Journal of Intelligent Manufacturing (2022) 33:425–439

https://doi.org/10.1007/s10845-021-01885-x

Trends in intelligent manufacturing research: a keyword co-occurrence

network based review
Chenxi Yuan1 · Guoyan Li1 · Sagar Kamarthi1 · Xiaoning Jin1 · Mohsen Moghaddam1

Received: 5 March 2021 / Accepted: 21 November 2021 / Published online: 5 January 2022
© The Author(s) 2022

Abstract
In recent years, driven by Industry 4.0 wave, academic research has focused on the science, engineering, and enabling
technologies for intelligent and cyber manufacturing. Using a network science and data mining-based Keyword Co-occurrence
Network (KCN) methodology, this work analyzes the trends in data science topics in the manufacturing literature over the
past two decades to inform the researchers, educators, industry leaders of knowledge trends in intelligent manufacturing. It
studies the evolution of research topics and methods in data science, Internet of Things (IoT), cloud computing, and cyber
manufacturing. The KCN methodology is applied to systematically analyze the keywords collected from 84,041 papers
published in top-tier manufacturing journals between 2000 and 2020. It is not practically feasible to review this large body
of literature through tradition manual approaches like systematic review and scoping review to discover insights. The results
of network modeling and data analysis reveal important knowledge components and structure of the intelligent and cyber
manufacturing literature, implicit the research interests switch and provide the insights for industry development. This paper
maps the high frequency keywords in the recent literature to nine pillars of Industry 4.0 to help manufacturing community
identify research and education directions for emerging technologies in intelligent manufacturing.

Keywords Keyword co-occurrence network · Industry 4.0 · Intelligent manufacturing · Data science · Machine learning ·
Artificial intelligence

Introduction teria such as efficiency, throughput, and cost to emerging

criteria like adaptability, resilience, autonomy, interoper-
The manufacturing industry has always been significantly ability, and cyber-security. With the growing application of
influenced by technological revolutions—from the invention intelligent and digital technologies, factories of the future
of steam engines and electricity to the rise of robotics and are shaping into networks of cyber-physical machines with
automation, the Internet of things (IoT), and cyber-physical embedded sensing, computing, and communication capabil-
systems. The major drivers for integrating these technologies ities, enabling adaptive and intelligent automation (Kusiak,
into manufacturing range from traditional performance cri- 2019; Moghaddam & Nof, 2017). The scientific and tech-
nological advancements (Moghaddam et al., 2018) have
B Mohsen Moghaddam enabled emerging paradigms, such as smart manufacturing
[email protected] (Kusiak, 2018; Lu et al., 2016), cyber-physical production
Chenxi Yuan systems (Monostori et al., 2016), Industry 4.0 (Oztemel
[email protected] & Gursev, 2020), and cloud-enabled manufacturing (Chen,
Guoyan Li 2017). The main objective of this paper is to identified the
[email protected] knowledge components, knowledge structure, and research
Sagar Kamarthi trends using KCN approach (Duvvuru & Kamarthi, 2012;
[email protected] Radhakrishnan et al., 2017), which would allow us to explore
Xiaoning Jin and analyze a vast amount of literature covering 84,041
[email protected] papers published in top-tier manufacturing journals between
2000 and 2020. It is not practically feasible to review this
1 Department of Mechanical and Industrial Engineering, many papers through a traditional systematic review process.
Northeastern University, Boston, MA 02115, USA

123
426 Journal of Intelligent Manufacturing (2022) 33:425–439

Motivation: technology trends in manufacturing

The fourth industrial revolution (or Industry 4.0) is driven by

the emergence of new technologies such as cyber-physical
systems and Internet of Things (IoT) that focus heavily on
interconnectivity, automation, machine learning, and real-
time analytics. As a result, several paradigms have emerged
in recent years to characterize the architectures and require-
ments for next-generation manufacturing systems:

