damage [11]–[13]. To attain significant environmental and technological transformation
sustainable expansion in urban areas and harbors, employing [23]–[25]. The market places a considerable emphasis on

1
alternative fuels, energy derived from renewable sources, and industry. This study aims to understand better the obstacles
artificial intelligence (AI), among other essential components. and possibilities associated with using these cutting-edge
By enhancing sustainability, efficiency, and environmental technologies and processes. The paper informs policymakers,
responsibility, these three pillars have the potential to provide industry stakeholders, and academics on how these insights
the maritime industry with more significant opportunities for might enhance maritime operations' sustainability, efficiency,
improvement [26], [27]. Traditional fossil fuels, such as oil and environmental stewardship. We wish to contribute to
and diesel, have been depended on by the marine industry for developing well-informed policies and initiatives that will
a very long time [28], [29]. These fuels not only contribute to propel the marine industry toward a more environmentally
the emission of greenhouse gases but also present the friendly, technologically advanced future.
possibility of oil leaks. Alternative fuels, which include
liquefied natural gas (LNG), LPG, hydrogen, and biofuels II. MATERIALS AND METHOD
[30]–[34], provide options that are both sustainable and
beneficial to the environment [35]–[38]. If the marine sector A. Literature review
were to make use of alternative fuels and renewable energy, The evolution of green marine policies throughout history
it would be possible to drastically cut greenhouse gas has been defined by a growing understanding of
emissions while simultaneously improving air quality in port environmental concerns, the establishment of more
regions and coastal zones [39], [40]. Two advantages may be sophisticated standards, and a shared determination to limit
gained by utilizing renewable energy sources such as solar the ecological effect of the maritime sector. These three
and wind power to generate electricity [41]–[46]. By factors have been integral to the growth of green marine
installing solar panels and wind turbines on ships or harbors, policies. Concerns about the environmental impact of the
it is possible to reduce the need for fossil fuels while shipping sector first surfaced in the later part of the 20th
simultaneously providing an environmentally friendly and century, when the roots of environmentally aware maritime
dependable source of energy consumption [47]–[50]. It is also legislation can be traced back to its beginnings [78]. The
possible to deploy technologies that generate renewable increase in pollution, oil spills, and rubbish dumping in seas
energy for emission-free powering systems, which is an has been a concern for environmentalists and governments
essential step toward achieving zero-emission shipping [51]– worldwide.
[53]. The MARPOL Convention is an international convention
Artificial Intelligence (AI) has the potential to created to save the maritime environment and prevent
revolutionize several aspects of many sectors, such as pollution caused by ships. The International Convention for
maritime industry, energy, transportation, manufacturing, and the Prevention of Pollution from Ships (MARPOL),
agriculture [54]–[63]. In addition, AI has been found to have implemented by the International Marine Organization (IMO)
great potential in other fields such as medicine, society, in 1973, is considered one of the first and most important
education, and economy [64]–[68]. With autonomous boats, international accords that address the issue of pollution in the
predictive maintenance, and route optimization, AI may aquatic environment. Several laws were put into place by
enhance safety, productivity, and decision-making processes MARPOL to limit the pollution produced by ships. These
[69]–[71]. AI can optimize routes, reduce fuel consumption, regulations focused on oil pollution, dangerous liquid
and improve operational efficiency by analyzing data chemicals, garbage, and air emissions [79], [80]. The
gathered from sensors, weather reports, and historical travel futuristic IMPO plan for green shipping is depicted in Fig. 1.
records. These three factors must interact for the marine On November 2, 1973, the International Maritime
industry to be viable over the long run. For instance, AI can Organization (IMO) approved the MARPOL Convention. A
help integrate renewable energy sources and alternative fuels string of tanker accidents between 1976 and 1977 led to
[72]–[74]. Power systems may be efficiently managed by it, establishing the Protocol of 1978, which directly responded to
ensuring that ships use alternative fuels when available and those disasters. Because the MARPOL Convention of 1973
switch to traditional fuels when needed [75]–[77]. This had not been implemented, the MARPOL Protocol of 1978
technique lowers emissions without sacrificing operational effectively absorbed and superseded the original Convention.
stability. Utilizing these technologies also aligns with The combined document entered into force by the law on
international efforts to reduce greenhouse gas emissions and October 2, 1983. The Convention was modified in 1997 by
address the environmental effects of the marine sector. The the approval of a Protocol, which included the establishment
International Maritime Organization (IMO) has set aggressive of a new Annex VI [21], [22]. This was one of the
goals to improve energy efficiency and reduce emissions; amendments that was made. The date when this modification
alternative fuels, AI, and renewable energy will be critical to became effective was May 19th, 2005. Throughout its
achieving these objectives. Ultimately, the maritime sector is existence, the MARPOL convention has been subjected to
pivotal in its history, and renewable energy, artificial alterations in the form of amendments that have been
intelligence, and alternative fuels might significantly implemented. The Convention includes provisions intended
influence its future. By adding these elements, the sector may to prevent and decrease pollution produced by ships. These
move closer to a more ecologically responsible and laws include accidental contamination and pollution that
sustainable future, reducing its ecological footprint while result from ships' everyday activities. Currently, there are six
increasing its efficiency and competitiveness in a constantly technical Annexes inside it. Particular Areas, which are
changing global market. This research will thoroughly usually included in the majority of Annexes, are responsible
investigate and explain how alternative fuels, renewable for enforcing tight rules on operational discharges [78], [82],
energy, and artificial intelligence will disrupt the marine [83].

2
 Fig. 1 The IMO plan for green shipping [81]

Throughout its existence, the MARPOL Convention has of greenhouse gases that are caused by global shipping.
been subjected to many updates, some of which have been Compared to the levels recorded in 2008, the level of
particularly noteworthy, such as Annex VI, which addresses greenhouse gas emissions from the transportation sector is
the problem of air pollution created by ships. The expected to fall by a minimum of fifty percent by the year
implementation of Annex VI in 1997, followed by its 2050. The plan aims to lower the amount of carbon emissions
strengthening in 2005, resulted in the establishment of produced for each unit of transportation labor. National and
stringent rules for the emission of sulfur oxides (SOx) and regional bodies have each developed their own ecologically
nitrogen oxides (NOx) from commercial vessels. According friendly marine regulations in addition to the worldwide
to these laws, using fuels with low levels of sulfur and standards that have been established [87], [88]. Many of these
installing catalytic converters are encouraged. Annex VI of programs include incentives for using alternative fuels,
the MARPOL Convention was the source of inspiration for renewable energy, and technology that are efficient in energy
the concept of Emission Control Areas, sometimes known as consumption. To ensure compliance with environmental rules,
ECAs. More rigorous limits on ship emissions have been deploying and adopting innovative technology, such as
implemented in some places, including the Baltic and North engines that are less harmful to the environment, ballast water
Seas. Some laws encourage the deployment of emission- management systems, and enhanced waste treatment facilities,
reduction technologies or the use of cleaner fuels. A have been essential. The historical history of green maritime
worldwide sulfur limitation was adopted in 2020 as part of legislation illustrates the continual development of the marine
MARPOL Annex VI. This cap is designed to limit the amount industry. These regulations are evidence of the industry's
of sulfur in marine fuels to 50%. The use of low-sulfur fuels commitment to minimizing its environmental impact. As a
or the deployment of exhaust gas cleaning equipment, also result of the persistent efforts to implement sustainable
referred to as scrubbers, was encouraged due to this event, practices, the incorporation of alternative fuels, renewable
which constituted a significant milestone in reducing sulfur energy, and artificial intelligence, as well as the enforcement
dioxide emissions [84]–[86]. of harsher standards, the marine industry has become more
The Initial Greenhouse Gas (GHG) Strategy was environmentally sensitive and efficient [89]–[91]. The policy
implemented by the International Maritime Organization framework for green shipping and its effects is depicted in Fig.
(IMO) in 2018. The goal of this plan is to reduce the emissions 2.

Fig. 2 Policy framework for green shipping and its effects [92]

The International Maritime Organization (IMO), a in creating and enforcing regulations that apply to the marine
specialized entity of the United Nations, plays a crucial role sector globally. Many rules apply to environmentally friendly

3
practices in the maritime industry that are enacted by the service [95]–[97]. Ship operators can maximize their boats'
International Maritime Organization (IMO): performance using artificial intelligence (AI) technologies to
Annex VI of the MARPOL Convention limits the emission meet these objectives.
of air pollutants from boats. These pollutants include sulfur A framework known as the Carbon Intensity Indicator (CII)
oxides (SOx), nitrogen oxides (NOx), and greenhouse gases was put into place by the IMO as a component of their Initial
(GHGs) [93], [94]. It is possible to include artificial GHG Strategy. This project aims to create a meter that can
intelligence in vessel operations to enhance the effectiveness measure energy efficiency. Calculating and monitoring the
of monitoring and controlling compliance with these CII of a vessel and identifying possible areas for improvement
standards. The goal of the Energy Efficiency Existing Ship may be accomplished with artificial intelligence [86]. The
Index (EEXI), which is scheduled to be introduced in 2023, is different technologies that can be used to improve the green
to improve the energy efficiency of ships that are currently in maritime sector are depicted in Fig. 3.

Fig. 3 Various technologies for improvement of the green maritime sector [98]

Fig. 4 The geographic distribution of mean green marine transport efficiency per nation [92]

The European Union (EU) has been at the forefront of monitor and regulate emissions is becoming increasingly vital
implementing ecologically friendly maritime rules to cut [99], [100].
emissions and advance the cause of sustainability. Among the A significant number of countries have enacted their very
topics under the purview of EU regulation are those own laws and regulations concerning emissions, pollution,
concerning energy efficiency, carbon dioxide emissions, and and environmentally responsible shipping operations. The use
sulfur emissions [88]. Technologies based on artificial of artificial intelligence is highly advantageous since it
intelligence (AI) have the potential to assist in achieving and enables warships to satisfy a wide variety of needs that are
exceeding these stringent criteria. The Baltic and North Seas constantly evolving. The IMO has recognized the need to
are two examples of areas recognized as Emission Control address cybersecurity concerns in this age of digitalization.
Areas (ECAs), designated with more rigorous emission One of the most critical aspects of the digitization of marine
regulations. Because ships operating in these locations are operations is using artificial intelligence (AI). It aims to
expected to comply with higher limitations on emissions, the strengthen cybersecurity procedures and protect boats from
utilization of technology driven by artificial intelligence to potential cyber hazards. Due to the growing popularity of

4
autonomous shipping, efforts are being made to build [101], [102]. The geographic distribution of mean green
worldwide standards and laws that will ensure the safe and marine transport efficiency per nation is depicted in Fig. 4. A
environmentally responsible use of AI-driven autonomous flow chart of the ML application process is described in Fig.
boats. These standards and rules are being developed in 5. At the same time, Table 1 is a summary list of recent studies
response to the growing demand for autonomous shipping using AI & ML for green shipping.

