Journal of Optimization in Industrial Engineering,

Vol.17, Issue 2, Summer & Autumn 2024, 243-258

DOI: 10.22094/QJIE.2024.1107423

Industry 4.0- Applications of machine learning in the field of

industrial engineering: Systematic review of the literature
Marco Antonio Diaz Martinez a,*, Reina Veronica Roman Salinas b , Santos Ruiz Hernandez c,
Gabriela Cervantes Zubirias d, Mario Alberto Morales Rodriguez e
a
Posgraduate area- TecNM- Higher Technological Institute of Pánuco (ITSP)- Veracruz, México
b
Department of industrial engineering-TecNM- Higher Technological Institute of Pánuco (ITSP)- Veracruz, México
c
Posgraduate area- TecNM- Higher Technological Institute of Pánuco (ITSP)- Veracruz, México
d
Department of industrial engineering - Multidisciplinary Academic Unit, Reynosa-Aztlan-UAT- Tamaulipas, México
e
Department of industrial engineering - Multidisciplinary Academic Unit, Reynosa-Aztlan-UAT- Tamaulipas, México
Received 08 April 2024; Revised 05 August 2024; Accepted 15 August 2024

Abstract
The aim of this research is to determine how the implementation of machine learning has generated advantages in the field of engineering.
Through a systematic review of the literature, it seeks to understand the importance of machine learning and its various applications in
engineering, such as equipment maintenance, business demand forecasting, production chain optimization, customer service, and quality
control. In this article, we conduct a systematic review and bibliometric analysis to explore the current state of research on machine
learning and Industry 4.0 applications in the field of industrial engineering. Our goal is to identify established and emerging fields of
research to guide future research. To carry out this study, we initially identified 926 scientific journal publications indexed by publishers
such as Ebsco essentials, ScienceDirect, IEEEXplore, and MDPI, collected from 1 January of 2015 to June 2024. Subsequently, a group of
specialists evaluated these publications, carefully selecting 88 of them that were placed in the literature review section and that were
considered relevant to our analysis. In a second stage, we applied a detailed analysis using MAXQDA v.2020 software on our collected
data, focusing on citation and keyword evaluation. This approach allowed us to gain a deeper understanding of trends and connections in
existing research in this field.

Keywords: Machine Learning; Industry 4.0; Supply Chain; maintenance; Artificial intelligence; Deep Learning; Additive Manufacturing.
The intersection between engineering and machine
1. Introduction
learning technologies is reshaping the way business
Industry 4.0 has unleashed an unprecedented revolution in models are developed. Examples such as the Marketplace
engineering, completely redefining the conception, illustrate how better benefits can be achieved at lower
design, and operation of production and manufacturing cost by implementing ML-based strategies (Romero et al.,
systems. In this context, machine learning, also known as 2021). In the environment of the Industrial Revolution 4.0
machine learning (ML), emerges as an essential tool that (I4.0), characterized by interconnection, big data, cyber-
drives efficiency, accuracy, and innovation in a wide physical systems and machine learning, machine learning
range of engineering applications. From process and the Internet of Things (IoT) emerge as crucial
optimization to predicting machine failures, the use of elements for business transformation. Sectors such as e-
machine learning is transforming the way engineers commerce have seen remarkable progress thanks to these
approach challenges across industries, delivering smarter, technologies (Sazzadur et al., 2023; Hermann, 2015).
more adaptive solutions. The technological advancements of Industry 4.0 are
Machine learning has emerged as a critical component in redefining industrial production by meeting the demands
the Industry 4.0 framework, thanks to its ability to of smart technology. Early detection of critical events,
identify patterns in massive data sets and perform such as fatigue cracks in machine components, is made
predictive analysis without the need for manual possible by machine learning (Drakaki et al., 2022). This
programming. This revolutionary approach extends from capability has positioned ML as an essential component in
supply chain optimization to improving teaching by the digitalization of production system operations,
creating learning models. The support of researchers such highlighting its applicability in materials engineering and
as Duong et al. (2023) and Choudhary et al. (2022) shows other fields (Chen et al., 2023; Amin et al., 2021).
how ML can simplify entire experimental cycles and Although ML promises significant improvements, real-
reduce errors, thus promoting greater efficiency in time fault detection remains a challenge due to the
industrial processes.
*Corresponding author Email address: [Link]@[Link]

243
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…

complexity of mechatronic control systems (Velásquez et occupational health. This also helps companies reduce
al., 2022). costs, minimize downtime, and protect worker integrity.
Given its dynamism, Industry 4.0 represents a new Field 4- Machine Learning in Additive
paradigm where value chains are driven by intelligent Manufacturing: Optimizes and improves manufacturing
connectivity and automation. Machine learning emerges processes through 3D printing. Machine learning can be
as a critical tool for improving problem environments on used to predict and prevent part defects, optimize resource
a large scale, thereby increasing the performance of allocation, and reduce production times. Machine learning
industrial operations (Mazzei & Tancredi, 2022). This also allows businesses to take full advantage of the
approach opens new opportunities for resource capabilities of this emerging technology.
optimization and continuous improvement in industrial Field 5- Machine Learning and inventory control: It
processes, supported by digital advances and device involves the use of advanced machine learning techniques
interconnection (Kumar, 2023). and algorithms to manage and optimize inventories across
The purpose of this work is to investigate the progress of industries. This includes improving inventory
Machine Learning in industrial environments and its
replenishment, reducing storage costs, decreasing excess
impact on the optimization of production processes. In
inventory, and preventing product shortages. In addition,
addition, it analyzes how this technology contributes to
it enables automated, data-driven decisions, increasing the
improving supply chain management, predicting
efficiency and profitability of the inventory management
occupational risks and suggesting preventive measures to
promote occupational safety. system.
It also examines its influence on optimizing and refining Field 6- Machine Learning and marketing: It refers to
manufacturing using 3D printing, as well as its ability to the application of machine learning techniques and
help companies reduce costs. This research offers algorithms to improve and optimize marketing strategies.
recommendations that may be of interest to students and By analyzing large volumes of data, machine learning
academics seeking to better understand ML concepts and allows you to personalize advertising campaigns, segment
applications in various fields of study, as well as to audiences more precisely, predict consumer behavior, and
software engineers and developers, entrepreneurs, and optimize ad spend. In addition, it helps identify market
managers looking to develop new research and projects in trends, improve customer experience, and increase the
this ever-evolving area. effectiveness of marketing campaigns through automation
This study is based on a comprehensive systematic review and predictive analytics.
of the literature, details of which are presented in section
2. The methodology used for this review is described in 2.1. Machine learning in production systems
detail in section 3. Finally, in section 4, a detailed analysis
of the information collected is carried out to identify ML has made its presence known in different fields of
patterns and trends in publications related to the engineering such as the mining industry, where it is used
application of Machine Learning in the field of industrial to analyze the feasibility of production systems through
engineering. the intervention of intelligent technology. This application
has resulted in a significant improvement in production
2. Literature Review efficiency.
The use of imputation algorithms, such as Miceforest,
Before starting with the literature review, it is important have helped to quantify mining feasibility and establish
to describe each of the fields of application that will be optimal production conditions, contributing to an
addressed to have a better clarity of the research topic. accuracy rate of 96% (Kang et al., 2023). Another area of
Field 1- Machine Learning in Production Systems: engineering that has benefited is the oil industry, where
Refers to the applications of machine learning techniques the exploitation of tight oil reservoirs has become
within industrial or production environments. It also economically viable, thanks to the combination of
involves learning models used in industrial environments horizontal drilling and multistage hydraulic fracturing.
with the aim of optimizing processes. The incorporation of machine learning and big data has
Field 2- Machine Learning in the Supply Chain: Refers enabled the use of evolutionary algorithms to optimize
to the application of ML techniques to analyze data and fracturing parameters, such as fracture length, fracture
improve supply chain management. By integrating ML mean, fracture permeability, fracture spacing, and fracture
into the supply chain, businesses can reduce costs, height (Dong et al, 2022). In the petroleum industry,
improve operational efficiency, and more effectively meet
various geological factors are critical to the production of
customer needs.
oil in a well. Machine learning is a key tool for
Field 3- Machine learning in industrial safety: ML in
identifying these critical factors. Using techniques such as
industrial safety seeks to improve safety in work
Pearson's correlation tests and sensitivity analysis, aspects
environments by analyzing large volumes of data, these
systems can identify patterns, predict risks and suggest such as fluid volumes, saturation, gas condensation, and
preventive measures to avoid accidents and improve carbon content are examined. These statistical analyses
allow oil companies to develop predictive models and