• Cyber-physical manufacturing lies at the intersection of

IoT, cyber-physical systems, and manufacturing science
and technologies (Fig. 1), and is generally characterized
as “physical and engineered systems whose operations are
monitored, controlled, coordinated, and integrated by a
computing and communicating core” (Rajkumar et al.,
2010). This paradigm advocates capabilities for collecting Fig. 1 The convergence of cyber and physical technologies in future
data accurately and efficiently, analyzing large amounts of manufacturing (adapted from (Monostori et al., 2016))
data in real-time, making decisions based on real-time data
in an instant and distributed manner, and supporting self-
ufacturing services via networks (e.g., the Internet) and
configuration and self-adaption (Majstorović et al., 2018).
cloud manufacturing service platforms” (Xu., 2012).
• Industry 4.0 characterizes a new era of industrial revo-
lutions with emphasis on interconnectivity, automation,
machine learning, and real-time data analytics (Rec- Despite the differences between the objectives of the intel-
ommendations for implementing the strategic initiative ligent manufacturing facets mentioned above, their common
INDUSTRIE 4.0 April, Securing the future of German vernacular is to enable (Moghaddam et al., 2018): (1) inte-
manufacturing industry Final report of the Industrie 4.0 grated and collaborative manufacturing systems and value
Working Group 2013), with the vision of enabling the networks connected via IoT, (2) digitalization and integra-
notion of lot-size of one by addressing current challenges tion of manufacturing resources on the cloud as secure
of shortened product lifecycles, growing demand for per- and on-demand services, and (3) connected, intelligent, and
sonalized products, and heightened global competition autonomous cyber-physical machines enabled by cloud/edge
(Xu, 2012). It is characterized by horizontal integration computing and machine learning technologies (Tian et al.,
through supply and value chains, vertical integration of the 2002). While different segments or domains of the man-
automation hierarchy and industrial control systems, and ufacturing industry may have different needs, they are all
end-to-end integration of product and manufacturing engi- expected to leverage connectedness and access to real-time
neering throughout their entire value stream (Hofmann & insights across systems, processes, partners, products, and
Rüsch, 2017; Järvenpää et al., 2019). people. The goal is to enable mass-customization of prod-
• Smart manufacturing. This paradigm advocates use ucts and services; adopt automatic and flexible production
of information and communication technologies and lines (Escobar et al., 2021); track parts and products in real-
advanced data analytics to deal with the uncertain and time; facilitate communication among parts, products, and
dynamic behaviors of demand, factory conditions, and machines; enhance human-machine collaboration(Nguyen
supply networks (Thoben et al., 2017). NIST defines Ngoc et al., 2021); achieve IoT-enabled production opti-
smart manufacturing as “fully-integrated, collaborative mization in smart factories (Yao et al., 2019); create new
manufacturing systems that respond in real-time to meet types of services and business models of interaction in value
changing demands and conditions in the factory, in the chains (Alcácer & Cruz-Machado, 2019); achieve a higher
supply network, and in customer needs” (Lu et al., 2016). level of intelligent automation; and ultimately increase qual-
• Cloud manufacturing is built upon the technological ity, productivity, flexibility, lower costs, and higher efficiency
advances in hardware, virtualization technology, dis- (Barari et al., 2021).
tributed computing, and service delivery over the Internet Figure 1 illustrates the convergence of various cyber-
(Ooi et al., 2018). it is “a new networked manufactur- physical manufacturing technologies that collectively aim at
ing paradigm that organizes manufacturing resources over materializing the vision of future manufacturing described
networks (manufacturing clouds) according to consumers’ above. This convergence, in turn, is leading to the develop-
needs and demand to provide a variety of on-demand man- ment of new knowledge and methods at the intersection of