Fig. 5 Flow chart of ML application process [103]

TABLE I
RECENT STUDIES USING AI & ML FOR GREEN SHIPPING

Objectives ML / AI used Main outcomes Source

Model forecasting of sulfur Multiple linear regression (MLR), Automated The AutoML was superior in data prediction for SO2 [17]
dioxide emission from shipping machine learning regressor (AutoML), Gradient emission.
descent, and Artificial neural network
Prediction model for onboard Extreme learning machines (ELM), kernel ELM could provide at least 5.6% more precise [104]
photovoltaic system density estimation (KDE), and K-mean predictions
clustering
Machine learning-based Principal component regression, Partial least The ML techniques have demonstrated effective [105]
prediction of ship’s performance square regression, probabilistic ANN, and ANN performance in simulating ship hydrodynamic
conditions, with the probabilistic ANN model being the
most effective.
Prediction optimization of trim ANN, Decision tree (DT Random Forest (RF), The prediction precision of RF was the best, with [106]
in case of container ship and K-nearest neighbor 88.89% accuracy.
ML with physics interpretation eXtreme Gradient Boosting (XGBoost) and ML may improve speed prediction by 30% when [107]
for ship speed prediction Physics-informed neural networks (PINNs) enough data is available for modeling.
Model-prediction of Multiple machine learning techniques A maneuvering emission impact forecasting model for [108]
Manoeuvring emissions pilots is constructed with a 73% consistency.
ML-based prediction of ship’s Gaussian process regression (GPR) The proposed ML-based methodology could estimate [109]
trajectory the probabilistic pattern of grounding risk and ship
dynamics.

5
 Objectives ML / AI used Main outcomes Source
Model prediction of ship RF, bagging, and boosting Tree-based modeling offers significant advantages since [9]
performance with tree-based it delivers excellent accuracy and simplicity of
methods comprehension.
Hybrid propeller-engine Multiple ML approaches The groundwork for sophisticated data analytics will be [110]
diagram for green shipping used to determine optimal vessel navigation.

AI integration within the framework of these policies and

Fuel cost =
regulations involves the utilization of AI for:
 Strategies for monitoring and reducing emissions. ∑ ∆ + 1− + (1)
 Implementing predictive maintenance to optimize
vessel performance and minimize operational Eq. (2) represents the cost that was incurred as a result of
interruptions. the carbon tax measures:
 Optimization of vessel routes to achieve fuel efficiency
Cost of carbon =
and minimize emissions. ! " Δ + − $ . &' (
 Improved safety is achieved by utilizing artificial (2)
intelligence-powered autonomous navigation systems.
 The integration of AI in the maritime sector addresses The amount of time spent sailing determines a portion of
environmental concerns and facilitates compliance with the cost. To begin, ships are rented and given a fee
regulations. Furthermore, it offers potential for proportional to their use time. In the second place, the
enhancing the industry's sustainability, efficiency, and navigation of vessels requires both labor and equipment as
safety. essential components. The costs consist of the salary of the
 AI's capacity to foster innovation and promote staff members who are on board, the expenses that are
environmental responsibility in maritime operations is essential for living, the fees for maintaining the equipment,
increasingly acknowledged by policymakers and the depreciation charge, and the insurance payments. It is
stakeholders. calculated with Eq. (3):
B. Method Cost time = ) ∑ Δ (3)
1) Analytical description Cost of incentive for speed reduction of the ship needs that
Prescribing the main terms and definitions employed the shipping company is entitled to monetary compensation
during the study is essential. The “Lane” here denotes the by the VSRIP if the vessel deviates from the designated speed
entire route of a ship transporting cargo via various ports. The limit within VSRZs in the vicinity of the departure port A and
term “Route” represents the distance between two fixed ports. the arrival port B. By diminishing the overall expenses, this
Each route has several alternatives, such as curved or straight. component of the cost acquires a negative value. That is
The term “Stretch” means that a route can be broken into articulated as:
many lengths based on the region established by the policy. Cost owing to VSRIP =
+,-.
− ∑+,-. 3,5 *+,-. /0 < 0 2 . 6+,-. .
For example, ECA's policy divides the sea surface into two (4)
areas: within the emission control region and outside the 7

control area. There is a path both inside and outside of the While a ship is cruising, fuel oil produces SOx and CO2.
emission control area. The route can then be broken into many Here's why SOx and CO2 are used to investigate emission
sections along the boundary of the emission control region. A reductions. CO2 emissions contribute to global warming,
route can be split into many segments of a defined length. In while pollutant SOx affect the environment. SOx and CO2
other words, every section is the same length. To be sure, emissions are the primary targets of emission reduction
repeatedly changing speed or heading harms the ship's measures. In this research, ECAs aim to decrease SOx
operation and may increase fuel consumption and emissions. emissions, while carbon pricing schemes aim to reduce CO2
Thus, ships move in a straight line with a constant velocity emissions. Emission levels are determined by the ship's power
throughout each segment [111]. In the present study, the ship's and gas emission characteristics. This is mentioned in [113]:
sailing speed is initially optimized through the joint
application of virtual arrival methods with a carbon tax, 89: = " Δ < = >: + 1 − >: ?
emission control area (ECA), and Vessel Speed Reduction
Incentive Program (VSRIP). This technique presupposes that + >: @ (5)
the departure and arrival ports are known and that the route
In this case, j denotes the type of gas, j = SOx or CO2. Eq.
has been planned. The speed model's optimization aim is to
(5) is used to compute gas emissions once the sailing speed
minimize overall cost. The primary purpose of this section is
has been calculated to reduce the overall cost under various
to determine the objective function.
policies and techniques.
Fuel expenditures are incurred on all ship journeys. Unit
In the above discussion, ‘i' denotes the segmental index, ‘n’
costs vary depending on the kind of fuel. Ships utilize marine
represents the total number of segments in a single route, va is
gas oil (MGO) throughout ECAs and low sulfur fuel oil the threshold of sailing speed, vi denotes the segmental speed
(LSFO) outside ECAs for their principal engine fuel. The of a ship, vmax is the peak speed of the boat, Δti is segmental
sulfur content of gasoline used by auxiliary engines is often sailing time, T denotes ship’s total sailing time on the whole
relatively low, and its unit price is comparable to MGO.c In
route, ΔT represents the time delay in implementing the
summary, fuel costs are stated as [111], [112]:

6
Virtual Arrival strategy. αi is either 1 or 0 for emission control successively corrects mistakes caused by earlier models, and
areas and outside emission control areas, respectively. τm approximation tree learning, which accelerates training by
denotes fuel consumption for the main engine's fuel creating trees level-wise, are two of its main properties. Both
consumption in ton/kW·h while τa denotes fuel consumption of these elements are among its most essential characteristics.
in the case of the auxiliary engine's fuel consumption in The success of XGBoost may be attributed to its capacity to
ton/kW·h. WL denotes the cost of marine fuel in dollars per generate very accurate predictions and its adaptability in
ton, while WH represents the cost of low-sulfur fuel oil in various areas [116], [117]. As a result, it is quickly becoming
dollars per ton. the tool preferred by data scientists and machine learning
practitioners. The following is the procedure followed:
2) Data analysis
If the training data set is defined as:
A= B ,C
The data collected from the ship’s logbook through
several trips was used as a case study. The data was cleaned (6)

In this case, the is feature vector while C denotes the

and arranged, and descriptive statistical analysis was
conducted [114], [115]. The descriptive statistics of the data
corresponding label for ith sample.
are listed in Table 2. The descriptive statistical analysis shows
The objective of XGBoost is to learn the additive ensemble
that the mean value for Distance is 824.783 nm, the standard
of weak learners
D B = ∑G EF B ; herein, each denotes a
deviation is 836.41 nm, and the range of values for Distance
is from 0 nm to 6870.7 nm. The average number of hours daily F
is 78.828 hours, and the standard deviation is 96.61 hours. regression tree.
The objective function of XGBoost is given as:
TABLE II
DESCRIPTIVE STATISTICAL ANALYSIS G
HI = " J C ,⏞
C + " Ω EF (7)
Distance, Time, LFO, DO, F
nm Hr Gallons Gallons
count 100 100 100 100 Herein, J C , C⏞ the loss function is the actual value
mean 824.783 78.828 66.6992 1.48816 while denoting the predicted value. The model is trained
std 836.41 96.61 102.84 4.96 repeatedly by introducing weak learners into the ensemble. In
min 0 0 12.955 0.029 each cycle, a new regression tree is trained with the aim of
25% 461.125 44.9 32.74375 0.09425 minimizing the following objective function:
50% 653.6 64.05 42.639 0.197
.
.
75% 864.925 80.875 59.99075 2.0265 HIF = " J C ,⏞
C + " Ω E + $. &
max 6870.7 717 839.884 48.551
1 ) (8)
Kurtosis 29.22 32.15 34.97 84.23
+ N " O:!
Skewnes
4.67 5.343 5.45 8.84 2 :
s
In this expression, T denotes the number of leaves and N
The range of time is from 0 hours to 717 hours. LFO has a represents the regularization term defining the complexity of
mean of 66.6992 gallons, a standard deviation of 102.84 the tree. Following training, the final prediction for a fresh
gallons, and a range from 12.955 gallons to 839.884 gallons. sample x is derived by adding the forecasts of all regression
The DO has a mean value of 1.48816 gallons, with a standard trees in the ensemble: herein, the forecast of the kth tree.
deviation of 4.96. The DO may range anywhere from 0.029
gallons to 48.551 gallons. Given that the kurtosis values for III. RESULTS AND DISCUSSION
Distance, Time, LFO, and DO are 29.22, 32.15, 34.97, and
84.23, respectively, these variables' probability distributions A. Results
are tail-like. The skewness values for Distance, Time, LFO, 1) Preprocessing
and DO are 4.67, 5.343, 5.45, and 8.84, respectively,
Data preparation, which includes developing correlation
illustrating the level of asymmetry around their respective
heatmaps, ensures analytical integrity, trustworthiness, and
means. The presence of a positive skew indicates a longer
efficacy. These processes are critical to research and analysis
right tail, while the presence of a negative skew indicates a
longer left tail. Higher absolute values indicate a higher for the following reasons. Correlation heatmaps demonstrate
deviation from the mean. the correlations between dataset variables. They let
researchers determine how variables interact, whether they
3) Machine learning for model-prediction: are positively or negatively related, and how much. This
Extreme Gradient Boosting, or XGBoost, is a robust and information is essential for choosing model predictors and
frequently used machine learning method well-known for its identifying multicollinearity concerns. Identifying strong
effectiveness and simplicity in managing structured and connections might help you choose features by reducing
tabular data. It is a member of the family of methods known duplicate data points irrelevant to the investigation. Fig. 6
as ensemble learning, and it works by iteratively constructing displays the correlation heatmap for the research data. The
an ensemble of weak learners, often decision trees, to generate data is about green shipping using a mix of low-sulfur oil
a robust predictive model. Using regularization to minimize (LSO) and diesel oil (DO).
overfitting and the minimization of a particular loss function,
XGBoost can perform very well in regression and
classification exercises. Gradient boosting, which

7
 The results of the experiments that were carried out to
demonstrate the prediction capabilities of the machine
learning model XGBoost on ship fuel consumption data are
shown in this part of the findings. In this section, the model
development process is broken down in great depth, and a
comprehensive analysis of the advantages and disadvantages
of each model is presented. The prediction model was
developed using a systematic process comprising data
preparation, feature engineering, model selection,
hyperparameter tuning, and rigorous evaluation. This method
was followed for every one of the models. Ship-related data,
such as the distance traveled by the vessel, the time it took,
and the amount of fuel used in the past, was gathered and
analyzed for training purposes. To consider the time
dependency, the feature engineering process included the
Fig. 6 Correlational heatmap
development of lag features and the data standardization to
2) Model-prediction achieve consistent scaling, as shown in Fig. 7.