244
 Journal of Optimization in Industrial Engineering, Vol.17, Issue 2, Summer & Autumn 2024, 243-258

better understand how these factors influence the algorithms (Xu, 2023). Another aspect benefited by
performance of an oil well (Guo et al., 2021). Machine Learning is the maintenance of production
In the field of manufacturing engineering, various equipment, this process strengthens the continuity of
conventional machine tools such as lathes, milling manufacturing or service operations. For the
machines and grinding machines are used to produce implementation of preventive maintenance to be effective,
mechanical parts. However, innovative technologies have reliable information from sensors is required to be used in
been introduced such as autonomous production the algorithms and give effective responses, given
machines, to perform the operations automatically. previous simulations with real data. It is important to
To improve the ability to forecast predictive maintenance consider that the greater the precision required in the
on equipment and components such as the spindle, processes, the more complex and expensive the solutions
intelligent learning-based defect detection modules are should be (Rojek et al., 2023).
employed. These modules can anticipate the severity of In the automotive field, ML is used in the production of
faults and failures in machines (Schlangenhauf, 2021). In vehicles with lightweight materials such as high-strength
the engineering field, one of the most critical performance steels, aluminum, and reinforced thermoplastics (FRP).
indicators for production control is overall equipment These materials have contributed to the reduction of fuel
effectiveness (OEE). This indicator can combine data consumption and, in the case of electric vehicles, have
from equipment and process performance to improve the significantly improved the autonomy (Hürkamp et al.,
quality of the final product. The incorporation of cyber- 2020). Thermoplastic materials possess specific
physical production systems (CPPS) and sensors for data mechanical properties that allow large-scale production
storage facilitates machine learning, maintains optimal thanks to short production cycles and efficient cost-
performance, improves security, and ensures the quality benefit ratio (Liebsch et al., 2019).
of results. The accuracy of ML results also depends on the Industry 4.0 and ML have expanded the possibilities in
suitability of the algorithms and the quality of the industrial control systems, enabling a high degree of
information used (Engelman et al, 2020; Wiemer, 2021). automation through process modeling and real-time
Artificial intelligence (AI) and ML have increased control. This promises to address challenges at all levels
production capabilities in additive manufacturing in three of automation, with the goal of controlling industrial
important areas; One of them is the manufacture of production systems (Jamwal et al., 2021). The use of
products with complex geometry and shapes, standing out machine learning (ML) algorithms has emerged as a
for the effective integration of components, improved powerful tool for addressing problems and exploring
performance, and development of compact designs for applications in a wide range of areas. The accelerated
mobile devices. Another field refers to large-scale progress of Industry 4.0 and cloud computing has made it
customization, applied in the manufacture of small possible to handle huge amounts of data and ML models,
quantities of devices that incorporate high-value thus improving the control of information in production
components, such as medical devices. The systems and optimizing the identification of parameters in
disintermediation of the supply chain is another area, processes (Pereira & Constantin, 2024). Within the
which focuses on the direct production of components, framework of Industry 4.0 and its integration with
highlighting on-site manufacturing as a crucial element. production systems, metal additive manufacturing (AM)
This enables rapid production in cases of unpredictable stands out, offering a wide variety of industrial
demand, especially in products intended for military or applications. Broadly speaking, AM consists of the
medical applications. It is essential to keep in mind that creation of components by fusing a base material, known
machines used in machine learning use technology that as raw material, using a high-powered, localized heat
requires supervision and must operate in controlled source, which is guided by the information contained in a
environments, subject to periodic maintenance schedules computer-aided design (CAD) file (Mo et al., 2024).
(Carpanzo & Togo, 2022). Monitored systems enable
efficient capture of system parameters and states in real 2.2. Machine learning in the supply chain
time, ensuring effective control in manufacturing
processes, resulting in an improvement in reliability, data The advent of Industry 4.0 has led to the development of
accuracy, and satisfaction for companies (Lee, 2021). technologies that improve supply chain management.
The internet of things (IoT) has transformed industries These technologies allow manufacturers to track the flow
into smart ones, contributing to the economic of information using tools that take advantage of available
development of modern nations by raising living data sources, in addition to using Big Data analytics
standards and improving working conditions. (BDA) and ML. These tools are essential in supply chain
Technologies such as 5G, block chain, edge computing management (Barzizza et al., 2023). The success of
and fog computing are helping to solve security organizations in supply chain management depends on
challenges by enabling efficient transmission of collaboration among team members and efficient
communication of information. This collaboration has
information, as they make it difficult for digital device
motivated companies to adopt improvements in their
systems to be compromised, thanks to data and learning
operations, such as data analysis, event prediction, and the

245
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…

proper handling of large volumes of information (Ali et For inventory management control, machine learning is
al., 2022). used in the use of drones. These devices can locate
Another aspect that has been incorporated into ML is products on shelves quickly, move light goods, detect
cloud-based blockchain technology, which provides a cycle inventories, make RFID readings, which increases
machine learning approach for supply chain management value in the supply chain. Delivery drones are emerging
applications, ensuring the sustainability of the process. as a reliable and promising means of transport in this
The adoption of cloud computing has played an essential chain (Lei, 2022; Bridgeball, 2023).
role in the scalability of blockchains, addressing issues Another aspect that has been present along with machine
such as availability, storage capacity, latency, and learning is the Internet of Things, which has made notable
throughput (Wong et al., 2023; Kim et al., 2019). progress in the logistics field, specifically in warehouse
The current relevance of ML stems from the convergence management, communication systems, quality of service
of various mutually reinforcing trends, making ML an and procurement. The application of recurrent neural
effective tool thanks to the abundance of digital data, network models with gates (GRU) and convolutional
which is fundamental to most deep learning algorithms. In neural networks (CNNs) has enhanced the ability to
addition, another factor that contributes to safer learning anticipate and manage information effectively, reducing
and effective decision-making is advanced natural users' dependence on constant monitoring of the data
language processing neural networks (Schroeder, 2021; sequence (Alzahrani, 2023).
Roh et al., 2019). For many SMEs (Small and medium-sized enterprises),
Enterprises should be aware of the challenges related to implementing the use of technology for supply chain
data transformation and manipulation as data transactions management, as well as to maintain it efficiently, is a
in supply chain management networks become more challenge and requires the granting of financing. In this
immutable. To effectively manage the large amount of sense, in the absence of credit risk predictions from
data generated, it is crucial to have a robust cloud traditional methods, the implementation of machine
infrastructure that offers scalability and Big Data learning algorithms such as the recurrent neural network
processing. In the maritime logistics domain, systems of and the short-term memory model, has made it possible to
this type contribute to risk reduction by guiding cargo efficiently predict to grant and in turn monitor credits,
ships along planned and safe routes, avoiding deviations since they can identify fraud, through financial data
and dangerous areas (Wong et al., 2021). To achieve analysis, economic reports and trends, as well as
optimal supply chain management, it is essential to rely forecasting consumer and market behavior (Xia et al.,
on statistical evidence and operational research to 2023). Machine learning also plays a critical role in
optimize the fulfillment of product or service demand optimizing the supply chain in the field of biofuels and
targets. While the adoption of technologies related to seed cultivation. Its purpose is to identify methods that
blockchain enhancement is critical (see Figure 1), so is ensure efficient material flows between suppliers,
the implementation of a supply chain network designed to warehouses, production facilities, and end customers.
predict consumer perception and meet the needs of Supply chain management has become an essential tool
specific events (Khchine et al., 2018). for coordinating all phases of the process, recognizing
that any disruption at a specific point can significantly
affect the overall performance of the chain (Kim, 2024).
In terms of seed cultivation, it is crucial to develop
effective non-destructive methods for the classification of
millet seed cultivars, with applications in the detection of
cultivar contamination that could occur at various points
along the cereal supply chain (Ekramirad et al., 2024).