123
Journal of Intelligent Manufacturing (2022) 33:425–439 427

the physical world technologies (e.g., robotics, automation, Section presents the results and analyses of the experiments.
product-service systems, cloud manufacturing) and the cyber Discussion Section discusses the implications of the results
world technologies (e.g., AI, machine learning, multi-agent for identifying key knowledge trends and workforce skills
systems, cloud computing). As the emerging cyber-physical requirements in future manufacturing systems. Conclusions
manufacturing technologies grow more complex and more Section concludes the article and summarizes the findings
knowledge-intensive, the workforce is expected to be adept and directions for future research.
at handling these technologies. For example, big data analyt-
ics and management (e.g., data mining, data classification,
data storage) are becoming a critical challenge for manu- Background
facturers (Choudhary et al., 2009) (Yao et al., 2019). New
cloud architectures and services are needed for analyzing data This section presents a review of existing studies that inves-
under certain security and privacy protocols(Alam & Saddik, tigate various aspects of technologies in cyber-physical
2017). New machine learning algorithms and cloud services manufacturing in a systematic fashion. The current literature
are required for optimizing manufacturing processes and sys- reviews are grouped into two categories: (1) concepts and
tems (Usuga Cadavid et al., 2020). The rise of deep learning technology trends, and (2) applications and future trends.
poses new opportunities and challenges for manufacturers to Concepts and Technology Trends The articles in this cat-
decode the increasingly complex manufacturing processes egory investigate the key technologies for cyber-physical
and large-scale production systems to be competitive in manufacturing from diverse perspectives. DeFelice and
their rapidly changing industry (Miškuf & Zolotová, 2016; Petrillo (2018) conduct a literature review on smart manu-
Oztemel & Gursev, 2020). In this context, both researchers facturing using a multicriteria decision model. They identify
and practitioners need to identify emerging research trends, cyber-physical systems, IoT, and big data as the most fre-
technologies, and workforce knowledge requirements. This quently used terms in smart manufacturing based on a large
is the focal issue investigated in this paper. number of publications in years 2014–2016. Alcacer and
Cruz-Machado (2019) present a literature review of new
Objectives technologies for manufacturing systems and conclude that
Industry 4.0 has accelerated the progress of manufactur-
This paper is motivated by an urgent need for observing the ing digitalization. Henzel and Herzwurm (2018) identify
research trends in rapidly changing interelligent manufac- cloud manufacturing as an emerging technology by deliv-
turing landscape(Radhakrishnan et al., 2017). We conducted ering a state-of-the-art survey of current issues on cloud
a systematic review of 84,041 recent articles published in manufacturing. Franco and Ganga (2020) explore additive
leading journals ranked Q1 by the Scimago Journal & Coun- manufacturing as an emerging technology for manufactur-
try Rank (SJR). A Keyword Co-occurrence Network (KCN) ing, describe the framework for additive manufacturing, and
methodology (Duvvuru & Kamarthi, 2012; Radhakrishnan explore the effects of adopting additive manufacturing. Big
et al., 2017) to identify knowledge components, knowl- data analytics is essential for generating the digital twins
edge structure, and research trends associated with emerging of manufacturing equipment, and subsequently, enabling
cyber-physical manufacturing technology areas including the advanced predictive capacity to identify events that can affect
items listed in Fig. 1 and beyond. A KCN is created by treat- production before they occur (Schuh et al., 2017). Shrouf and
ing the keywords of the articles as individual nodes and each Ordiers (Shrouf et al., 2014) define the main characteristics
co-occurrence of a pair of keywords is modeled as a link of smart factories with the focus on sustainability. They antic-
between their respective nodes. The co-occurrence frequency ipate IoT to play a major role in manufacturing.
of each pair of keywords is represented as the weight of the Cyber-physical manufacturing requires advanced analyt-
link connecting the pair. The proposed KCN methodology is ical capabilities for manufacturers to turn data into insights
implemented on the keywords listed in the 84,041 articles to and decisions in real-time to achieve the required level of
obtain key topological parameters and perform a statistical adaptability for efficient and personalized production, where
and visual analysis of the network to reveal potential pat- big data is targeted as an essential technology (Mourtzis et al.,
terns and evolution of keywords, providing insight into the 2016).
emerging knowledge trends in the manufacturing sector. Applications and Future Trends Several researchers have
The remainder of this article is organized as follows. explored applications of key technologies and potentials
Background Section discusses the background and related of characterizing next-generation cyber-physical manufac-
work on the identification of emerging knowledge trends and turing systems. Cui and Cara (2020) present a systematic
skills requirements in the manufacturing industry. The KCN literature review on big data analytics in manufacturing and
Methodology Section introduces the KCN methodology and identify six key drivers: system integration, data, predic-
elaborates on the design of experiments. Results andAnalyses tion, sustainability, resource sharing, and hardware. They