(a) (b)
Fig. 7 XGBoost-based model’s actual vs predicted (a) LFO data (b) DO data

R-squared, Mean Squared Error (MSE), and Mean Moving on to the DO model, their R-squared values were
Absolute Error (MAE) are three metrics that may describe the even higher than those of LSO. This is especially true in the
LSO and DO models based on XGBoost. These metrics can training dataset, where it reaches 0.9998, which indicates that
be utilized for both the training and test datasets. Beginning there is a robust correlation between the values that were
with the LSO model, it was observed that the R-squared predicted and the values that were observed. The R-squared
values are high in both the training dataset and the test dataset. value, on the other hand, lowers to 0.8006 in the test dataset,
This indicates a significant connection between the predicted which indicates that the model's performance on data that has
values of the fuel-related variable and the actual values of the not yet been seen has somewhat decreased but has still
variable. The value of R squared in the training dataset is 0.96, retained a pretty high degree of correlation. Within the
which indicates that the model can explain about 96% of the training dataset, the MSE values for DO are very near zero,
variation in the variable that is associated with fuel. Similarly, indicating very few mistakes between the predicted and actual
the R-squared score in the test dataset is slightly lower at 0.92, values. Furthermore, the MSE value in the test dataset is
which indicates that the model continues to perform well on 0.0045, which is also considered extremely low. While the
data it has not seen before. There are quite a few mistakes MAE values in the training dataset are near zero (0.0016),
between the predicted values and those observed, as shown by they are reasonably tiny in the test dataset (0.0536),
the MSE values for LSO, which are 498.55 for the training demonstrating that the model accurately predicts fuel-related
dataset and 396.26 for the test dataset. The MAE values for variables when diesel oil or distillate fuel oil is used. When
the training dataset are 6.59, whereas the MAE values for the employed as predictors for the fuel-related variable, both
test dataset are 9.68. This further substantiates the model's kinds of ship fuels, LSO and DO, have solid correlations and
accuracy in predicting fuel-related variables when using low- low prediction errors. The high R-squared values and the low
sulfur fuel oil.

8
MSE and MAE values across the training and test datasets extraction, production, and disposal of resources and waste
demonstrate this. [87], [91].
Long-term sustainability: In the framework of long-term
B. Policy suggestions sustainability, it is necessary to take into consideration the
1) Environmental implications environmental advantages that would result from the marine
industry's use of alternative fuels, various forms of renewable
The marine sector may see significant environmental
energy, and artificial intelligence. The widespread adoption
implications, good and bad, due to adopting technology such
of these technologies should align with broader sustainability
as artificial intelligence (AI), renewable energy sources, and
alternative fuels. An analysis of these effects is presented as goals, such as preserving biodiversity, protecting ecosystems,
follows: and equitable distribution of environmental benefits and
burdens. Although these technologies present opportunities to
GHG reduction: One of the most important advantages of
reduce environmental harm in the short term, their widespread
implementing alternative fuels and renewable energy sources
adoption should align with these broader sustainability goals
in the marine sector is the massive decrease in greenhouse gas
[128], [129].
emissions. This is one of the critical benefits of this adoption.
Using alternative fuels, renewable energy sources, and
Traditional marine fuels, such as heavy fuel oil, are
artificial intelligence technologies in the maritime sector can
responsible for the emission of significant quantities of certain
drastically cut greenhouse gas emissions, lessen marine
pollutants, including carbon dioxide (CO2), sulfur oxides
pollution, and increase energy efficiency. However, to
(SOx), and nitrogen oxides (NOx) [118], [119]. The adoption
guarantee that these technologies effectively contribute to
of more environmentally friendly alternatives, such as
liquefied natural gas (LNG), hydrogen, or biofuels, has the achieving long-term sustainability objectives, seriously
potential to reduce these emissions significantly, therefore considering the potential environmental repercussions is
necessary.
contributing to the mitigation of climate change and the
improvement of air quality [120], [121]. 2) Economic implications
Marine pollution mitigation: Traditional marine fuels Cost Savings: Artificial intelligence technologies can
contribute to air pollution and carry concerns of oil spills and potentially optimize various nautical activities, resulting in
water contamination. Air pollution is a significant contributor business cost savings. As an illustration, predictive
to marine pollution. Alternative fuels, such as LNG, offer a maintenance systems powered by artificial intelligence can
decreased risk of leaks and can contribute to reducing assist in predicting the breakdown of equipment, enabling
pollution in maritime environments. A further reduction in prompt repairs and reducing downtime. Similarly, AI-driven
dependency on fossil fuels and the danger of oil spills can be route optimization algorithms can minimize the amount of
achieved by using renewable energy sources such as wind and fuel consumed and the duration of voyages, resulting in
solar power to provide auxiliary power [122], [123]. cheaper operational costs for shipping corporations [82],
Energy efficiency enhancement: Artificial intelligence [130].
technology may be utilized to optimize vessel operations, Efficiency enhancements: Artificial intelligence can
resulting in enhanced energy efficiency and decreased fuel improve operational efficiency throughout the marine supply
usage and consumption. Intelligent systems powered by chain. AI technologies can potentially enhance overall
artificial intelligence can optimize ship routes, modify speeds efficiency and resource utilization, which may result in cost
based on weather conditions, and operate onboard systems savings for enterprises. This can be accomplished by
more effectively, ultimately resulting in substantial fuel automating repetitive jobs, optimizing logistics and inventory
savings and emissions reductions [124], [125]. management, and expediting cargo handling procedures [131],
Impact on marine ecosystem: Despite the potential of [132].
alternative fuels and renewable energy sources to lessen the Market opportunities: The implementation of artificial
amount of pollution caused by maritime activities, intelligence in the marine sector has the potential to provide
implementing these technologies may have an indirect effect new income streams and market opportunities. There is a
on marine ecosystems. The manufacturing and distributing possibility that businesses that utilize artificial intelligence for
biofuels and installing offshore wind farms are two examples predictive analytics, autonomous navigation, or intelligent
of activities that can potentially damage aquatic ecosystems port management may gain a competitive advantage and
and habitats. Careful planning and extensive environmental attract new consumers who are looking for creative and
impact studies are required to minimize these consequences efficient marine solutions [133], [134].
and ensure the deployment of these technologies in a Reduction of risk: Risk management systems driven by
sustainable manner [126], [127]. artificial intelligence can analyze vast volumes of data to
Consumption of resources: The creation and deployment of detect possible safety hazards, security threats, and
artificial intelligence technology, alternative fuels, and operational risks. Marine firms can reduce disruptions, avoid
infrastructure for renewable energy all demand significant costly accidents, and preserve regulatory compliance by
resources, such as land, water, and materials. Depending on taking proactive measures to handle these risks. This will
how these resources are obtained and maintained, reduce financial obligations and insurance premiums [135],
environmental implications may be connected with them. [136].
These impacts may include the loss of habitats, the use of
water, and increased carbon emissions related to 3) Society impacts
manufacturing operations. To lessen the severity of these Job creation: Artificial intelligence has the potential to
effects, it is necessary to encourage sustainable methods in the generate new employment possibilities, even though there are

9
worries over the displacement of people in the marine sector possible good influence on society, addressing potential
by automation. There will be an increase in the need for issues such as the displacement of workforces and ethical
skilled individuals in industries such as data science, software considerations is vital.
development, robotics, and AI system maintenance due to the
4) Challenges and barriers
increasing number of organizations investing in artificial
intelligence technology. Additionally, breakthroughs that are Putting into practice environmentally friendly policies and
driven by artificial intelligence have the potential to artificial intelligence solutions in the marine sector is met with
encourage employment development in similar areas, such as several hurdles and impediments, which may be broken down
engineering, technology, and logistics [137], [138]. into four categories: technological, regulatory, economic, and
Better health results: Artificial intelligence applications in cultural. A summary in this regard is depicted in Fig. 8.
the marine industry have the potential to improve crew safety Technological obstacles:
and well-being, which in turn can contribute to better health  The process of retrofitting older vessels with
results. For instance, through predictive analytics driven by environmentally friendly technologies or adopting
artificial intelligence, it is possible to identify potential health artificial intelligence solutions can be technically
problems among crew members, enabling preemptive demanding and expensive simultaneously. Because of
interventions and preventative actions. Additionally, this, it could be necessary to make considerable
autonomous and remotely operated vessels equipped with adjustments to the infrastructure and systems on board.
collision avoidance systems based on artificial intelligence  The importance of ensuring the dependability and
can potentially lower the likelihood of accidents occurring at safety of artificial intelligence (AI) systems and
sea, therefore protecting the lives and health of maritime environmentally friendly technology in marine
personnel [137], [139]. operations cannot be overstated. Accidents or
Environmental benefits: When it comes to marine disruptions in operations might be caused by technical
operations, using artificial intelligence may result in faults or malfunctions, which would pose a threat to the
significant environmental benefits that, in turn, improve safety of the crew as well as the environmental
social well-being. Although these benefits are not directly tied sustainability of the environment.
to economic gains, they are still important. Innovations in the  Artificial intelligence systems depend on having access
maritime industry driven by artificial intelligence assist in to vast amounts of high-quality data to train and make
preserving ecosystems, maintaining natural resources, and decisions. Interoperability problems and data silos
protecting the health of populations that rely on marine within the marine sector, on the other hand, might be
ecosystems for their sustenance and happiness. These obstacles to the successful adoption of AI-driven
innovations help reduce greenhouse gas emissions, mitigate solutions
marine pollution, and promote sustainable practices [78], Regulatory obstacles:
[140].  Maritime stakeholders must comply with green policies
In conclusion, AI provides various economic benefits, such and regulations to reduce emissions and promote
as cost reductions, increases in efficiency, and possibilities in sustainable practices. These policies and laws place
the market. Additionally, it delivers social benefits, such as compliance duties on maritime stakeholders.
creating jobs and improved health outcomes for marine Implementing these requirements may require
workers and communities. To ensure that the implementation substantial financial investments in technological
of artificial intelligence in the marine sector has the greatest improvements and operational modifications.