2.3. Machine learning and industrial security

Machine learning has played a critical role in advancing

industrial risk assessments during the early exploration,
stable development, and high-speed development phases.
Fig. 1. Blockchain technology. This has had a significant impact on the safety of workers
and the work environment, by allowing more effective
Supply chain management faces the complex challenges monitoring of risks associated with construction and
of meeting each phase of supply and demand, machinery areas, as well as environmental protection
compounded by globalization. In response to these (Smith, 2020; Wei et al., 2023).
challenges, companies are adopting data-driven Another positive aspect in terms of safety for companies
technologies and machine learning to overcome obstacles lies in the detection of hazards related to landslides. This
in communication channels, distribution times, inventory is achieved through the application of advanced
management, among others (Sani et al., & Kumar, 2023). techniques such as neural networks, classification

246
 Journal of Optimization in Industrial Engineering, Vol.17, Issue 2, Summer & Autumn 2024, 243-258

algorithms, and principal component analysis, as well as

the use of inferential and deductive statistical methods. 2.4. Machine learning and the additive manufacturing
These strategies make it possible to develop an effective (AM)
methodology for the creation of susceptibility maps,
taking into account the specific characteristics of the In the engineering domain, the expansion of IoT and ML
terrain, which is essential in the proper planning of their has significantly evolved the demand for location-based
use (Tien et al, 2015). services (LBS), machine learning and industrial
Observation of operations in industrial processes provides manufacturing equipment (Rathnayake et al., 2023).
the rapid identification of potential faults, which helps to Additive manufacturing (AM) is one of the most
ensure safety in production. Approaches have been innovative processes that is based on the extrusion and
developed that integrate techniques such as Fisher design of complex 3D parts. In this sense, ML presents a
discrimination (VWDA), distributed dimensionality promising approach to identify suitable parameters in the
reduction in t (t-SNE) and extreme learning (ELM) to manufacture of parts with the most suitable properties,
carry out visual monitoring of processes. VWDA assigns without the need to analytically model the entire process
weights to variables based on their impact on potential (Pelzer et al., 2023; Gibson, 2015 & Heng et al., 2024).
failures, represented as vectors. The t-SNE visualizes the Additive manufacturing offers broad design flexibility
characteristics of these vectors, and finally, ELM is used and the ability to modify material properties. It is
to map the input data to their respective mapping points, important to mention that, in additive manufacturing, a
thus allowing real-time monitoring of the process (Lu, powder bed laser fusion process is used for metals
2021). (Khorasani et al., 2020; Ngo et al., 2018), which can give
rise to imperfections such as porosity, which can
The use of learning algorithms in security applications
influence the mechanical characteristics of the metal
must address various vulnerabilities, including the components. Therefore, to improve the accuracy of defect
verification and interpretation of performance constraints. detection, Random Forest models have been developed
There are also training approaches and methods, such as that are capable of classifying images from micrographs,
ML Security Taxonomy, that can decrease generalization achieving an impressive 95% accuracy (Altmann et al.,
errors, detect attacks in production environments, employ 2023). AM has driven notable advances in design,
advanced machine learning techniques and diversified industrial processes, and manufacturing technologies. ML
strategies (Mohseni et al., 2022). has allowed people to surpass themselves, as they are
Another aspect associated with ML is the internet of adapting to new technologies and algorithms that are
things (IoT), which is used to identify intruders and providing benefits in the execution of tasks and in the
threats in the IoT working environment by using deep analysis of effectiveness in real time for more informed
decision making.
learning (DLIS). It is important to note that as the number
Although additive manufacturing is one of the most
of IoT-connected devices increases, security becomes
significant technologies that Industry 4.0 presents to
even more crucial due to the nature of its architecture.
engineering, it is not exempt from failures due to fatigue
Special modules exist that combine optimization
of equipment components or the presence of defects in the
algorithms and stacked deep polynomial type networks
processed parts, which complicates the design process and
(SDPNs) to achieve more accurate detection of
makes it less precise. These situations, by default, would
anomalous data and better performance in terms of
imply carrying out additional evaluations to determine the
accuracy and recall (Otoum, 2021; Kouicem, 2018 &
existence of said problems in the internal structure of the
Tariq et al., 2019).
materials, which would imply expensive experimental
While the implementation of ML and technologies
tests. All these challenges can be effectively addressed by
associated with Industry 4.0 has provided benefits that are
ML algorithms, which would not only evaluate the limit
useful for workers involved in the design and processes of
fatigue but also predict the fatigue of the components.
manufacturing and industrial machine systems, it is
This is how AM has driven remarkable advances in
important to consider the risks associated with artificial
design, industrial processes and ML-driven manufacturing
intelligence (AI) and ML, along with health requirements.
technologies, bringing benefits in the execution of tasks
safety and ergonomics that should be considered as part
and in the analysis of efficiency in real time (Esoso et al.,
of the safety risk assessment process (Anastasi, 2021).
& Tridello et al., 2023; Sanaei; 2020 & Awd, 2022).
There are various information compaction techniques,
Artificial intelligence (AI) is one of the tools used by
also known as custom-made compacted data, that have
researchers to predict the surface quality of materials.
been applied in various areas of research, including
Roughness is a complicated and crucial parameter for
machine learning-based biometrics and statistical analysis
meeting quality standards, and it has posed challenges in
for the development of artificial intelligence. This type of
industries such as automotive, aerospace, energy, and
compacted data design allows to achieve high accuracy in
product manufacturing, where surface quality influences
the data and a reduction in the time in the analysis of the
performance and functionality. While AI helps anticipate
information, adapting to different domains of knowledge.
surface defects in materials, ML is effective in reducing
(Yoon et al., 2023 & Song, 2024).

247
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…

costs and production time, emerging as a promising and associated with SAP, which take into account the overall
novel technique (Batu, 2023; Pereira, 2019). structure of the application and how its components
Laser-directed fabrications in 3D printing, with a central interact. This type of protocol allows the execution of the
focus on neural network architecture (CNN), such as software to be parallelized and tasks to be distributed,
VGG16, AlexNet, GooLeNet, and ResNet, are promising, thus achieving greater efficiency (Falkner et al., 2024;
as time information monitoring systems can establish Kannisto et al., 2021 & López, 2024). In business,
product quality. According to (Patil et al., 2023) the optimization models are used to solve programming
VGG16 architecture can achieve an accuracy of 80% and problems that include boundary conditions, decision
a recovery rate of 89.3%, making it a reliable option for variables, constraints, and target functions. Optimizing
information scheduling has become a crucial topic in
automating the detection and classification of defects such
recent years. Many organizations are looking to improve
as gaps, burr formation, and rough texture in parts with
their facilities to optimize processes and control
additive manufacturing (Zhang et al., 2019).
inventories, aligning with the Industry 4.0 paradigm and
Design for additive manufacturing (DfAM) is one of the optimally managing production systems. Scheduling
technologies linked to industry 4.0, which drives to algorithms and operations scheduling have contributed to
continue innovating in engineering, for the use of metal increasing the efficiency of organizational processes
alloys that offer solutions with impact in different areas (Vaccari et al., 2020 & Togo et al., 2022). The Internet of
such as aeronautics, robotics, automotive, among others, Things (IoT) provides solutions and opportunities for a
and that in turn reduces production costs (Egan, 2023). variety of applications and purposes in areas such as
Arc flash additive manufacturing (WAAM) can produce communications, cyberattack analysis, customer service,
large, complex metal components with a high deposition industrial process management, weather forecasting,
rate. This process can set the right parameters for your traffic flow, inventory control, and targeted marketing.
operation. In addition, the use of models such as SVR, These elements facilitate access to information, which
BPNN, and XGBoost, along with GA and PSO contributes to improving the quality of life of both
algorithms, facilitates the prediction of these parameters individuals and industrial workers in various settings,
(Zhang et al., 2024 & Sharma et al., 2024). especially in the industrial sector (Hossein et al., 2020 &
Ghasemkhani et al., 2022).
2.5. Machine learning and inventory control There are also inventory models for items with non-
instantaneous spoilage that optimize inventory costs,
The integration of machine learning with inventory order times, and maintenance costs in a supply chain.
control (Figure 2) can radically transform inventory There are also mathematical programming methods aimed
management by offering intelligent and adaptable at efficiently managing the inventory system (Mad et al.,
solutions that improve accuracy and efficiency in 2021).
decision-making. While machine learning is a widely Machine learning is becoming a crucial tool in inventory
used tool for solving problems, incorporating it into control, as it offers the ability to forecast product demand
specific use cases with business logic presents numerous with high accuracy and automate various aspects of
challenges. Training, managing, and storing multiple inventory management. This translates into a significant
models require the use of a variety of frameworks and reduction in operating costs, minimization of overstocks
programming languages. and stockouts, and improved overall supply chain
efficiency. By analysing large volumes of historical and
current data, machine learning algorithms can identify
patterns and trends that are difficult to detect manually,
allowing businesses to make more informed and timely
decisions. In addition, the integration of machine learning
systems with inventory management platforms facilitates
the automation of processes such as product
replenishment, stock level optimization, and price
adjustment, resulting in a more agile and adaptive
operation in the face of changes in market demand.