123
428 Journal of Intelligent Manufacturing (2022) 33:425–439

identify five potential directions for big data research in cation by analyzing links between citations in the literature
manufacturing: modeling and simulation, connectivity and during a short period (Onel et al., 2011) (Shibata et al., 2011).
interoperability, standardized big data platform design, real- This method is not suitable for the purpose of current work,
time big data analytics, and cybersecurity. Baroroh and because it focuses on the relative impact of published sci-
Wang (2020) conduct a systematic review of recent appli- entific work rather than on the emerging technology trends.
cations of augmented reality in smart manufacturing from The other frequently used method is keyword network anal-
the human-machine interaction perspective and discuss how ysis (KCN) (Su & Lee, 2010), which explores the links
augmented reality works as an interface between human and between keywords in the literature to reveal the knowledge
artificial intelligence to assist manual operations in manu- components and structure of a scientific field. Keywords can
facturing. Xu and He (2014) conduct a survey of current provide a concise overview of the important content and
research of the Internet of things, which refers to a system key points of a body of articles as an essential textual ele-
of integrated devices (e.g., machines, equipment, sensors, ment (Li et al., 2016). The KCN methodology captures the
terminals) equipped with sensing, identification, processing, connections between different concepts at a micro-level and
communication, and networking capabilities. IoT is believed provides insights about respective roles and importance. Fur-
to be developed in inventory and production management ther, the KCN methodology can deliver network attributes
and transportation. Vaidya and Ambad (2018) present an on articles published over a long period and hence reveal
overview of Industry 4.0 and the state of current manu- the evolution of the topics over time (Zhong et al., 2014).
facturing systems. They impose the issues and challenges Therefore, this paper applies the concept of KCN to analyze
of cyber-physical manufacturing including cybersecurity, the linkage of concepts and investigate the research topics,
modularized and flexible physical artifacts, manufacturing their relationships, and development trends in manufactur-
systems ensuring high quality and data integrity. Com- ing science and engineering. The rationale behind the KCN
prehensive systematic reviews have also been conducted methodology is that the keywords of top-tier, peer-reviewed
to identify opportunities for research in artificial intelli- journal articles represent the most important areas of research
gence applications in manufacturing (Aggour et al., 2019; and development in their respective area of study and that
Kutschenreiter-Praszkiewicz, 2008; Nti et al., 2021). their co-occurrence represents their relationships and relative
This study addresses the pressing need for an exhaustive importance. This technique can analyze the content of a large
review of emerging manufacturing technologies and their number of papers by quantifying the associations between
impact on the manufacturing workforce in terms of the new keywords, reflecting on the respective role and evolution of
domain knowledge and skills required. We speculate that the keywords, and identify the overall structure of the research
scientific research publications are representative of tech- field (Lozano et al., 2019).
nology and skills requirements trends in the broader area of
manufacturing science and engineering. Thus, this paper con- Data collection and processing procedure
ducts a comprehensive analysis of keywords of Q1 articles in
JSR’s category of Manufacturing and Industrial Engineering To build the KCN for manufacturing science and engineer-
to reach this end. ing, we collected the keywords of top-tier journals published
in the period betwen2000 and 2019 in 43 journals that are
ranked as SJR Q1 (Table 1). All papers published in each
The KCN methodology journal were accessed through the Web of Science database
and the author-specified keywords from all these articles
This section provides a detailed description of the Keywork were collected for further processing. A total of 84,041 arti-
Co-occurrence Network (KCN) methodology, including the cles were identified, which collectively comprise a total of
underlying rationale for conducting this type of analysis, data 258,882 standard keywords defined by their authors (Table
collection and processing procedure, and network parameters 2). The number of publications increased from 1638 papers in
for extracting insights on emerging technologies and skills 2,000 to 7,887 papers in 2019. These keywords were recon-
requirements. ciled to eliminate redundancy, by unifying singular and plural
variants (e.g., “algorithms” and “algorithm”), hyphened and
Rationale non-hyphenated phrases (e.g., “multi agent system” and
“multi-agent system”), synonyms (e.g., “statistical methods”
Two main network-based methods have been widely used and “statistical modeling”), and acronym variants (e.g., “arti-
for the review of scientific and technical publications (Li ficial neural networks (ANN)”, “artificial neural networks”,
et al., 2016). The first is the citation network, which identifies and “ANN”). Next, keywords relevant to the cyber-world
important academic articles according to citation frequency, and physical-world manufacturing science, engineering, and
and focuses on studying the structure of scientific communi- technology were selected by three authors knowledgeable