Fig. 8 Summary of challenges and opportunities [98]

10
 Regulation uncertainty: The rapid growth of artificial In conclusion, to overcome the issues and hurdles that the
intelligence and environmentally friendly technology maritime sector is encountering when attempting to
presents regulatory organizations with difficulty in implement green policies and AI solutions, it is necessary to
defining clear and consistent standards. Regulatory take a multidimensional strategy that includes technology
ambiguity can create obstacles that prevent innovation innovation, legislative reform, economic incentives, and
and investment in environmentally friendly solutions. cultural transformation. It is vital to collaborate among
 International Coordination Due to the maritime stakeholders and take proactive actions to overcome technical,
industry's global nature, multiple stakeholders, such as regulatory, financial, and organizational constraints to
governments, international organizations, and industry achieve the full potential of technology-driven solutions that
associations, must coordinate efforts to harmonize are sustainable and environmentally friendly in the marine
regulatory frameworks and guarantee that industry.
environmental and safety standards are consistently
5) Policy recommendations for promoting green
enforced.
maritime practices and integrating AI solutions
Economic challenges:
 The implementation of green policies and artificial Establish clear regulatory frameworks:
intelligence solutions sometimes necessitates  To encourage the marine sector to use environmentally
considerable initial investments. This may discourage friendly technology and artificial intelligence solutions,
specific marine organizations, tiny and medium-sized regulatory frameworks that are both comprehensive
enterprises (SMEs), from embracing these technologies. and uniform should be developed.
 Evaluating the long-term economic advantages of  To drive decarbonization activities effectively, it is
green technologies and artificial intelligence solutions necessary to establish emission reduction objectives
can be difficult owing to the uncertainties surrounding and put environmental solid regulations in place, such
fuel pricing, improvements in regulatory policies, and as imposing limits on sulfur and nitrogen oxide
technological breakthroughs. Without clear proof of emissions.
return on investment, businesses could hesitate to  Legislative incentives should be provided to stimulate
invest. investment in environmentally friendly technology and
 It may be difficult to obtain finance for artificial AI-driven efficiency gains. These incentives might be
intelligence and green projects, particularly for small tax breaks, subsidies, or carbon trading programs.
and medium-sized enterprises (SMEs) and businesses Investment in research and development:
that operate in emerging countries. To encourage the  Funding should be allocated for research and
widespread adoption of environmentally friendly and development projects that aim to enhance artificial
technology-driven solutions in the marine industry, it is intelligence solutions and environmentally friendly
vital to address the limits that may be imposed by technology for the marine industry.
funding.  To speed up innovation in alternative fuels,
Organizational and cultural structure: autonomous shipping, and predictive maintenance, it is
 The adoption of green policies and artificial essential to encourage collaboration between industry
intelligence solutions can be hampered by resistance to players, academic institutions, and research groups.
change, which can be considered cultural opposition Collaborative learning and development of capabilities:
inside marine companies and conventional views  to enhance the skills and capacities of marine
toward innovation and sustainability. To successfully professionals in artificial intelligence, data analytics,
overcome opposition to change, leadership must and sustainability, it is necessary to establish training
demonstrate commitment, employees must be trained, programs and platforms for knowledge sharing.
and stakeholders must be engaged.  Make sure that there is a steady supply of qualified
 Data analytics, machine learning, and cybersecurity are individuals for the marine sector by providing financial
all areas in which specialist expertise is required to assistance to workforce development projects and
implement artificial intelligence solutions in the marine educational institutions that offer courses in disciplines
business successfully. To guarantee the effective relevant to the industry.
deployment and exploitation of artificial intelligence Encourage public-private partnerships:
technologies, it is vital to address the skills gap by  The development and implementation of
implementing education and training programs. environmentally friendly marine solutions should be
 Collaboration across stakeholders: Effective facilitated by public-private partnerships (PPPs), which
collaboration across a wide range of stakeholders is should encourage collaboration between governments,
required to successfully implement green policies and industry participants, and technology suppliers.
artificial intelligence solutions. These stakeholders  To encourage the adoption of AI-driven technologies
include shipowners, operators, regulators, technology and best practices, it is necessary to establish platforms
suppliers, and research institutes. Establishing trust and for sharing information and cooperation. Some
cultivating relationships among various stakeholders is examples of such platforms should include industry
paramount to overcome obstacles and propel collective consortia, innovation centers, and demonstration
action toward sustainability and innovation. projects.

11
 Facilitate the exchange of Information and interoperability: facilitate the incorporation of AI, these steps can be adapted
 Establishing standards and procedures for collecting, in the following manner:
storing, and exchanging data is a great way to Financial incentives for the implementation of AI:
encourage data sharing and interoperability across  Provide marine enterprises investing in artificial
several stakeholders in the marine industry. intelligence technology with financial incentives, such
 Create frameworks that will allow for the exchange of as grants, subsidies, or tax credits, that are expressly
anonymized data for the sake of research while at the targeted toward them.
same time preserving sensitive business information  It would be beneficial to provide money for pilot
and maintaining the privacy and security of data. projects and demonstrations demonstrating artificial
Financial incentives and support mechanisms: intelligence's advantages in boosting efficiency, safety,
 The use of environmentally friendly technology and and environmental performance within the marine
artificial intelligence solutions by marine firms should industry.
be supported by providing financial incentives such as  Establish incentives that are based on performance and
grants, loans with low interest rates, and investment tax are related to outcomes that are driven by artificial
credits. intelligence, such as increases in fuel economy,
 To finance sustainability initiatives and technological reductions in emissions, or savings in operating costs.
improvements in the marine industry, green financing Support for research and development via subsidies:
methods should be established. Some examples of such  Allocate subsidies and research grants to assist research
mechanisms are green bonds and revolving funds. and development efforts focused on creating artificial
Promote international collaboration and knowledge intelligence solutions suited to the marine sector's
exchange: specific difficulties and requirements.
 Multilateral forums, such as the International Maritime  Encourage collaboration between public research
Organization (IMO) and regional maritime organizations, universities, and entities from the
organizations, should be utilized to facilitate business sector through research partnerships that are
international collaboration and the exchange of financially supported to enhance artificial intelligence
information. applications in marine operations.
 To promote the worldwide shift towards Frameworks for the regulation of Artificial Intelligence:
environmentally responsible and technology-driven  To address the ethical, safety, and liability problems
marine activities, sharing best practices, lessons learned, that are linked with the use of artificial intelligence in
and newly developed technical advances is essential. In marine operations, regulatory frameworks should be
addition, it should enhance students' and learners' developed.
awareness of the importance of protecting the  To guarantee that artificial intelligence technologies are
environment [141]. used responsibly within the maritime sector, it is
Monitoring and evaluating progress: necessary to establish principles for AI governance.
 Establish systems for monitoring and assessing the These guidelines should include algorithm
application of green marine policies and artificial transparency, data privacy, cybersecurity, and
intelligence solutions. These mechanisms should accountability criteria.
include the tracking of emission reductions,  To facilitate the worldwide adoption of AI-driven
improvements in energy efficiency, and the deployment solutions in marine shipping, it is essential to
of technology [142]. collaborate with international organizations and
 To ensure that regulatory frameworks and policy industry players to unify artificial intelligence rules
measures are successful, identifying areas in which across different jurisdictions.
they may be improved, and adapting to changing Developing capabilities and providing training programs:
technical and market trends, it is essential to conduct  Training programs and capacity-building projects
periodic reviews. centered on developing artificial intelligence
Implementing these policy proposals will enable capabilities should be made available to marine
policymakers to create an atmosphere favorable to including professionals such as ship operators, engineers, and
artificial intelligence solutions and promoting data analysts [143], [144].
environmentally friendly maritime practices. This  For marine enterprises engaging in staff training and
environment will be conducive to the development of upskilling programs connected to artificial intelligence
ecologically friendly maritime practices. As a consequence of technology and data analytics, provide financial help
this, the marine sector will be able to experience sustainable through subsidies or tuition aid.
expansion and innovation while also addressing concerns  Work with educational institutions and training
regarding the environment. providers to produce customized training programs and
certification courses specifically designed to meet the
6) Need for incentives, subsidies, and regulatory
artificial intelligence requirements of the marine sector
frameworks that consider AI
[145], [146]. In addition,
To speed the adoption of technology-driven solutions and Innovation in Artificial Intelligence through public-private
environmentally responsible practices, it is vital to partnerships:
incorporate regulatory frameworks, incentives, and subsidies  To co-fund artificial intelligence (AI) innovation
that consider artificial intelligence in the marine industry. To initiatives and technology pilots that solve critical