2.6. Machine learning and marketing

Machine learning and marketing have merged to

transform the way businesses interact with their
customers and optimize their business strategies. Machine
Fig. 2. Inventory control elements. learning is a branch of artificial intelligence that allows
machines to learn from data and make predictions or
Tools such as SAP and Flextory allow you to calculate decisions without being explicitly programmed to do so.
orders and play a crucial role in this context. In addition, In marketing, these techniques are used to analyse large
there are architectures and protocols such as RabbitMQ volumes of consumer data, identify patterns and trends,

248
 Journal of Optimization in Industrial Engineering, Vol.17, Issue 2, Summer & Autumn 2024, 243-258

personalize advertising campaigns, improve market where profit opportunities are brief. One of the most
segmentation, and optimize return on investment. This prominent advances in AI is risk management. Advanced
synergy allows companies to anticipate customer needs algorithms can analyze large volumes of data and detect
and behaviours, offering more relevant and effective patterns that might go unnoticed by humans, thereby
experiences. improving risk assessment and mitigation (Chopra, 2021
The fields of artificial intelligence (AI) and machine & El, 2023).
learning (ML) have had a significant impact on the The incorporation of artificial intelligence (AI) and
financial markets thanks to advances in computing and machine learning (ML) into risk management has brought
algorithms. These technologies are transforming various about transformative change in financial markets, with
sectors, especially the financial sector, optimizing applications including credit risk assessment, market risk
business processes and improving risk management (El, management, and operational risk management. One of
2023 & Chui, 2016). the most prominent applications of AI and ML in risk
Artificial intelligence (AI) has revolutionized high- management is credit risk assessment. Traditional models,
frequency trading (HFT) by providing advanced tools and based on predefined parameters and thresholds, often
algorithms capable of analysing large amounts of data in classify people into rigid credit categories, which can lead
real-time and making split-second trading decisions. AI to inaccurate evaluations or inappropriate credit decisions.
enables HFT traders to identify market patterns and trends In contrast, AI and ML offer a more dynamic and
with unprecedented accuracy and speed, improving the adaptable approach (Berg, 2019).
efficiency and effectiveness of their strategies. In Machine learning in marketing has revolutionized the way
addition, the AI's machine learning capability allows these businesses understand, segment, and communicate with
algorithms to continuously adapt and evolve, optimizing their audiences. By leveraging advanced algorithms, they
their trades and reducing the risk of losses. However, the can personalize experiences, predict behaviours, and
use of AI in HFT also poses regulatory and ethical optimize strategies more effectively, resulting in more
challenges, as its speed and complexity can contribute to relevant campaigns and a higher return on investment.
market volatility and raise concerns about fairness and
transparency in financial transactions. There are machine
learning algorithms that use deep learning techniques to 3. Materials and methods
analyse historical market data, considering factors such as
trading volume, differences between supply and demand, For the development of this research, we follow a
and volatility (Arifovic, 2022 & El, 2023). systematic and transparent process to identify research
The complexities of financial markets are often found in focused on the target topic. It also involved consulting
patterns and subtle details that conventional statistical information from scientific articles, repositories and
models might miss. Artificial intelligence, particularly publishers specialized in the field of machine learning and
deep learning models, can identify non-linear its applications in engineering. We searched a variety of
relationships in data, which can be crucial for strategies
sources, both in English and Spanish, covering publishers
such as high-frequency trading (HFT) (Gomber et al.,
such as Ebsco-essentials, ScienceDirect, IEEEXplore, and
2018).
MDPI. Table 1 provides a detailed breakdown of the
On the other hand, AI algorithms can identify market
trends and patterns more quickly than humans, which is search mechanism used to collect data.
particularly advantageous in high-frequency trading,

Table 1
Search process to obtain information.
Editorial
Ebsco-Essentials Science-Direct IEEEXplore MDPI
Search topics
Machine Machine Machine
Machine learning in production Machine Learning in the Machine learning in Learning in Learning and Learning and
systems Supply Chain security Additive inventory control marketing
Manufacturing
Literature Research Languages
English Spanish
Information analysis
Search Logical
Country of publication
Operators Author's name Magazine name Year of publication Keywords
or origin
(Algorithms)

Step 1: Scope of the review. A comprehensive analysis that are contributing to continued advancement in
of the relevant literature was carried out, with the aim of engineering. Following the established guidelines for
exploring existing research on the topic at hand, conducting a critical review, our methodological approach
specifically about reviews on Industry 4.0 and machine was designed to encompass a broad and representative
learning applications. This process focused on identifying selection of articles, as recommended by Paré et al.
both the inherent challenges and innovative applications (2015). In addition, given the dynamic and ever-evolving

249
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…
nature of 4.0 technologies, we limited the scope of the were employed to retrieve relevant documents from the
review to the most recent publications, covering the various relevant publishers, as described in detail in Table
period from 1 January 2015 to 03 June 2024. 2.
Step 2: Search procedures. To ensure comprehensive
coverage of the research topic, specialized algorithms

Table 2
Search algorithms for document extraction.
Ebsco
Search Topics Search Logical Operators (Algorithms) MDPI Science Direct IEEE-Xplore
Essential
AND machine learning and production systems Title AND
machine learning Abstract AND machine learning AllFields AND
Machine learning
production systems AllFields
in production 55 6 36 14
systems

AND machine learning and supply chain Title AND machine

learning Abstract AND machine learning AllFields AND supply
Machine learning chain AllFields
200 12 41 31
in the supply chain

AND machine learning and industrial security Title AND

machine learning Abstract AND machine learning AllFields AND
Machine learning
industrial security AllFields
and industrial 10 13 1 17
security

AND machine learning and additive manufacturing Title AND

Machine learning additive manufacturing Abstract AND machine learning AllFields
and additive AND additive manufacturing AllFields 182 5 109 2
manufacturing

AND machine learning and inventory control Title AND

Machine learning inventory control Abstract AND machine learning AllFields AND
and inventory inventory control AllFields 52 20 10 25
control

AND machine learning and marketing Title AND marketing

Machine learning
Abstract AND machine learning AllFields AND marketing 50 20 10 5
and marketing
AllFields

Fig. 3. PRISMA flowchart used to determine the number of papers used in the literature review.