123
Journal of Intelligent Manufacturing (2022) 33:425–439 429

Fig. 2 The procedure of data

collection, processing and
cleaning

about these areas. The selection of keywords by three authors to other nodes in a network. The degree of node i represents
was compiled into the final keyword list with 178,111 key- its relative importance in the network, as follows:
words in total. Through this process, the keywords that 
had occurred with high frequency but were irrelevant to ki  ei j , (1)
cyber-physical manufacturing technologies were eliminated. j∈N
Co-occurrence of all pairs of keywords was then calculated
by counting the co-appearance of word-pairs in all the papers. where N denotes the set of all keywords in the keyword
The procedure of data collection, processing and cleaning are network, and ei j ∈ {0, 1} denotes the existence of an edge
described in Fig. 2. between nodes i and j (i, j ∈ N ). In the KCN methodology,
The papers published between 2000 and 2019 were par- the higher the degree of a node, the greater the centrality of
titioned into four 5-year windows: 2000–2004, 2005–2009, the node in the keyword network. Centrality of a keyword is
2010–2014, and 2015–2019. A separate KCN is constructed an indicator of its importance.
for each window to explore the temporal evolution of the The keyword network is a weighted network. The weight
trends in manufacturing science and engineering research. wi j of an edge ei j represents the count of nodes i and j co-
The attributes of the KCNs are then imported into the Net- occurring. The strength of node i is the sum of the weights of
work Workbench software (NWB Team, 2006) to retrieve all edges connected to node i. It is, like centrality, an indicator
various network parameters to explore co-occurrence pat- of the relative importance and connectivity of a keyword in
terns among keywords, and then generate insights on the the network. The strength of node i is calculated as follow:
potential knowledge structure of the emerging trends in man- 
ufacturing. si  ei j wi j . (2)
j∈N

Another key parameter for KCN analysis is the aver-

age weight of end point degrees, which measures the co-
Network analysis parameters occurrence of the edges between pairs of nodes as the degrees
of the nodes change. It is defined as follows:
This section describes the network analysis parameters (Rad-  
hakrishnan et al., 2017) used in the KCN methodology for wi j ∼ ki k j , (3)
analyzing the network connection weights and the relation-
ship between different nodes. The first parameter is degree, where ki and k j are degrees of nodes i and j, respectively, and
which is the total number of links that connect a given node · denotes average value. In the context of this research, the

123
430 Journal of Intelligent Manufacturing (2022) 33:425–439

Table 1 List of the selected 43 journals positioned top in the SJR ranking Table 2 The total number of papers and keywords each year from 2000
to 2019
Journal of operations CAD Computer Aided Design
management Year # of paper # of Year # of paper # of
International Journal of Food and Bioprocess keywords keywords
Machine Tools & Technology
Manufacture 2019 7887 23,704 2009 4028 12,592