12
 difficulties in marine transportation, such as engine. Part 1: Modeling aspect,” Appl. Therm. Eng., vol. 29, no. 11–
12, pp. 2153–2159, Aug. 2009,
autonomous navigation, predictive maintenance, and doi:10.1016/j.applthermaleng.2008.06.032.
emissions monitoring, it is highly recommended that [5] A. Podsiadlo and W. Tarelko, “Modelling and developing a decision-
public-private partnerships (PPPs) be facilitated. making process of hazard zone identification in ship power plants,” Int.
 To make the most of the aggregate experience and J. Press. Vessel. Pip., vol. 83, no. 4, pp. 287–298, Apr. 2006,
doi:10.1016/j.ijpvp.2006.02.017.
resources available, it is essential to encourage [6] N. D. K. Pham, G. H. Dinh, H. T. Pham, J. Kozak, and H. P. Nguyen,
collaboration between government agencies, industry “Role of Green Logistics in the Construction of Sustainable Supply
groups, technology suppliers, and academic institutions Chains,” Polish Marit. Res., vol. 30, no. 3, pp. 191–211, Sep. 2023,
to create artificial intelligence solutions for the marine doi: 10.2478/pomr-2023-0052.
[7] M. D. Nguyen, K. T. Yeon, K. Rudzki, H. P. Nguyen, and N. D. K.
industry toward the goals of era 4.0 and 5.0 for Pham, “Strategies for Developing Logistics Centres: Technological
transportation sector [147], [148]. Trends and Policy Implications,” Polish Marit. Res., vol. 30, no. 4, pp.
To prototype and verify novel artificial intelligence 129–147, 2023, doi: 10.2478/pomr-2023-0066.
applications in real-world marine contexts, it is essential to [8] A. Al-Enazi, E. C. Okonkwo, Y. Bicer, and T. Al-Ansari, “A review
of cleaner alternative fuels for maritime transportation,” Energy
establish innovation clusters or testbeds where maritime Reports, vol. 7, pp. 1962–1985, Nov. 2021,
enterprises may work with artificial intelligence startups and doi:10.1016/j.egyr.2021.03.036.
researchers. To encourage investment, innovation, and [9] O. Soner, E. Akyuz, and M. Celik, “Use of tree-based methods in ship
collaboration within the maritime industry, policymakers can performance monitoring under operating conditions,” Ocean Eng., vol.
166, pp. 302–310, Oct. 2018, doi:10.1016/j.oceaneng.2018.07.061.
incorporate incentives, subsidies, and regulatory frameworks [10] S. Vakili, A. I. Ölçer, A. Schönborn, F. Ballini, and A. T. Hoang,
specifically designed to consider artificial intelligence [149], “Energy‐related clean and green framework for shipbuilding
[150]. This will ultimately lead to adopting AI-driven community towards zero‐emissions: A strategic analysis from concept
solutions that improve maritime operations' sustainability, to case study,” Int. J. Energy Res., vol. 46, no. 14, pp. 20624–20649,
Nov. 2022, doi: 10.1002/er.7649.
efficiency, and safety. [11] B. Hu, “Application of Evaluation Algorithm for Port Logistics Park
Based on Pca-Svm Model,” Polish Marit. Res., vol. 25, no. s3, pp. 29–
IV. CONCLUSION 35, Dec. 2018, doi: 10.2478/pomr-2018-0109.
[12] Z. Domachowski, “Minimizing Greenhouse Gas Emissions From
In conclusion, our investigation sheds light on the Ships Using a Pareto Multi-Objective Optimization Approach,” Polish
significant part that artificial intelligence plays in facilitating Marit. Res., vol. 28, no. 2, pp. 96–101, Jun. 2021, doi:10.2478/pomr-
the shift to environmentally responsible practices in the 2021-0026.
[13] M. I. Lamas, R. C.G., T. J., and R. J.D., “Numerical Analysis of
marine industry, in conjunction with using alternative fuels Emissions from Marine Engines Using Alternative Fuels,” Polish
and renewable energy sources. Marit. Res., vol. 22, no. 4, pp. 48–52, Dec. 2015, doi: 10.1515/pomr-
 AI facilitates a shift to environmentally responsible 2015-0070.
practices in the marine industry. [14] H. P. Nguyen, P. Q. P. Nguyen, D. K. P. Nguyen, V. D. Bui, and D. T.
Nguyen, “Application of IoT Technologies in Seaport Management,”
 Utilizing predictive maintenance systems, route
JOIV Int. J. Informatics Vis., vol. 7, no. 1, pp. 228–240, Mar. 2023,
optimization algorithms, and autonomous shipping doi: 10.30630/joiv.7.1.1697.
capabilities. [15] H. P. Nguyen, N. D. K. Pham, and V. D. Bui, “Technical-
 Enhancing decision-making processes and vessel Environmental Assessment of Energy Management Systems in Smart
Ports,” Int. J. Renew. Energy Dev., vol. 11, no. 4, pp. 889–901, Nov.
operations and reducing pollutants and fuel
2022, doi: 10.14710/ijred.2022.46300.
consumption. [16] C. Wang, J. Shen, P. Vijayakumar, and B. B. Gupta, “Attribute-Based
 Recommendations for public policy include financial Secure Data Aggregation for Isolated IoT-Enabled Maritime
incentives, legal frameworks, and capacity-building Transportation Systems,” IEEE Trans. Intell. Transp. Syst., pp. 1–10,
2021, doi: 10.1109/TITS.2021.3127436.
activities.
[17] C. D. Paternina-Arboleda, D. Agudelo-Castañeda, S. Voß, and S. Das,
 Urgent need for action to address climate change and “Towards Cleaner Ports: Predictive Modeling of Sulfur Dioxide
environmental deterioration in the marine industry. Shipping Emissions in Maritime Facilities Using Machine Learning,”
 Shifting to alternate fuels, renewable energy, and AI to Sustainability, vol. 15, no. 16, p. 12171, Aug. 2023,
doi:10.3390/su151612171.
reduce environmental impact and pollution.
[18] V. G. Nguyen et al., “Using Artificial Neural Networks for Predicting
 Future success requires a holistic strategy incorporating Ship Fuel Consumption,” Polish Marit. Res., vol. 30, no. 2, pp. 39–60,
AI, alternative fuels, and renewable energy. Jun. 2023, doi: 10.2478/pomr-2023-0020.
 Policymakers should prioritize AI research, encourage [19] V. D. Bui and H. P. Nguyen, “Role of Inland Container Depot System
in Developing the Sustainable Transport System,” Int. J. Knowledge-
stakeholder cooperation, and create regulatory settings
Based Dev., vol. 12, no. 3/4, p. 1, 2022,
conducive to collaboration. doi:10.1504/IJKBD.2022.10053121.
[20] H. Geerlings and R. Van Duin, “A new method for assessing CO2-
REFERENCES emissions from container terminals: a promising approach applied in
Rotterdam,” J. Clean. Prod., vol. 19, no. 6–7, pp. 657–666, 2011.
[1] X. P. Nguyen, “Solutions for navigated safety of super-tankers [21] M. Fitzmaurice, “The International Convention for the Prevention of
operating on Dinh River traffic lanes and PTSC port,” 2020, p. 020017. Pollution from Ships (MARPOL),” in Research Handbook on Ocean
doi: 10.1063/5.0000141. Governance Law, Edward Elgar Publishing, 2023, pp. 91–108.
[2] A. T. Hoang et al., “Technological solutions for boosting hydrogen doi:10.4337/9781839107696.00019.
role in decarbonization strategies and net-zero goals of world shipping: [22] P. Brodie, “International Convention for the Prevention of Pollution
Challenges and perspectives,” Renew. Sustain. Energy Rev., vol. 188, from Ships (MARPOL),” in Commercial Shipping Handbook, Informa
p. 113790, Dec. 2023, doi: 10.1016/j.rser.2023.113790. Law from Routledge, 2014, pp. 219–221.
[3] A. T. Hoang et al., “Energy-related approach for reduction of CO2 doi:10.4324/9781315774695-85.
emissions: A critical strategy on the port-to-ship pathway,” J. Clean. [23] X. P. Nguyen and D. K. Pham Nguyen, “Experimental Research on
Prod., vol. 355, p. 131772, Jun. 2022, the Impact of Anchor-Cable Tensions in Mooring Ship at Vung Tau
doi:10.1016/j.jclepro.2022.131772. Anchorage Area,” Int. J. Adv. Sci. Eng. Inf. Technol., vol. 9, no. 6, pp.
[4] J. Kowalski and W. Tarelko, “NOx emission from a two-stroke ship 1892–1899, 2019.