250
 Journal of Optimization in Industrial Engineering, Vol.17, Issue 2, Summer & Autumn 2024, 243-258

Step 3: Article review and selection process. We chose frequency can be considered as a valid measure of the
to use the PRISMA method (Haddaway, 2020) as a relative importance of words in documents. The word
starting point to carry out the systematic literature review cloud, as a text visualization technique, is especially
process. This method uses a flowchart (see Figure 3) as a useful for visualizing the prominence of "keywords"
model for the organized presentation of the information according to their frequency in the text (Xu, 2016).
examined. For this research, a review of the initial 926 Figure 4 shows the 40 most frequent words that were used
documents was carried out, prioritizing their importance in the search for information in the scientific literature
in relation to the research topic. During this process, 243 applied in this study.
documents were discarded due to duplication or
similarities, leaving 680 documents. Of these, a
comprehensive review was conducted by experts in the
field, taking into account criteria such as the objective of
the research, the target population, the lack of meaningful
application, and its relevance; and 290 documents were
excluded because they did not meet the objectives of the
research, 125 because they did not fit the target
population, 90 because they lacked significant
applications in the area of engineering, and 87 because
they were not relevant in terms of the information
provided. Consequently, only 88 texts were assessed for
suitability and inclusion in the systematic review.
The search was based on the collection of information
Fig. 4. Word cloud associated with machine learning.
from various sources, such as articles, books, and
specialized pages in the fields of research related to
Machine Learning (ML) and Industry 4.0 (see Appendix- After identifying the most common words, a second filter
A). Special attention was paid to publications from the was applied to select those that are most closely related to
last 10 years, prioritizing the most recent. the concept of ML, resulting in 8 keywords. These words
Step 4: Bibliometric analysis process and results. In the are accompanied by information on the number of
bibliometric analysis, the word cloud creation approach documents in which they appear, their percentage of
was applied using the MAXQDA v.2020 software. impact, and the total count of words detected (see Table
Literature search is a common practice in current 3). The most prevalent word is "Machine learning", which
research, as it allows researchers to explore the theoretical is found in 88 documents with an impact of 89% and is
underpinnings related to study variables. In the same way, mentioned 6359 times, while the least frequent is
the words that have a greater association with the topic of "manufacturing", which appears in only 59 documents,
machine learning are shown, along with the number of with an impact of 53% and 2785 mentions in total.
documents in which they appear, the impact and the
number of highlighted words in total. Table 3
For the presentation of the results on the number of Number of articles and citations
publications per year, the statistical software Origin 2022 %
was used, which is a program with a high quality in the Number of
Keywords Documents Documentary
design of statistical graphs and presentation of results (see Words
Impact
Figure 4). Data 58 66 6603
Finally, the number of articles published, the name of the Machine
journal where the documents were published and the 78 89 6359
Learning
corresponding citations by year were analyzed. Citation
Model 67 76 3019
data was obtained through Google Scholar. Of the 88
Used 61 69 2644
articles, the 10 most cited articles were identified.
Process 65 74 2827
Using 61 69 2639
4. Results
4.1. Data review and analysis System 58 66 1448
Manufacturing 59 53 2785
At this point, the 2020 edition of the MAXQDA software
was used to perform a portion of the qualitative In another analysis, data from the 88 articles were
examination of the data. This stage of the analysis focused grouped with data from the years 2022 and 2024, as a
on generating a visual representation of a set of words reference of the most current. The research focused on
from the text, organizing them in a hierarchical structure machine learning as an essential component, its
according to their frequency of use. application in engineering has seen significant growth
The use of the word cloud has made it possible to identify year after year. It is relevant to note that 53% of the total
and discern which words are most closely related to the research belongs to the years 2022 and 2023, suggesting
research topic. According to (Chung et al., 2008), word that machine learning is an innovative trend in the field of

251
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…

engineering and is being widely adopted as part of available in a

manufacturing
Industry 4.0 technologies in various organizations (see environment
Figure 5). Introducing a
novel way of
looking at
quantum
Springer Quantum machine
2017 physics using
Nature learning
machine
learning and
open-source
algorithms
Relevant
aspects such as
cryptographic
approaches,
Internet of things safety and
Computer
2018 security: A top-down security in the
networks
survey field of
engineering
and private
security of the
Fig. 5. Number of publications on machine learning and its IoT
To provide a
different applications in engineering. comprehensive
The year of publication, the name of the journal, the title review of 3D
of the article and the objectives of each of the documents printing
related to the topics associated with machine learning are techniques in
Additive manufacturing
terms of the
presented below (see Table 4). Citation data was obtained Composites (3D printing): A review
main methods
2018 Part B: of materials, methods,
through Google Scholar. Of the 88 articles, the 10 most Engineering applications, and
employed,
cited were identified. materials used,
challenges
their current
status, and
Table 4 applications in
Most cited articles related to machine learning various
Research industries.
Year Journal Article name Data
Objective
acquisition
Spatial prediction Machine and
models for shallow Learning augmentation
landslide hazards: a Optimization models such as
comparative assessment with an
of the efficacy of Susceptibility A Survey on Data improvement
2015 Landslides
support vector Mapping 2019 IEEEXplore Collection for Machine in information
machines, artificial Learning management
neural networks, kernel and
logistic regression, and improvement
logistic model tree of data quality
Trends and through
developments cleaning
that comply processes.
with ASTM Analysis and
and review of the
Additive manufacturing regulations. capability of
2015 Springer A comparison of
technologies Additive additive
Manufacturing traditional
manufacturing
Technology Procedia manufacturing vs
2019 and its current
and manufacturing additive manufacturing,
development
Understanding the best method for the
in establishing
of Concepts job
quality in
Analysis of today's
Industry 4.0 industry.
scenarios Continuous
based on a improvement
Design principles for
2016 IEEEXplore qualitative in additive
industrie 4.0 scenarios
review Defects in additive manufacturing
identifying Progress in manufactured metals as an ideal aid
design 4.0 2020 materials and their effect on in the
principles science fatigue performance: A machining of
Presentation of State-of-the-art review parts with a
Machine learning in
Production & an overview of high
manufacturing:
2016 Manufacturing machine complexity in
advantages, challenges,
Research learning their shape
and applications
techniques

252
 Journal of Optimization in Industrial Engineering, Vol.17, Issue 2, Summer & Autumn 2024, 243-258

By contrast, in 2015 there were 12,050 machine learning- complex model interpretation. However, the combination
related citations in papers on neural networks, decision of ML and additive manufacturing promises a profound
trees, additive manufacturing, and direct digital transformation in the industry, driving innovation and
manufacturing. The year 2018 stood out with 7,190 efficiency in the production of high-quality, customized
citations that addressed topics such as the supply chain, products.
the Internet of Things, blockchain, biomaterials, ceramic The use of machine learning in inventory control is
materials, and metal alloys. Importantly, in the latest revolutionizing supply chain management by providing
publications of 2024, research on artificial intelligence, accurate predictions and automated decisions. Machine
machine learning, and innovative technologies is already learning algorithms analyze large volumes of historical
published, which is expected to have a significant impact sales data, market trends, and demand patterns to
globally, with numerous citations in future research in the anticipate future inventory needs with unprecedented
field of engineering (See Figure 6). accuracy. This allows companies to optimize their stock
levels, reducing both excess and shortage of inventory. In
addition, machine learning techniques can be
continuously adapted and improved, learning from new
data inputs to adjust their predictive models. As a result,
businesses can improve their operational efficiency,
reduce costs, and increase customer satisfaction by
ensuring products are available when and where they are
needed.
Machine learning is transforming marketing by enabling
highly personalized and efficient strategies. By analyzing
large volumes of consumer behavior data, such as
purchase histories, social media interactions, and web
browsing, machine learning algorithms can identify
patterns and segment audiences with great accuracy. This
Fig. 6. Number of publications on machine learning and its makes it easier to create targeted marketing campaigns
different applications in engineering. that resonate more with consumers' individual interests
and needs, increasing conversion rate and customer
5. Discussion loyalty. In addition, machine learning optimizes the
allocation of advertising budgets by predicting which
The discussion on improving production systems in channels and messages will be most effective. Real-time
companies reveals the crucial interdependence between analytics capabilities enable dynamic adjustments to
machine learning (ML) and mechatronics within the marketing strategies, continuously improving their
context of manufacturing systems. As noted by effectiveness.
Constantin et al. (2024), automating manufacturing
processes has significant potential to increase productivity 6. Conclusion
in various industries. The application of algorithms such
as Miceforest, highlighted by Akande (2016) and Li In engineering, machine learning has become a disruptive
(2020), highlights the fundamental role of accuracy and technological innovation that is changing the way
efficiency in optimizing production conditions. In businesses face various challenges in different fields. This
addition, tool condition monitoring (TCM) systems are revolutionary branch of artificial intelligence is based on
essential to ensure the efficiency of machining processes, the ability of computers to process large amounts of data,
as Failing et al. (2023) point out. allowing patterns to be identified and better decisions to
However, it is in the realm of additive manufacturing that be made, rather than relying on explicit programming. In
ML shows its true potential. Schlagenhauf (2022) this context, ML has become an essential tool for
highlights ML's ability to detect anomalies in materials engineers, helping to optimize processes and create
such as 16MnCr5 steel, indicating a significant smarter and more efficient systems.
improvement in error detection and production In areas such as manufacturing, logistics, and supply
optimization. In addition, the interconnectedness between chain, ML algorithms can analyze data in real-time to
technologies such as Blockchain, cybersecurity, AI, and identify inefficiencies, forecast machinery failures, and
ML, highlighted by Mishara (2023) and Ahmad et al. thus intervene effectively so as not to affect product
(2022), demonstrates the importance of a comprehensive quality. This promotes the execution of preventive
digital transformation strategy to ensure data security and maintenance, which translates into significant resource
reliability in the manufacturing framework. savings and increased productivity.
While the implementation of ML and other emerging In the field of mechanical engineering, machine learning
technologies presents promising opportunities to improve is applied in product design and manufacturing, where
efficiency and personalization in manufacturing, it also computer-aided design creates more accurate and resilient
poses significant challenges, such as data quality and