Production and Operations Journal of Manufacturing 2018 7552 22,781 2008 3781 11,549
Management Processes 2017 7082 21,687 2007 3436 10,734
Additive Manufacturing Industrial Management and 2016 5859 18,288 2006 3124 9349
Data Systems 2015 5793 18,048 2005 2537 7877
International Journal of CIRP Journal of Manufacturing 2014 5219 16,512 2004 2228 6753
Production Economics Science & Technology
2013 5203 16,116 2003 2469 7377
CIRP Annals—Manufacturing Materials and Manufacturing
2012 4346 13,672 2002 2379 7138
Technology Processes
2011 3742 11,936 2001 2274 6952
Virtual and Physical IEEE Transactions on Industry
Prototyping Applications 2010 3464 10,803 2000 1638 5014
Journal of Quality Technology Rapid Prototyping Journal
Reliability Engineering and International Journal of
System Safety Bioprinting
where the exponent θ is a positive constant. Positive correla-
Manufacturing Letters International Journal of
Advanced Mfg. Technology tion confirms that highly connected keywords are more likely
Journal of Materials Processing Journal of Process Control
to have stronger connections.
Technology The next network parameter examined in this study is the
Advanced Materials Journal of Manufacturing average weighted nearest neighbor’s degree of node i, for
Technologies Technology Management all i ∈ N . It measures the tendency of a node to link with
Journal of Manufacturing Sustainable Production and its neighbors with similar degree characteristics: high degree
Systems Consumption nodes connecting with high degree neighbors and low degree
International Journal of Machining Science and nodes connecting with low degree neighbors. This parameter
Production Research Technology can be calculated as follows:
Journal of Industrial 3D Printing and Additive
Information Integration Manufacturing
Sustainable Materials and Advances in Production 1 
Technologies Engineering and Management kin  ei j wi j k j . (5)
si
j∈N
Production Planning and Journal of Computing and
Control Information Science in Eng
Robotics and International Journal of
Computer-Integrated Precision Engineering and If this value increases with the keyword degrees, it
Manufacturing Mfg represents the following assortative behavior: high-degree
Journal of Intelligent International Journal of keywords are associated with other high-degree keywords
Manufacturing Industrial Eng. Computations in the papers. Inversely, the disassortative behavior indicates
IISE Transactions Advances in Natural Sciences that high-degree keywords co-occur with low-degree key-
Flexible Services and Journal of Intelligent and words.
Manufacturing Journal Robotic Systems The final parameter is the weighted clustering coefficient,
Operations Management Advances in Manufacturing which measures the local cohesiveness of node i among its
Research neighbors. It represents the importance of the structure clus-
tered around a certain keyword on the basis of the interaction
intensity found on the local triplets (Barrat et al., 2004). This
parameter is formulated as follows:
average weight of end point degrees answers the following
question: Are the connections between high-degree keywords  
1  wi j + wi h
stronger than of the connections between low-degree key- ci  ei j ei h e j h . (6)
words? This parameter can be approximated by a power-law: si (ki − 1) 2
j,h∈N

 θ A greater ci indicates that keyword i has better connectiv-

wi j ∼ ki k j , (4) ity and cohesiveness to other keywords.

123
Journal of Intelligent Manufacturing (2022) 33:425–439 431

Table 3 Topological
characteristics of KCN for the Characteristics 2000–2004 2005–2009 2010–2014 2015–2019
manufacturing science and
Number of nodes 112 134 157 189
engineering literature
Number of edges 129 221 274 573
Average degree 2.3 3.3 3.49 6.06
Maximum degree 20 32 34 48
Average weight 1.55 1.95 1.91 1.91
Maximum weight 24 51 44 29