13
[24] X. P. Nguyen, “A simulation study on the effects of hull form on [43] O. K. Ahmed, R. W. Daoud, S. M. Bawa, and A. H. Ahmed,
aerodynamic performances of the ships,” 2020, p. 020015. “Optimization of PV/T Solar Water Collector based on Fuzzy Logic
doi:10.1063/5.0000140. Control.,” Int. J. Renew. Energy Dev., vol. 9, no. 2, pp. 303–310, 2020.
[25] D. T. Nguyen and H. C. Le, “Potential of jute fiber-reinforced [44] A. Y. Kian and S. C. Lim, “On the Potential of Solar Energy for
composites in the manufacture of components and equipment used on Chemical and Metal Manufacturing Plants in Malaysia,” Int. J. Adv.
ships and hulls,” J. Emerg. Sci. Eng., vol. 1, no. 1, pp. 14–21, Sep. Sci. Eng. Inf. Technol., vol. 13, no. 5, pp. 1898–1904, 2023,
2023, doi: 10.61435/jese.2023.3. doi:10.18517/ijaseit.13.5.19052.
[26] V. D. Bui and H. P. Nguyen, “Sustainable development of Vietnam’s [45] I. Ismail, A. H. Ismail, and G. H. N. Nur Rahayu, “Wind Energy
transportation from analysis of car freight management,” Int. J. Feasibility Study of Seven Potential Locations in Indonesia,” Int. J.
Knowledge-Based Dev., vol. 12, no. 2, p. 77, 2021, Adv. Sci. Eng. Inf. Technol., vol. 10, no. 5, pp. 1970–1978, 2020,
doi:10.1504/IJKBD.2021.10045882. doi:10.18517/ijaseit.10.5.10389.
[27] H. P. Nguyen, P. Q. P. Nguyen, and T. P. Nguyen, “Green Port [46] D. H. Barus and R. Dalimi, “Multi-Stage Statistical Approach to Wind
Strategies in Developed Coastal Countries as Useful Lessons for the Power Forecast Errors Evaluation: A Southern Sulawesi Case Study,”
Path of Sustainable Development: A case study in Vietnam,” Int. J. Int. J. Adv. Sci. Eng. Inf. Technol., vol. 11, no. 2, pp. 633–641, 2021,
Renew. Energy Dev., vol. 11, no. 4, pp. 950–962, Nov. 2022, doi: 10.18517/ijaseit.11.2.12385.
doi:10.14710/ijred.2022.46539. [47] H. Zhang, Y. Hu, and J. He, “Wind Tunnel Experiment of Multi-Mode
[28] K. Rudzki, P. Gomulka, and A. T. Hoang, “Optimization Model to ARC Sail Device,” Polish Marit. Res., vol. 28, no. 4, pp. 20–29, Jan.
Manage Ship Fuel Consumption and Navigation Time,” Polish Marit. 2022, doi: 10.2478/pomr-2021-0046.
Res., vol. 29, no. 3, pp. 141–153, Sep. 2022, doi: 10.2478/pomr-2022- [48] Y. Shi and W. Luo, “Application of Solar Photovoltaic Power
0034. Generation System in Maritime Vessels and Development of Maritime
[29] T. Chu Van, J. Ramirez, T. Rainey, Z. Ristovski, and R. J. Brown, Tourism,” Polish Marit. Res., vol. 25, no. s2, pp. 176–181, Aug. 2018,
“Global impacts of recent IMO regulations on marine fuel oil refining doi: 10.2478/pomr-2018-0090.
processes and ship emissions,” Transp. Res. Part D Transp. Environ., [49] A. A. Salem and I. S. Seddiek, “Techno-Economic Approach to Solar
vol. 70, pp. 123–134, May 2019, doi:10.1016/j.trd.2019.04.001. Energy Systems Onboard Marine Vehicles,” Polish Marit. Res., vol.
[30] P. Sharma and A. K. Sharma, “AI-Based Prognostic Modeling and 23, no. 3, pp. 64–71, Sep. 2016, doi: 10.1515/pomr-2016-0033.
Performance Optimization of CI Engine Using Biodiesel-Diesel [50] X. P. Nguyen and V. H. Dong, “A study on traction control system for
Blends,” Int. J. Renew. Energy Res., no. v11i2, 2021, solar panel on vessels,” in International Conference on Emerging
doi:10.20508/ijrer.v11i2.11854.g8191. Applications in Material Science and Technology, ICEAMST 2020,
[31] S. Serbin, B. Diasamidze, V. Gorbov, and J. Kowalski, “Investigations 2020, p. 020016. doi: 10.1063/5.0007708.
of the Emission Characteristics of a Dual-Fuel Gas Turbine [51] M. F. Fazri, Lintang Bayu Kusuma, Risa Burhani Rahmawan,
Combustion Chamber Operating Simultaneously on Liquid and Hardiana Nur Fauji, and Castarica Camille, “Implementing Artificial
Gaseous Fuels,” Polish Marit. Res., vol. 28, no. 2, pp. 85–95, Jun. 2021, Intelligence to Reduce Marine Ecosystem Pollution,” IAIC Trans.
doi: 10.2478/pomr-2021-0025. Sustain. Digit. Innov., vol. 4, no. 2, pp. 101–108, Feb. 2023,
[32] Z. Stelmasiak, J. Larisch, J. Pielecha, and D. Pietras, “Particulate doi:10.34306/itsdi.v4i2.579.
Matter Emission from Dual Fuel Diesel Engine Fuelled with Natural [52] K. Bakker, “Smart Oceans: Artificial intelligence and marine protected
Gas,” Polish Marit. Res., vol. 24, no. 2, pp. 96–104, Jun. 2017, area governance,” Earth Syst. Gov., vol. 13, p. 100141, Aug. 2022, doi:
doi:10.1515/pomr-2017-0055. 10.1016/j.esg.2022.100141.
[33] Y. Li, B. Li, F. Deng, Q. Yang, and B. Zhang, “Research on the [53] Ni. Agarwala, “Managing Marine Environmental Pollution using
Application of Cold Energy of Largescale Lng-Powered Container Artificial Intelligence,” Marit. Technol. Res., vol. 3, no. 2, p.
Ships to Refrigerated Containers,” Polish Marit. Res., vol. 28, no. 4, Manuscript, Jan. 2021, doi: 10.33175/mtr.2021.248053.
pp. 107–121, Jan. 2022, doi: 10.2478/pomr-2021-0053. [54] A. Kheiter, S. Souag, A. Chaouch, A. Boukortt, B. Bekkouche, and M.
[34] V. N. Nguyen et al., “Understanding fuel saving and clean fuel Guezgouz, “Energy Management Strategy Based on Marine Predators
strategies towards green maritime,” Polish Marit. Res., vol. 30, no. 2, Algorithm for Grid-Connected Microgrid,” Int. J. Renew. Energy Dev.,
pp. 146–164, 2023, doi: 10.2478/pomr-2023-0030. vol. 11, no. 3, pp. 751–765, Aug. 2022, doi:10.14710/ijred.2022.42797.
[35] S. reza seyyedi, E. Kowsari, S. Ramakrishna, M. Gheibi, and A. [55] Y. Nain and O. Chi, “Application of Nonlinear Autoregressive Neural
Chinnappan, “Marine plastics, circular economy, and artificial Network to Model and Forecast Time Series Global Price of Bananas,”
intelligence: A comprehensive review of challenges, solutions, and Int. J. Data Sci., vol. 2, no. 1, pp. 19–37, 2021.
policies,” J. Environ. Manage., vol. 345, p. 118591, Nov. 2023, [56] B. V. Duc and H. P. Nguyen, “A Comprehensive Review on Big Data-
doi:10.1016/j.jenvman.2023.118591. Based Potential Applications in Marine Shipping Management,” Int. J.
[36] F. P. García Márquez, M. Papaelias, and S. Marini, “Artificial Adv. Sci. Eng. Inf. Technol., vol. 11, no. 3, pp. 1067–1077, 2021.
Intelligence in Marine Science and Engineering,” J. Mar. Sci. Eng., [57] T. T. Le et al., “Management strategy for seaports aspiring to green
vol. 10, no. 6, p. 711, May 2022, doi: 10.3390/jmse10060711. logistical goals of IMO: Technology and policy solutions,” Polish
[37] E. M. Ditria, C. A. Buelow, M. Gonzalez-Rivero, and R. M. Connolly, Marit. Res., vol. 30, no. 2, pp. 165–187, 2023, doi: 10.2478/pomr-
“Artificial intelligence and automated monitoring for assisting 2023-0031.
conservation of marine ecosystems: A perspective,” Front. Mar. Sci., [58] H. P. Nguyen, P. Q. P. Nguyen, and V. D. Bui, “Applications of Big
vol. 9, Jul. 2022, doi: 10.3389/fmars.2022.918104. Data Analytics in Traffic Management in Intelligent Transportation
[38] V. H. Dong and P. Sharma, “Optimized conversion of waste vegetable Systems,” JOIV Int. J. Informatics Vis., vol. 6, no. 1–2, pp. 177–187,
oil to biofuel with Meta heuristic methods and design of experiments,” May 2022, doi: 10.30630/joiv.6.1-2.882.
J. Emerg. Sci. Eng., vol. 1, no. 1, pp. 22–28, Sep. 2023, doi: [59] B. Aprilia, M. Marzuki, I. Taufiq, and F. Renggono, “Development of
10.61435/jese.2023.4. a Method for Classifying Convective and Stratiform Rains from Micro
[39] S. Zhu et al., “A Triboelectric Nanogenerator Based on a Pendulum- Rain Radar ( MRR ) Observation Data Using Artificial Neural
Plate Wave Energy Converter,” Polish Marit. Res., vol. 29, no. 4, pp. Network,” Int. J. Data Sci., vol. 3, no. 2, pp. 71–79, 2022.
155–161, Dec. 2022, doi: 10.2478/pomr-2022-0053. [60] Y. Antonisfia, Y. Silvia, and M. Botto-tobar, “E-Nose Application for
[40] P. Geng, X. Xu, and T. Tarasiuk, “State of Charge Estimation Method Detecting Banana Fruit Ripe Levels Using Artificial Neural Network
for Lithium-Ion Batteries in All-Electric Ships Based on LSTM Neural Backpropagation Method,” Int. J. Data Sci., vol. 3, no. 1, pp. 11–18,
Network,” Polish Marit. Res., vol. 27, no. 3, pp. 100–108, Sep. 2020, 2022.
doi: 10.2478/pomr-2020-0051. [61] B. Aruwa, A. Taye, and O. Adegoke, “Adaptive Android APKs
[41] J. Ugwu, K. C. Odo, L. O. Oluka, and K. O. Salami, “A Systematic Reverse Engineering for Features Processing in Machine Learning
Review on the Renewable Energy Development, Policies and Malware Detection,” Int. J. Data Sci., vol. 4, no. 1, pp. 10–25, 2023.
Challenges in Nigeria with an International Perspective and Public [62] V. G. Nguyen et al., “An extensive investigation on leveraging
Opinions,” Int. J. Renew. Energy Dev., vol. 11, no. 1, pp. 287–308, machine learning techniques for high-precision predictive modeling of
Feb. 2022, doi: 10.14710/ijred.2022.40359. CO 2 emission,” Energy Sources, Part A Recover. Util. Environ. Eff.,
[42] D. D. S. Garcia-Marquez, I. Andrade-Gonzalez, A.-M. Chavez- vol. 45, no. 3, pp. 9149–9177, Aug. 2023,
Rodriguez, M. I. Montero-Cortes, and V. S. Farias-Cervantes, doi:10.1080/15567036.2023.2231898.
“Prototype of a Solar Collector with the Recirculation of Nanofluids [63] R. V. N. Seutche, M. Sawadogo, and F. N. Ngassam, “Valuation of
for a Convective Dryer,” Int. J. Renew. Energy Dev., vol. 11, no. 4, pp. CO2 Emissions Reduction from Renewable Energy and Energy
1124–1133, Nov. 2022, doi: 10.14710/ijred.2022.44221. Efficiency Projects in Africa: A Case Study of Burkina Faso,” Int. J.