253
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…

products, considering various factors, such as Akande, O., Li, F., & Reiter, J. (2017). An empirical
performance, durability, and cost. comparison of multiple imputation methods of
In addition, machine learning plays a critical role in categorical data. The American Statistician, 71 (2),
software engineering, helping to detect vulnerabilities, 162-170.
contributing to the development of more secure systems, [Link]
maximizing performance, and automating testing and Ali, N., Ghazal, T., Ahmed, A.., Abbas, S., Khan, M.,
debugging processes. Alzoubi, M., Farooq, U., Ahmad, M., & Khan, M.
In short, machine learning has transformed engineering by (2022). Fusion-based supply chain collaboration using
providing machine scientists with powerful tools that machine learning techniques. Tech Science Press, 31,
allow them to free themselves from monotonous tasks, 1-18. [Link]
focusing on those that are most relevant and with more Altmann, M., Benthien, T., Ellendt, N., & Toenjes, A.
insightful solutions. It is crucial to note that artificial (2023). Defect classification for additive
intelligence may not currently address all scenarios in the manufacturing with machine learning. Materials, 16
same way as a human being, but it continues to advance (18), 1-16. [Link]
by leaps and bounds improving the quality of life Alzahrani, A., & Asghar, M. (2023) Intelligent risk
globally. Collaboration between engineers and data prediction system in IoT-based supply chain
scientists is essential to realize the full potential of this management in logistics sector. Electronics, 12 (13),
technology and successfully address future challenges. 1-24. [Link]
Machine learning stands out as a crucial tool in inventory Amin, R., Saeed, N., Masood, M., Moradi, Y., & Slim, M.,
control, as it makes it possible to forecast product demand Al, T. (2021) Deep learning in the industrial internet
with high accuracy and automate key aspects of inventory of things: potentials, challenges, and emerging
management, reducing operating costs, minimizing applications. IEEEXplore, 10, 1-29.
overstocking and shortages, and improving overall supply [Link]
chain efficiency. By analysing large volumes of data, Anastasi, S., Madonna, M., & Monica, L. (2021).
these algorithms identify patterns and trends that are Implications of embedded artificial intelligence-
difficult to detect manually, facilitating more informed machine learning on safety of machinery. Procedia
and timely decisions. The integration of machine learning compute science, 180, 338-343.
systems with inventory management platforms automates [Link]
processes such as product replenishment, stock level Arifovic, J., Zhong, X., & Wei, L. (2022). Machine
optimization, and price adjustment, resulting in a more learning and speed in high frequency trading. Journal
agile and adaptive operation in the face of changes in of economic dynamics and control. 139.
market demand. [Link]
In the financial markets, artificial intelligence (AI) and Awd, M., Münstermann, S., & Walther, F. (2022). Effect of
machine learning have revolutionized areas such as high- microstructural heterogeneity on fatigue strength
frequency trading (HFT) and risk management. Advanced predicted by reinforcement machine learning. Fatigue
AI tools and algorithms make it possible to analyse large & fracture of engineering materials & structure, 1, 1-
amounts of data in real-time and make trading decisions 25. [Link]
in fractions of a second, improving the efficiency and Barzizza, E., Biasetton, N., Ceccato, R., & Salmaso, L.
effectiveness of trading strategies. In addition, these (2023). Big data analytics and machine learning in
technologies make it possible to identify patterns and supply chain 4.0: a literature review. Stats, 6, 1-22.
trends more accurately, optimize operations, and reduce [Link]
the risk of losses. However, the use of AI in HFT also Batu, T., Lemu, H., & Shimels, H. (2023). Application of
poses regulatory and ethical challenges, such as market artificial intelligence for surface roughness prediction
volatility and concerns about fairness and transparency in of additively manufactured components. Materials, 16
financial transactions. (18), 1-34. [Link]
The application of machine learning and artificial Berg, T., Burg, V., Gombović., & Puri, M. (2023). On the
intelligence in marketing and finance has allowed rise of findtechs: Credit scoring using digital
companies and markets to improve their strategies, footprints. The review of financial studies, 33 (7),
optimize processes and offer more personalized and 2845–2897. [Link]
efficient experiences, although it also brings with it Biamonte, J; Wittek, P; Pancotti, N; Rebentrost, P; Wiebe,
important challenges that must be addressed. & N; Lloyd, S. (2017). Quantum machine learning.
Nature, 549, 195-202.
References [Link]
Bridgeball, R. (2023). Unlocking drone potential in the
Ahmad, T., Zhu, H., Zhang, D., Tariq, R., Bassam, A., pharma supply chain: A hybrid machine learning and
Ullah, F., Alghamdi, A., & Alshamrani, S. (2022). GIS approach. Stand, 3, 283-296.
Energetics systems and artificial intelligence: [Link]
Applications of industry 4.0. Energy reports, 8, 334- Carpanzo, E., & Knüttel, D. (2022). Advances in artificial
361. [Link] intelligence methods applications in industrial control
systems: Toward cognitive self-optimizing