Results and analyses in cyber-physical manufacturing. However, except for the top
five keywords, the others shift a lot in their rankings. To better
The main topological characteristics of the keyword net- observe the evolution of keywords for the past twenty years,
works for the four-time windows are presented in Table 3. two slope charts are presented in Fig. 3.
Over the two decades, the number of keywords increased Figure 3. shows shifts in the strenght-based ranking of key-
approximately by 68%, which shows that the knowledge words from 2000–2004 to 2015–2019, where a node strength
structure in the manufacturing science and engineering reflects the frequency of its occurrence in the literature. The
literature has experienced a significant expansion by the left side slope-chart with orange line captures the emerging
introduction of new concepts and topics. Over the same topics, which were ranked low in 2000–2004 but rose to top
period, the number of edges connecting the nodes has dramat- ranks in 2015–2019, as shown on the right side of the figure
ically grown by 340%, which implies the significant increase with blue line. Three main categories can be defined based on
in the co-occurrence of keywords in recent papers relative the trends of all keywords in recent two decades, as follows:
to the older ones. The growth in the number of edges also
reflects increasing synergy between different topics, meth- • Fundamental topics: It is evident that genetic algorithm,
ods, and concepts. The number of edges in the 2015–2019 optimization, heuristics, fault diagnosis, and data mining
time window is two times the number of edges in 2005–2009 remain in the top five positions throughout the 2000–2019
or 2010–2014, and almost five times the number of edges in period, as shown in Fig. 3. These rankings indicate the
2000–2004. These observations again indicate rapid growth fundamental role these topics have played in developing
in both the number of technical concepts and topics as well intelligent manufacturing research. Thus, these keywords
as their convergence, as they co-occur more frequently in the are referred to as fundamental topics in intelligent manu-
publications. These observations are also evident from the facturing, which still remain active and important areas of
average degree results (in Table 3), which show the emer- research in manufacture.
gence of diverse research topics described by new keywords • Emerging topics: It can be observed from the left side
or old ones that were not extensively discussed in the earlier slope-chart that the emerging topics are mostly related to
literature. The average and maximum weight indicate that two classes: cyber technologies and data science-oriented
low-degree nodes in the first decade may have been candi- topics. The cyber technologies include Industry 4.0, smart
dates for further investigation in the second decade and that manufacturing, cloud manufacturing, Internet of things,
researchers are working on a combination of emerging and augmented reality and digital manufacturing. The data
relatively more established concepts and topics. science-oriented topics include machine learning, multi-
Table 4 presents the top twenty keywords with the highest objective optimization, Bayesian networks, big data, and
strength (si ) in each time window. It informs which keywords cloud computing. The emerging keywords project the
remained dominant or became obsolete over the years. It also prospects of intelligent manufacturing research and possi-
points out which keywords have been gaining importance in bly the future needs of the industry from both technological
recent years. The strength of most keywords took a big jump and workforce perspectives. New technologies and meth-
from 2000 to 2004 to 2005–2009 but remained relatively sta- ods, especially those that are data science-oriented, have
ble until 2010–2015. In 2015–2019, the keyword strengths become the focus of recent research in manufacturing. The
became three times greater than the strengths in 2000–2004. emerging topics also underscore important technologies
The close examination of emerging keywords indicates that (e.g., prognostics and health management), which are not
cyber-physical manufacturing has gained a lot of traction necessarily”hot topics” like deep learning and augmented
with researchers in 2015–2019. As can be seen, genetic algo- reality.
rithm, optimization, heuristics, fault diagnosis, data mining • Maturing topics: The last category of research topics
are the five most frequently used keywords across all time includes keywords that used to be hot topics last decad
windows in two decades. These keywords describe methods but their rankings have significantly declined since then.

123
432 Journal of Intelligent Manufacturing (2022) 33:425–439

Table 4 Top 20 keywords with the highest strength (si )

2000–2004 2005–2009 2010–2014 2015–2019

Keyword si Keyword si Keyword si Keyword si

Genetic algorithm 227 Genetic algorithm 494 Genetic algorithm 499 Optimization 592
Optimization 184 Optimization 351 Optimization 451 Genetic algorithm 388
Heuristics 79 Heuristics 152 Heuristics 178 Fault diagnosis 240
Fault diagnosis 76 Neural networks 133 Neural networks 129 Data mining 168
Data mining 59 Statistical method 105 Statistical methods 120 Heuristics 157
Neural networks 54 ANOVA 103 Fault diagnosis 119 Multi-objective 142
optimization
Statistical methods 45 Fault diagnosis 101 Model predictive control 98 Neural networks 138
Computer-aided 44 Data mining 100 Data mining 96 ANOVA 119
manufacture
Expert system 40 Image processing 54 ANOVA 95 Industry 4.0 118
Multi-agent systems 35 Ant colony optimization 53 Multi-objective 92 Statistical methods 114
optimization
ANOVA 30 Multi-agent system 53 Dynamic programming 75 Big data 113
Model predictive control 30 Model predictive control 51 Multi-agent systems 67 Model predictive control 112
Mobile robot 28 Multi-objective 47 Ant colony optimization 62 Cloud manufacturing 102
optimization
Image processing 28 Prediction 42 Principal component 58 Machine learning 99
analysis
Condition monitoring 22 Computer-aided 42 Image processing 54 Image processing 97
manufacturing
Linear programming 21 Mobile robot 41 Prediction 50 Smart manufacturing 85
Case-based reasoning 21 Dynamic programming 38 Mobile robot 46 Condition monitoring 84
Prediction 20 Principal component 31 Support vector machine 43 Prediction 83
analysis
Algorithm 19 Condition monitoring 28 Condition-based 40 Internet of things 81
maintenance