14
 Renew. Energy Dev., vol. 10, no. 4, pp. 713–729, Nov. 2021, [84] P.-C. Wu and C.-Y. Lin, “Cost-Benefit Evaluation on Promising
doi:10.14710/ijred.2021.34566. Strategies in Compliance with Low Sulfur Policy of IMO,” J. Mar. Sci.
[64] H. W. Asrini, G. W. Wicaksono, and B. Budiono, “Curriculum Eng., vol. 9, no. 1, p. 3, Dec. 2020, doi: 10.3390/jmse9010003.
Management Systems for Blended Learning Support,” JOIV Int. J. [85] O. Schinas and C. N. Stefanakos, “Selecting technologies towards
Informatics Vis., vol. 7, no. 4, pp. 2189–2197, 2023. compliance with MARPOL Annex VI: The perspective of operators,”
[65] R. Kesavan, N. Palanichamy, and T. Thirumurugan, “IoT and Deep Transp. Res. Part D Transp. Environ., vol. 28, pp. 28–40, May 2014,
Learning Enabled Smart Solutions for Assisting Menstrual Health doi: 10.1016/j.trd.2013.12.006.
Management for Rural Women in India: A Review,” JOIV Int. J. [86] IMO, “Guidelines for Consistent Implementation of the 0.50%
Informatics Vis., vol. 7, no. 4, pp. 2198–2205, 2023. Sulphur Limit Under MARPOL Annex VI,” 2019.
[66] A. H. Rangkuti, V. H. Athala, F. H. Indallah, E. Tanuar, and J. M. [87] J. Chen, Y. Fei, and Z. Wan, “The relationship between the
Kerta, “Optimization of Historic Buildings Recognition: CNN Model development of global maritime fleets and GHG emission from
and Supported by Pre-processing Methods,” JOIV Int. J. Informatics shipping,” J. Environ. Manage., vol. 242, pp. 31–39, Jul. 2019,
Vis., vol. 7, no. 4, pp. 2230–2239, 2023. doi:10.1016/j.jenvman.2019.03.136.
[67] I. Huda, A. A. Suhendra, and M. A. Bijaksana, “Design of Prediction [88] K. Aalbu and T. Longva, “From Progress to Delay: The Quest for Data
Model using Data Mining for Segmentation and Classification in the Negotiations on Greenhouse Gases in the International Maritime
Customer Churn in E-Commerce Mall in Mall,” JOIV Int. J. Organization,” Glob. Environ. Polit., vol. 22, no. 2, pp. 136–155, May
Informatics Vis., vol. 7, no. 4, pp. 2280–2289, 2023. 2022, doi: 10.1162/glep_a_00653.
[68] A. E. K. Gunawan and A. Wibowo, “Stock Price Movement [89] P. Serra and G. Fancello, “Towards the IMO’s GHG Goals: A Critical
Classification Using Ensembled Model of Long Short-Term Memory Overview of the Perspectives and Challenges of the Main Options for
(LSTM) and Random Forest (RF),” JOIV Int. J. Informatics Vis., vol. Decarbonizing International Shipping,” Sustainability, vol. 12, no. 8,
7, no. 4, pp. 2255–2262, 2023. p. 3220, Apr. 2020, doi: 10.3390/su12083220.
[69] V. D. Bui and H. P. Nguyen, “A Comprehensive Review on Big Data- [90] T.-H. Joung, S.-G. Kang, J.-K. Lee, and J. Ahn, “The IMO initial
Based Potential Applications in Marine Shipping Management,” Int. J. strategy for reducing Greenhouse Gas(GHG) emissions, and its
Adv. Sci. Eng. Inf. Technol., vol. 11, no. 3, pp. 1067–1077, Jun. 2021, follow-up actions towards 2050,” J. Int. Marit. Safety, Environ. Aff.
doi: 10.18517/ijaseit.11.3.15350. Shipp., vol. 4, no. 1, pp. 1–7, Jan. 2020,
[70] M. Drzewiecki and J. Guziński, “Design of an Autonomous IoT Node doi:10.1080/25725084.2019.1707938.
Powered by a Perovskite-Based Wave Energy Converter,” Polish [91] A. Chircop, “The IMO Initial Strategy for the Reduction of GHGs
Marit. Res., vol. 30, no. 3, pp. 142–152, Sep. 2023, doi:10.2478/pomr- from International Shipping: A Commentary,” Int. J. Mar. Coast. Law,
2023-0047. vol. 34, no. 3, pp. 482–512, Aug. 2019, doi: 10.1163/15718085-
[71] P. J. B. Sanchez et al., “Use of UIoT for Offshore Surveys Through 13431093.
Autonomous Vehicles,” Polish Marit. Res., vol. 28, no. 3, pp. 175–189, [92] X. Guo, J. Li, and S. Huang, “Study on Trade Effects of Green
Sep. 2021, doi: 10.2478/pomr-2021-0044. Maritime Transport Efficiency: An Empirical Test for China Based on
[72] A. Al-Othman et al., “Artificial intelligence and numerical models in Trade Decision Model,” Sustainability, vol. 15, no. 16, p. 12327, Aug.
hybrid renewable energy systems with fuel cells: Advances and 2023, doi: 10.3390/su151612327.
prospects,” Energy Convers. Manag., vol. 253, p. 115154, 2022. [93] A. T. Hoang and V. D. Tran, “Experimental Analysis on the
[73] T. T. Le, J. C. Priya, H. C. Le, N. V. L. Le, T. B. N. Nguyen, and D. Ultrasound-based Mixing Technique Applied to Ultra-low Sulphur
N. Cao, “Harnessing artificial intelligence for data-driven energy Diesel and Bio-oils,” Int. J. Adv. Sci. Eng. Inf. Technol., vol. 9, no. 1,
predictive analytics: A systematic survey towards enhancing p. 307, Feb. 2019, doi: 10.18517/ijaseit.9.1.7890.
sustainability,” Int. J. Renew. Energy Dev., vol. 13, no. 2, 2024, [94] V. D. Tran, A. T. Le, and A. T. Hoang, “An Experimental Study on
doi:10.61435/ijred.2024.60119. the Performance Characteristics of a Diesel Engine Fueled with
[74] A. B. Kanase-Patil, A. P. Kaldate, S. D. Lokhande, H. Panchal, M. ULSD-Biodiesel Blends,” Int. J. Renew. Energy Dev., vol. 10, no. 2,
Suresh, and V. Priya, “A review of artificial intelligence-based pp. 183–190, May 2021, doi: 10.14710/ijred.2021.34022.
optimization techniques for the sizing of integrated renewable energy [95] V. V. Pham and A. T. Hoang, “Aalyzing and selecting the typical
systems in smart cities,” Environ. Technol. Rev., vol. 9, no. 1, pp. 111– propulsion systems for ocean supply vessels,” in 2020 6th
136, 2020. International Conference on Advanced Computing and
[75] T. Uyanık, Ç. Karatuğ, and Y. Arslanoğlu, “Machine learning based Communication Systems (ICACCS), IEEE, Mar. 2020, pp. 1349–1357.
visibility estimation to ensure safer navigation in strait of Istanbul,” doi: 10.1109/ICACCS48705.2020.9074276.
Appl. Ocean Res., vol. 112, p. 102693, Jul. 2021, [96] V. V. Pham, A. T. Hoang, and H. C. Do, “Analysis and evaluation of
doi:10.1016/j.apor.2021.102693. database for the selection of propulsion systems for tankers,” p.
[76] R. Sultanbekov, I. Beloglazov, S. Islamov, and M. Ong, “Exploring of 020034, 2020, doi: 10.1063/5.0007655.
the Incompatibility of Marine Residual Fuel: A Case Study Using [97] O. Kanifolskyi, “General Strength, Energy Efficiency (EEDI), and
Machine Learning Methods,” Energies, vol. 14, no. 24, p. 8422, Dec. Energy Wave Criterion (EWC) of Deadrise Hulls for Transitional
2021, doi: 10.3390/en14248422. Mode,” Polish Marit. Res., vol. 29, no. 3, pp. 4–10, Sep. 2022,
[77] T. Song et al., “A review of artificial intelligence in marine science,” doi:10.2478/pomr-2022-0021.
Front. Earth Sci., vol. 11, Feb. 2023, doi:10.3389/feart.2023.1090185. [98] T. Vidović, J. Šimunović, G. Radica, and Ž. Penga, “Systematic
[78] Z. Islam Rony et al., “Alternative fuels to reduce greenhouse gas Overview of Newly Available Technologies in the Green Maritime
emissions from marine transport and promote UN sustainable Sector,” Energies, vol. 16, no. 2, p. 641, Jan. 2023,
development goals,” Fuel, vol. 338, p. 127220, Apr. 2023, doi:10.3390/en16020641.
doi:10.1016/j.fuel.2022.127220. [99] K. Cullinane and R. Bergqvist, “Emission control areas and their
[79] C. D. Mantoju, “Analysis of MARPOL implementation based on port impact on maritime transport,” Transp. Res. Part D Transp. Environ.,
state control statistics,” J. Int. Marit. Safety, Environ. Aff. Shipp., vol. vol. 28, pp. 1–5, May 2014, doi: 10.1016/j.trd.2013.12.004.
5, no. 3, pp. 132–145, Jul. 2021, doi:10.1080/25725084.2021.1965281. [100] T. Brewer, “A Maritime Emission Control Area for the Mediterranean
[80] N. Hussain, A. Khan, Shumaila, and S. Memon, “Addressing Marine Sea? Technological Solutions and Policy Options for a ‘Med ECA,’”
Pollution: An Analysis of MARPOL 73/78 Regulations and Global Euro-Mediterranean J. Environ. Integr., vol. 5, no. 1, p. 15, Apr. 2020,
Implementation Efforts,” J. Soc. Sci. Rev., vol. 3, no. 1, pp. 572–589, doi: 10.1007/s41207-020-00155-1.
Mar. 2023, doi: 10.54183/jssr.v3i1.193. [101] S. Karamperidis, C. Kapalidis, and T. Watson, “Maritime Cyber
[81] M. Issa, A. Ilinca, and F. Martini, “Ship Energy Efficiency and Security: A Global Challenge Tackled through Distinct Regional
Maritime Sector Initiatives to Reduce Carbon Emissions,” Energies, Approaches,” J. Mar. Sci. Eng., vol. 9, no. 12, p. 1323, Nov. 2021, doi:
vol. 15, no. 21, p. 7910, Oct. 2022, doi: 10.3390/en15217910. 10.3390/jmse9121323.
[82] M. H. Moradi, M. Brutsche, M. Wenig, U. Wagner, and T. Koch, [102] R. Hopcraft and K. M. Martin, “Effective maritime cybersecurity
“Marine route optimization using reinforcement learning approach to regulation – the case for a cyber code,” J. Indian Ocean Reg., vol. 14,
reduce fuel consumption and consequently minimize CO2 emissions,” no. 3, pp. 354–366, Sep. 2018, doi: 10.1080/19480881.2018.1519056.
Ocean Eng., vol. 259, p. 111882, Sep. 2022, [103] P. Louvros, F. Stefanidis, E. Boulougouris, A. Komianos, and D.
doi:10.1016/j.oceaneng.2022.111882. Vassalos, “Machine Learning and Case-Based Reasoning for Real-
[83] Z. Korczewski, “Energy and Emission Quality Ranking of Newly Time Onboard Prediction of the Survivability of Ships,” J. Mar. Sci.
Produced Low-Sulphur Marine Fuels,” Polish Marit. Res., vol. 29, no. Eng., vol. 11, no. 5, p. 890, Apr. 2023, doi: 10.3390/jmse11050890.
4, pp. 77–87, Dec. 2022, doi: 10.2478/pomr-2022-0045. [104] H. Lan, J. Gao, Y.-Y. Hong, and H. Yin, “Interval forecasting of