254
 Journal of Optimization in Industrial Engineering, Vol.17, Issue 2, Summer & Autumn 2024, 243-258

manufacturing systems. Applied sciences, 12, 1-20. financial markets: A comprehensive analysis of AI
[Link] applications in trading, risks management, and
Chopra, R., & Deep, G. (2021). Application of artificial financial operations. Journal of risk and financial
intelligence in stock market forecasting: A critique, management, 16 (10), 1–16.
review, and research agenda. Journal of risk and [Link]
financial management, 14 (11), 1–34. Engelmann, B., Schmitt, S., Miller, E., Braütigam, V., &
[Link] Schmitt, J. (2020). Advances in machine learning
Choudhary, K., DeCost, B., Chen, C., Jain, A., Tavazza, F., detecting changeover processes in cyber physical
Cohn, R., Park, C., Choudhary, A., Agrawal, A., production systems. Journal of manufacturing and
Billinge, S., Holm, E., Ong, S., & Wolverton, C. materials processing, 4 (4), 1-20.
(2022). Recent advances and applications of deep [Link]
learning methods in material science. Npj Esoso, A., Ikumapayi, M., Jen, T; & Akinlabi, E. (2023).
computational materials, 8, 1-26. Exploring machine learning tools for enhancing
[Link] additive manufacturing: A comparative. International
Chui, M., Manyika, J., & Miremadi, M. (2016). Where information and engineering technology association,
Machines Could Replace Humans—And Where They 28, 535-544. [Link]
Can’t (Yet). McKinsey. Failing, J., Abellán, J., Benavent, S., Rosado, P., &
Chung, H., Pong, R., Fai, K., & Lam, K. (2008). Romero, F. (2023). A tool condition monitoring
Interpreting TF-IDF term weights as making relevance system based on low-cost sensors and an IoT platform
decisions. ACM Transactions on information systems, for rapid deployment. processes, 11 (3), 1-18.
26 (3), 1-37. [Link] [Link]
Constantin, C., Zichil, V., Andrei, V., & Marius, S. (2024). Falkner, D., Bögl, M., Gattinger, A., Stainko, R., Zenisek,
Machine learning mechatronics, and stretch forming: J., & Affenzeller, M. (2024). Integrating machine
A history of innovation in manufacturing engineering. learning into supply chain management: Challenges
Machines, 12 (3), 1-36. and opportunities. Procedia computer sciences. 232,
[Link] 1779-1788.
Constantin, C., Zichil, V., Andrei, V., & Marius, S. (2024). [Link]
Machine learning, mechatronics, and stretch forming: Ghasemkhani, B., Yilmaz, R., Birant, D., & Alp, R. (2022).
A History of innovation in manufacturing engineering. Machine learning models for the prediction of energy
Machines. 12 (3), 1-36. consumption based on cooling and heating loads in
[Link] internet of things based smart buildings. symmetry. 14
Dong, Z., Wu, L., Wang, L.., Li, W., Wang, Z., & Liu, Z. (8), 1-24. [Link]
(2022). Optimization of fracturing parameters with Gibson, I., Rosen, D., & Stucker, B. (2015). Additive
machine-learning and evolutionary algorithm manufacturing technologies- 3D printing, rapid
methods. Energies, 15, 1-23. prototyping, and direct digital manufacturing.
[Link] Springer, 1-510. [Link]
Drakaki, M., Karnavas, Y., Tziaffetas, I., Linardos, V., & 2113-3
Tzionas, P. (2022). Machine learning and deep Gomber, P., Kauffman, R. J., Parker, C., & Weber, B. W.
learning based methods toward industry 4.0 predictive (2018). On the Fintech Revolution: Interpreting the
maintenance in induction motors: A state of art Forces of Innovation, Disruption, and Transformation
survey. Journal of industrial engineering and in Financial Services. Journal of Management
management, 15 (1), 31-57. Information Systems, 35(1), 220–265.
[Link] [Link]
Duong, N., Born, S., Woo, J., Schermeyer, M., Paulick, K., Guo, Z., Wang, H., Kong, X., Shen, L., & Jia, Y. (2021).
Borisyak, M., Cruz, M., Werner, T., Scholz, R., Machine learning-based production prediction model
Schmidt, L., Neubauer, P., & Martínez, E. (2023). and its application in duvernay formation. Energies,
When bioprocess engineering meets machine learning: 14, 1-18. [Link]
A survey from perspective of automated bioprocess Haddaway, N., Page, M., Pritchard, C., & McGuiness, L.
development. Biochemical engineering journal, 190, (2020). An R package and shiny app for producing
1-85. [Link] PRISMA 2020-compliant flow diagrams with
Egan, P. (2023). Design for additive manufacturing: Recent interactivity for optimized digital transparency and
innovations and future directions. Designs, 7, 1-33. open synthesis. Campbell collaboration Wiley, 18, 1-
[Link] 12. [Link]
Ekramirad, N., Doyle, L., Loeb, J., Santra, D., & Adedeji, Heng, S., Tariq, U., Joy, R., Sparks, T., Flood, A., & Liou,
A. (2024). Hyperspectral imaging and machine F. (2024). Experimental, computational, and machine
learning as a nondestructive method for Proso millet learning methods for prediction of residual stresses in
seed detection and classification. Foods. 13 (9), 1-11. laser additive manufacturing: A critical review.
[Link] materials. 17 (7), 1-26.
El, M., & Hammoud, J. (2023). Unveiling the influence of [Link]
artificial intelligence and machine learning on

255
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…

Hermann, M., Pentek, T., & Otto, B. (2015). Design Lei, Y; Qiaoming, H; & Tong, Z. (2023). Research on
principles for industrie 4.0 scenarios. IEEEXplore, 1, supply chain financial risk prevention based on
1-16. [Link] machine learning. Computational intelligence and
Hürkamp, A., Gellrich, S., Ossowski, T., Beuscher, J., neuroscience, 1-15.
Thiede, S., Herrman, & C., Dröder. (2020). [Link]
Combining simulation and machine learning as digital Li, L; Prato, C; & Wang, Y. (2020). Ranking contributors
twin for the manufacturing of overmolded to traffic crashes on mountains freeways from an
thermoplastic composites. Journal of manufacturing incomplete dataset: A sequential approach of
and materials processing, 4 (3), 1-21. multivariate imputation by chained equations and
[Link] random forest classifier. Accident analysis &
Jamwal, A., Agrawal, R., Sharma, M., & Giallanza, A. prevention, 146, 1-11.
(2021). Industry 4.0 technologies for manufacturing [Link]
sustainability: a systematic review and future research Liebsch, A., Koshulow, W., Gabauer, J., Kupfer, R., &
directions. Applied sciences, 11, 1-27. Gude, M. (2019). Overmoulding of consolidated fibre-
[Link] reinforced thermoplastics-increasing the bonding
Kang, K., Chen, Q., Wang, K., Zhang, Y., Zhang, D., strength by physical surface pre-treatments. Procedia
Zheng, G., Xing, J., Long, T., Ren, X., Shang, C., & CIRP, 85, 212-217.
Cui, B. (2023). Application of interpretable machine [Link]
learning for production feasibility prediction of gold López, R., Panizo, L., & Del Mar, M. (2024). Flextory:
mine project. Applied sciences, 13, 1-17. Flexible software factory of IoT data consumers.
[Link] Sensors. 24 (8), 1-25.
Kannisto, P., Hästbacka, D., Gutiérrez, T., Suominen, O., [Link]
Vilkko, M, & Craamer, P. (2021). Plant-wide Lu, W., & Yan, X. (2021). Variable-weighted FDA
interoperability and decoupled, data-driven process combined with t-SNE and multiple extreme learning
control with message bus communication. Journal of machines for visual industrial process monitoring, ISA
industrial information integration. 26, 1-15. transactions 163-171.
[Link] [Link]
Khchine, R., Amar, A., Guennoun, Z., Bensouda, C., & Mad, M., Mohd, M., Bin, F., & Bin, E. (2021). An
Laaroussi, Y. (2018). Machine learning for supply unsupervised machine learning-based framework for
chain´s big data: State of the art and application to transferring local factories into supply chain networks.
social networks´ data. Matec web conf, 200, 1-7. mathematics. 9 (23), 1-31.
[Link] [Link]
Khorasani, A., Gibson, I., Veetil, J., & Ghasemi, A. (2020). Mazzei, D., & Ramjattan, R. (2022). Machine learning for
A review of technological improvements in laser- industry 4.0: A systematic review using deep learning-
based powder bed fusion of metal printers. The based topic modelling. Sensors, 22, 1-26.
international journal of advanced manufacturing [Link]
technology, 1-19. [Link] Mishara, S. (2022). Blockchain and machine learning-
05361-3 based hybrid IDS to protect smart networks and
Kim, M., Yun, J., Cho, Y., Shin, K., Jang, R., Bae, H., & preserve privacy. electronics, 12 (16), 1-23.
Kim, N. (2019). Deep learning in medical imaging. [Link]
Neurospine, 16, 657-668. Mo, H., Morin, D., Magne, K., He, J., & Zhang, Z. (2024).
[Link] Transferability of temperature evolution of dissimilar
Kim, S., Seo, J., & Kim, S. (2024). Machine learning wire-arc additively manufactured components by
technologies in the supply chain management research machine learning. Materials. 17 (3), 1-28.
of biodiesel: A review. Energies. 17 (6), 1-28. [Link]
[Link] Mohseni, S., Wang, H., Xiao, C., Yu, Z., Wang, Z., &
Kouicem, D., Bouabdallah, A., & Lakhlef, H. (2018). Yadawa, J. (2022). Taxonomy of machine learning
Internet of things security: A top-down survey. safety: A survey and primer. ACM Digital Library,
Computer networks, 1-24. 55, 1-38. [Link]
[Link] Ngo, T., Kashani, A., Imbalzano, G., Nguyen, K., & Hui,
Kumar, R. (2023). Improving quality of predictive D. (2018). Additive manufacturing (3D printing): A
maintenance through machine learning algorithms review of materials, methods, applications and
industry 4.0 environment. Proceedings on engineering challenges. Composites part B: Engineering, 143,
sciences, 5 (1), 1-10. 172-196.
[Link] [Link]
Lee, C., & Lim, C. (2021). From technological Otoum, Y., Liu, D., & Nayak, A. (2021). DL-IDS: A Deep
development to social advance: A review of industry learning-based intrusion detection framework for
4.0 through machine learning. Technological securing IoT. Trans. Emerg. Emerging
forecasting and social change, 167, 1-12. telecommunications technologies, 33 (3), 1-17.
[Link] [Link]

256
 Journal of Optimization in Industrial Engineering, Vol.17, Issue 2, Summer & Autumn 2024, 243-258

Paré, G., Trudel, M., Jaana, M., & Kitsiou, S. (2015). Measurement: sensors, 28, 1-19.
Synthesizing information systems knowledge: A [Link]
typology of literature reviews. Information & Schlagenhauf, T., & Burghard, N. (2021). Intelligent vision
Management, 52 (2), 183-199. based wear forecasting on surfaces of machine tool
[Link] elements. Sn. Applied sciences, 3 (858), 1-13.
Patil, D., Nigam, A., Mohapatra, S., & Nikam, S. (2023). A [Link]
deep learning approach to classify and detect defects Schlagenhauf, T., Jan, W., & Puchta, A. (2022).
in the components manufactured by laser directed Convolutional-based encoder-decoder network for
energy deposition process. Machines, 11, 1-19. time series anomaly detection during the milling of
[Link] 16MnCr15. data, 7 (12), 1-8.
Pelzer, L., Schulze, T., Buschmann, D., Enslin, C., Schmitt, [Link]
R., & Hopmann, C. (2023). Acquiring process Schroeder, M., & Lodeman, S. (2021). A systematic
knowledge in extrusion-based additive manufacturing investigation of the integration of machine learning
via interpretable machine learning. Polymers, 15, 1- into supply chain risk management. Logistics, 5, 1-16.
18. [Link] [Link]
Pereira, I., Madureira, A., Bettencourt, N., Coelho, D., Sharma, R., Raj, A., Mukherjee, M., Krishna, S., Kumar,
Ângelo, M., Araújo, C., & Alves, D. (2024). A R., & Kumar, A. (2024). Forecasting of process
machine learning as a servisse (MLaaS) Approach to parameters using machine learning techniques for wire
Improve marketing success. Informatics. 11 (2), 1-34. arc additive manufacturing process. Materialstoday:
[Link] proceedings. 80 (1), 248-253.
Pereira, T., Kennedy, J., & Potgieter, J. (2019). A [Link]
comparison of traditional manufacturing vs additive Smith, D., & Simpson, K. (2020). The safety critical
manufacturing, the best method for the job. Procedia systems handbook. Elsevier publishing, UK.
manufacturing, 30, 11-18. [Link]
[Link] Song, K. (2024). Compact data learning for machine
Rathnayake, R., Maduranga, M., Tilwari, V., & learning classifications. axions. 13 (3), 1-16.
Dissanayake, M. (2023). RSSI and machine learning- [Link]
based indoor localization systems for smart cities. Tancredi, G., Vignali, G., & Bottani, E. (2022). Integration
Eng, 4 (2), 1468-1494. of digital twin, machine-learning and industry 4.0
[Link] tools for anomaly detection: An application to a food
Roh, Y., Heo, G., Whang, S., & Whang, S. (2019). A plant. Sensors, 22, 1-23.
survey on data collection for machine learning: A big [Link]
data – AI integration perspective. IEEEXplore, 33, Tariq, N., Asim, M., Obeidat, F., Farooqi, M., Baker, T.,
1328-1347. Hammoudeh, M., & Ghafir, I. (2019). The security of
[Link] big data in fog-enable IoT applications including
Rojek, I., Jasiulewicz, M., Piechowski, M., & blockchain: A survey. Sensors, 19, 1-33.
Mikolajewski, D. (2023). An artificial intelligence [Link]
approach for improving maintenance to supervise Tien, D., Anh, T., Klempe, H., Pradhan, B., & Revhaug, I.
machine failures and support their repair. Applied (2015). Spatial prediction models for shallow
sciences, 33 (4), 1-17. landslide hazards: a comparative assessment of the
[Link] efficacy of support vector machines, artificial neural
Romero, J., Macgluf, A., Rodríguez, A., Espinoza, J., & networks, kernel logistic regression, and logistic
Suárez, A. (2021). Aplicación de Machine learning en model tree. Landslides, 13, 361-378.
la industria 4.0 en tiempos de pandemia. [Link]
Intercontectando saberes, 11 (6), 1-7. Togo, H., Asanuma, K., Nishi, T., & Liu, Z. (2022).
[Link] Machine learning and inverse optimization for
Sanaei, N., & Fatemi, A. (2020). Defects in additive estimation of weighting factors in multi-objective
manufactured metals and their effect on fatigue production scheduling problems. Applied. Sciences,
performance: A State of the art review. Progress in 12, 1-18. [Link]
materials science, Togo, H., Asanuma, K., Nishi, T., & Liu, Z. (2022).
[Link] Machine learning and inverse optimization for
Sani, S., Xia, H., Syed, J., & Salonitis, K. (2023). Supply estimation of weighting factors in multi-objective
chain 4.0: A machine learning-based Bayesian- production scheduling problems. Applied sciences. 12
optimized lightGBM model for predicting supply (19), 1-18. [Link]
chain risk. Machines , 11 (9), 1-20. Tridello, A., Ciampaglia, A., Berto, F., & Paolino, D.
[Link] (2023). Assessment of the critical defect in additive
Sazzadur, M., Tapotosh, G., Ferdous, N., Shamim, M., manufacturing components through machine learning
Anannya, M., & Sanwar, A. (2023). Machine learning algorithms. Applied sciences, 12 (19), 1-22.
and internet of things in industry 4.0: A review. [Link]

257
 Marco Antonio Diaz Martinez & et al. / Industry 4.0- Applications of machine learning in the field of…

Vaccari, M., di Capaci, R.B., Brunazzi, E., Tognotti, L., challenges, and applications. Production and
Pierno, P., Vagheggi, R., & Pannocchia, G. (2020). manufacturing research, 4, 1-24.
Implementation of an Industry 4.0 system to optimally [Link]
manage chemical plant operation. IFAC- Xia, Y., Xu, T., Wei, M., Wei, Z., & Tang, L. (2023).
PapersOnLine. 53 (2), 11545-11550. Predicting chain´s manufacturing SME credit risk in
[Link] supply chain finance based on machine learning
Velásquez, D., Pérez, E., Oregui, X., Artetxe, A., Manteca, methods. Sustainability, 15 (2), 1-18.
J., Mansilla, J., Toro, M., Maiza, M., & Sierra, B. [Link]
(2022). A Hybrid machine-learning ensemble for Xu, A., Darbandi, M., Javaheri, D., Navimipour, N., &
anomaly detection in real-time industry 4.0 systems. Yalcin, S. (2023). The management of IoT- based
IEEEXplore, 10, 72024-72036. organizational and industrial digitalization using
[Link] machine learning methods. Sustainability, 15 (7), 1-
Wei, Z., Liu, H., Tao, X., Pan, K., Huang, R., Ji, W., & 28. [Link]
Wang, J. (2023). Insights into the application of Xu, J., Tao, Y., & Lin, H. (2016). Semantic word cloud
machine learning in industrial risk assessment: A Generation based on word embeddings. IEEE Pacific
bibliometric mapping analysis. Sustainability, 15 (8), visualization symposium (PacificVis), 239-243.
1-30. [Link] [Link]
Wiemer, H., Dementyev, A., & Ihlenfeldt, S. (2021). A Yoon, S., Cantwell, W.J., Yeun, C., Cho, C., Byon, Y.,
holistic quality assurance approach for machine Kim, T. (2023). Defect detection in composites by
learning applications in cyber-physical production deep learning using highly nonlinear solitary waves.
systems. Applied sciences, 11 (20), 1-20. Int. J. Mech. Sci.
[Link] [Link]
Wong, S., Woon, J., Yip, Y., & Kawasaki, T. (2023). A Zhang, H., Bai, X., Dong, H., & Zhang, H. (2024).
case study of how Maersk adopts cloud-based Modelling and prediction of process parameters with
blockchain integrated with machine learning for low energy consumption in wire arc additive
sustainable practices. Sustainability, 15, 1-18. manufacturing based on machine learning. metals.
[Link] [Link]
Wong, S., Yeung, J., Lau, Y., & So, J. (2021). Technical Zhang, Y., Soon, H., Ye, D., Fuh, J., & Zhu, K. (2019).
sustainability of cloud-based blockchain integrated Powder-bed fusion process monitoring by machine
with machine learning for supply chain management. vision with hybrid convolutional neural networks.
Sustainability, 13 (15), 1-21. IEEEXplore, 16 (9), 5769-5779.
[Link] [Link]
Wuest, T., Weimer, D., Irgens, C., & Dieter, K. (2016).
Machine learning in manufacturing: advantages,

258