15
 photovoltaic power generation on green ship under Multi-factors [124] Z.-M. Yao, Z.-Q. Qian, R. Li, and E. Hu, “Energy efficiency analysis
coupling,” Sustain. Energy Technol. Assessments, vol. 56, p. 103088, of marine high-powered medium-speed diesel engine base on energy
Mar. 2023, doi: 10.1016/j.seta.2023.103088. balance and exergy,” Energy, vol. 176, pp. 991–1006, Jun. 2019,
[105] P. Gupta, A. Rasheed, and S. Steen, “Ship performance monitoring doi:10.1016/j.energy.2019.04.027.
using machine-learning,” Ocean Eng., vol. 254, p. 111094, Jun. 2022, [125] T. D. Eddy et al., “Energy Flow Through Marine Ecosystems:
doi: 10.1016/j.oceaneng.2022.111094. Confronting Transfer Efficiency,” Trends Ecol. Evol., vol. 36, no. 1,
[106] H. Tu, K. Xia, E. Zhao, L. Mu, and J. Sun, “Optimum trim prediction pp. 76–86, Jan. 2021, doi: 10.1016/j.tree.2020.09.006.
for container ships based on machine learning,” Ocean Eng., vol. 277, [126] L. Wang, J. Yao, H. Zhang, Q. Pang, and M. Fang, “A sustainable
p. 111322, Jun. 2023, doi: 10.1016/j.oceaneng.2022.111322. shipping management framework in the marine environment:
[107] X. Lang, D. Wu, and W. Mao, “Physics-informed machine learning Institutional pressure, eco-design, and cross-functional perspectives,”
models for ship speed prediction,” Expert Syst. Appl., vol. 238, p. Front. Mar. Sci., vol. 9, Jan. 2023, doi: 10.3389/fmars.2022.1070078.
121877, Mar. 2024, doi: 10.1016/j.eswa.2023.121877. [127] O. Oloruntobi, K. Mokhtar, A. Gohari, S. Asif, and L. F. Chuah,
[108] Y. E. Senol and A. Seyhan, “A novel machine-learning based “Sustainable transition towards greener and cleaner seaborne shipping
prediction model for ship manoeuvring emissions by using bridge industry: Challenges and opportunities,” Clean. Eng. Technol., vol. 13,
simulator,” Ocean Eng., vol. 291, p. 116411, Jan. 2024, doi: p. 100628, Apr. 2023, doi:10.1016/j.clet.2023.100628.
10.1016/j.oceaneng.2023.116411. [128] A. A. Yusuf, F. L. Inambao, and J. D. Ampah, “Evaluation of biodiesel
[109] M. Zhang, P. Kujala, M. Musharraf, J. Zhang, and S. Hirdaris, “A on speciated PM2.5, organic compound, ultrafine particle and gaseous
machine learning method for the prediction of ship motion trajectories emissions from a low-speed EPA Tier II marine diesel engine coupled
in real operational conditions,” Ocean Eng., vol. 283, p. 114905, Sep. with DPF, DEP and SCR filter at various loads,” Energy, vol. 239, p.
2023, doi: 10.1016/j.oceaneng.2023.114905. 121837, Jan. 2022, doi:10.1016/j.energy.2021.121837.
[110] L. P. Perera, K. Belibassakis, E. Filippas, and M. Premasiri, [129] C. W. Mohd Noor, M. M. Noor, and R. Mamat, “Biodiesel as
“Advanced Data Analytics Based Hybrid Engine-Propeller alternative fuel for marine diesel engine applications: A review,”
Combinator Diagram for Green Ship Operations,” in Volume 5A: Renew. Sustain. Energy Rev., vol. 94, pp. 127–142, Oct. 2018,
Ocean Engineering, American Society of Mechanical Engineers, Jun. doi:10.1016/j.rser.2018.05.031.
2022. doi: 10.1115/OMAE2022-79490. [130] M. Lashgari, A. A. Akbari, and S. Nasersarraf, “A new model for
[111] Y. Han, W. Ma, and D. Ma, “Green maritime: An improved quantum simultaneously optimizing ship route, sailing speed, and fuel
genetic algorithm-based ship speed optimization method considering consumption in a shipping problem under different price scenarios,”
various emission reduction regulations and strategies,” J. Clean. Prod., Appl. Ocean Res., vol. 113, p. 102725, Aug. 2021,
vol. 385, p. 135814, Jan. 2023, doi:10.1016/j.jclepro.2022.135814. doi:10.1016/j.apor.2021.102725.
[112] D. Ma, W. Ma, S. Hao, S. Jin, and F. Qu, “Ship’s response to low- [131] B. Rolf, I. Jackson, M. Müller, S. Lang, T. Reggelin, and D. Ivanov,
sulfur regulations: From the perspective of route, speed and refueling “A review on reinforcement learning algorithms and applications in
strategy,” Comput. Ind. Eng., vol. 155, p. 107140, May 2021, supply chain management,” Int. J. Prod. Res., vol. 61, no. 20, pp.
doi:10.1016/j.cie.2021.107140. 7151–7179, Oct. 2023, doi: 10.1080/00207543.2022.2140221.
[113] A. De, A. Choudhary, and M. K. Tiwari, “Multiobjective Approach for [132] S. El Mekkaoui, L. Benabbou, S. Caron, and A. Berrado, “Deep
Sustainable Ship Routing and Scheduling With Draft Restrictions,” Learning-Based Ship Speed Prediction for Intelligent Maritime Traffic
IEEE Trans. Eng. Manag., vol. 66, no. 1, pp. 35–51, Feb. 2019, doi: Management,” J. Mar. Sci. Eng., vol. 11, no. 1, p. 191, Jan. 2023, doi:
10.1109/TEM.2017.2766443. 10.3390/jmse11010191.
[114] Y. Yao, M. Wei, and B. Bai, “Descriptive statistical analysis of [133] W. Peng, X. Bai, D. Yang, K. F. Yuen, and J. Wu, “A deep learning
experimental data for wettability alteration with surfactants in approach for port congestion estimation and prediction,” Marit. Policy
carbonate reservoirs,” Fuel, vol. 310, p. 122110, Feb. 2022, Manag., vol. 50, no. 7, pp. 835–860, Oct. 2023,
doi:10.1016/j.fuel.2021.122110. doi:10.1080/03088839.2022.2057608.
[115] Y. Wang, X. Shao, C. Liu, G. Cai, L. Kou, and Z. Wu, “Analysis of [134] J. Jin, M. Ma, H. Jin, T. Cui, and R. Bai, “Container terminal daily gate
wind farm output characteristics based on descriptive statistical in and gate out forecasting using machine learning methods,” Transp.
analysis and envelope domain,” Energy, vol. 170, pp. 580–591, Mar. Policy, vol. 132, pp. 163–174, Mar. 2023,
2019, doi: 10.1016/j.energy.2018.12.156. doi:10.1016/j.tranpol.2022.11.010.
[116] Z. Said, P. Sharma, B. J. Bora, and A. K. Pandey, “Sonication impact [135] N. P. Ventikos and K. Louzis, “Developing next generation marine
on thermal conductivity of f-MWCNT nanofluids using XGBoost and risk analysis for ships: Bio-inspiration for building immunity,” Proc.
Gaussian process regression,” J. Taiwan Inst. Chem. Eng., vol. 145, p. Inst. Mech. Eng. Part O J. Risk Reliab., vol. 237, no. 2, pp. 405–424,
104818, Apr. 2023, doi: 10.1016/j.jtice.2023.104818. Apr. 2023, doi: 10.1177/1748006X221087501.
[117] T. Chen and C. Guestrin, “XGBoost,” in Proceedings of the 22nd ACM [136] A. Rawson and M. Brito, “A survey of the opportunities and
SIGKDD International Conference on Knowledge Discovery and Data challenges of supervised machine learning in maritime risk analysis,”
Mining, New York, NY, USA: ACM, Aug. 2016, pp. 785–794. doi: Transp. Rev., vol. 43, no. 1, pp. 108–130, Jan. 2023, doi:
10.1145/2939672.2939785. 10.1080/01441647.2022.2036864.
[118] A. T. Hoang and V. V. Pham, “A Review on Fuels Used for Marine [137] S. Filom, A. M. Amiri, and S. Razavi, “Applications of machine
Diesel Engines,” J. Mech. Eng. Res. Dev., vol. 41, no. 4, pp. 22–32, learning methods in port operations – A systematic literature review,”
Nov. 2018, doi: 10.26480/jmerd.04.2018.22.32. Transp. Res. Part E Logist. Transp. Rev., vol. 161, p. 102722, May
[119] V. V. Pham and A. T. Hoang, “Technological perspective for reducing 2022, doi: 10.1016/j.tre.2022.102722.
emissions from marine engines,” Int. J. Adv. Sci. Eng. Inf. Technol., [138] L. Barua, B. Zou, and Y. Zhou, “Machine learning for international
vol. 9, no. 6, pp. 1989–2000, 2019, doi: 10.18517/ijaseit.9.6.10429. freight transportation management: A comprehensive review,” Res.
[120] M. Rodriguez Valido, J. Perez Marrero, A. Mauro González, P. Transp. Bus. Manag., vol. 34, p. 100453, Mar. 2020, doi:
Fabiani Bendicho, and C. Efrem Mora, “Evaluation of the Potential of 10.1016/j.rtbm.2020.100453.
Sentinel-5P TROPOMI and AIS Marine Traffic Data for the [139] J. P. Panda, “Machine learning for naval architecture, ocean and
Monitoring of Anthropogenic Activity and Maritime Transport NOx- marine engineering,” J. Mar. Sci. Technol., vol. 28, no. 1, pp. 1–26,
Emissions in Canary Islands Waters,” Sustainability, vol. 15, no. 5, p. Mar. 2023, doi: 10.1007/s00773-022-00914-5.
4632, Mar. 2023, doi: 10.3390/su15054632. [140] Z. Li and Y. Liu, “An Improved Model for Marine Energy Price and
[121] R. Li, Y. Liu, and Q. Wang, “Emissions in maritime transport: A Efficiency Assessment,” J. Coast. Res., vol. 94, no. sp1, p. 640, Sep.
decomposition analysis from the perspective of production-based and 2019, doi: 10.2112/SI94-128.1.
consumption-based emissions,” Mar. Policy, vol. 143, p. 105125, Sep. [141] A. T. Hoang, A. R. Al-Tawaha, Lan Anh Vu, Van Viet Pham, A. M.
2022, doi: 10.1016/j.marpol.2022.105125. Qaisi, and J. Křeček, “Integrating Environmental Protection Education
[122] C. Sui, P. de Vos, D. Stapersma, K. Visser, and Y. Ding, “Fuel in the Curriculum: A Measure to Form Awareness of Environmental
Consumption and Emissions of Ocean-Going Cargo Ship with Hybrid Protection for the Community,” in Environmental Sustainability
Propulsion and Different Fuels over Voyage,” J. Mar. Sci. Eng., vol. Education for a Changing World, Cham: Springer International
8, no. 8, p. 588, Aug. 2020, doi: 10.3390/jmse8080588. Publishing, 2021, pp. 191–207. doi: 10.1007/978-3-030-66384-1_12.
[123] D. Huang, Y. Wang, and C. Yin, “Selection of CO2 Emission [142] T. A. T. Do, Q. T. Le, and T. N. D. Hoang, “Integration training
Reduction Measures Affecting the Maximum Annual Income of a information in Vietnam maritime university based on the Conceive-
Container Ship,” J. Mar. Sci. Eng., vol. 11, no. 3, p. 534, Mar. 2023, Design-Implement-Operate,” Int. J. Adv. Sci. Eng. Inf. Technol., vol.
doi: 10.3390/jmse11030534. 9, no. 3, pp. 1017–1024, 2019, doi: 10.18517/ijaseit.9.3.8260.

16
[143] T. Q. Le and T. T. A. Do, “Active teaching techniques for engineering the Development Strategy of Intelligent Transportation System in
students to ensure the learning outcomes of training programs by Vietnam,” Int. J. Adv. Sci. Eng. Inf. Technol., vol. 10, no. 2, pp. 520–
CDIO Approach,” Int. J. Adv. Sci. Eng. Inf. Technol., vol. 9, no. 1, pp. 528, Mar. 2020, doi: 10.18517/ijaseit.10.2.11129.
266–273, 2019. [148] V. T. Pham, “Critical information for vietnamese economy aiming at
[144] T. Q. Le, “Approaching CDIO to innovate the training program for a strategic breakthrough as approaching the industry 4.0,” Int. J. Adv.
seafarers to meet the requirements of the industrial revolution 4.0,” Int. Sci. Eng. Inf. Technol., vol. 9, no. 3, pp. 1008–1016, 2019.
J. Adv. Sci. Eng. Inf. Technol., vol. 9, no. 2, pp. 648–655, 2019. [149] A. Sharafi, H. Iranmanesh, M. S. Amalnick, and M. Abdollahzade,
[145] N. Assani, P. Matić, N. Kaštelan, and I. R. Ćavka, “A review of “Financial management of Public Private Partnership projects using
artificial neural networks applications in maritime industry,” IEEE artificial intelligence and fuzzy model,” Int. J. energy Stat., vol. 4, no.
access, 2023. 02, p. 1650007, 2016.
[146] S. Poongavanam, D. Rajesh, K. Viswanathan, and S. R. Banu, “Role [150] J. Baldoni, E. Begoli, D. F. Kusnezov, and J. MacWilliams, “Solving
and Challenges of Artificial Intelligence in the Maritime Industry,” J. hard problems with AI: dramatically accelerating drug discovery
Surv. Fish. Sci., vol. 10, no. 3S, pp. 6313–6317, 2023. through a unique public-private partnership,” J. Commer. Biotechnol.,
[147] H. P. Nguyen, “Core Orientations for 4.0 Technology Application on vol. 25, no. 4, 2020.

17