Keywords like expert systems, case-based reasoning, and small weights is significantly larger than the number of edges
automated manufacturing systems are examples of topics with large weights. This pattern implies that manufacturing
that have significantly matured over the past two decades literatures has only a few keywords that co-occur with many
and are currently at the backbone of some of the emerging other keywords and, in contrast, a large number of keywords
topics in manufacturing. It is observed that most keywords that co-occur with only a small set of other keywords.
in this category are relatively established manufacturing Figure 5 depicts the average weight of end point degrees.
areas such as computer-aided manufacturing, multi-agent This parameter reflects the correlations between the degrees
systems, tool wear monitoring, distributed manufacturing, of end nodes of links. The average edge weights show an
and multisensor fusion. It is worth noting that the descend- exponential increase with the end point degrees. This behav-
ing rankings of these topics only represent their maturity ior highlights the increase in the weights of connected hubs.
and not lower their importance in practice compared to the In other words, the keyword hubs (i.e., popular keywords)
new top-ranked technologies in recent years. occur in multiple articles dealing with different topics. The
average weight of end point degrees in the keyword network
for each period increases sharply for values of ki k j > 102 ,
Figure 4 shows the probability distribution functions of which indicates that the tendency of co-occurrence increases
weights of KCNs for four time periods, where both x and y sharply for high degree keywords. Except for the 2015–2019
axes are on a logarithmic scale. It is observed all four peri- period (red plot in Fig. 5), most of the weights of 2000–2004,
ods have the decaying pattern for weight distribution, which 2005–2009, and 2010–2014 were found to be close to one,
indicates that edges with large weights occur with low fre- which indicates the presence of a diverse set of keywords that
quency, and in contrast edges with small weights occur with are decentralized, whereas only a few being the hubs.
high frequency. In other words, the number of edges with

123
Journal of Intelligent Manufacturing (2022) 33:425–439 433

Fig. 3 Emerging keywords (left-side panel) and declining keywords (right-side panel) during 2000–2004 and 2015–2019. The number next to the
keyword indicate the ranking based on the keyword strength

Figure 6 indicates the behavior of the average weighted the trends of the four-time windows shows an uncorrelated
nearest-neighbor degrees. As observed, wi j  almost structure, which implies that there are correlations among
remained constant for two decades. The slight variation in different-degree keywords and that high-degree keywords

123
434 Journal of Intelligent Manufacturing (2022) 33:425–439

Fig. 4 Weight distribution of KCNs for the manufacturing science and

engineering literature Fig. 6 Average weighted nearest-neighbor degree of KCN keywords for
the manufacturing science and engineering literature

Fig. 5 Weighted end point degree of KCN keywords for the manufac- Fig. 7 Weighted clustering coefficient of KCN keywords for the man-
turing science and engineering literature ufacturing science and engineering literature

tend to link up with not only other high-degree keywords show an overall behavior similar to that of a structural organi-
but also with low-degree keywords. zation of networks. Exponential growth in average weighted
Figure 7 shows the plot of weighted clustering coeffi- end point degree indicates that keyword pairs with higher
cient versus degree. The continuously decaying coefficients degrees co-occur more frequently. Flat patterns in the aver-
indicate that keywords with high degree present a much age weighted nearest neighbor degree show that there is no
lower clustered neighborhood than low-degree keywords. clear tendency of low degree nodes attaching to low degree
Keywords with low degree co-occur within well-defined key- nodes. That is, there is a pattern of less frequently used
word groups (i.e., high clustering). In contrast, keywords keywords co-occurring with other less frequently used key-
with high degrees co-occur with different groups and com- words. Lastly, the average weighted clustering coefficients
munities, implying that keywords with larger degrees are explain the existence of edges between high degree anfd low
connected to several keywords that do not form cohesive degree keywords.
groups within themselves. Moreover, it is observed that the
trend line in 2015–2019 is in general above the trend lines for
other time windows. It implies that high-degree keywords in Discussion
recent papers tend to connect with relatively a large number
of keywords within interconnected groups. From 2000 to 2019, the number of articles published in the
In summary, the KCN perameters combining the topo- field of manufacturing has increased fivefold, but the num-
logical information with the weight distribution of networks ber of keywords has increased by only 68%. Though the

123
Journal of Intelligent Manufacturing (2022) 33:425–439 435

manufacturing literature has grown considerably in volume, inform knowledge and skills needs of intelligent manufactur-
it has not seen proportional growth in the variety of key- ing workforce, which has been experiencing high demand.
words, indicating that new topics have been introduced at a Influenced by Industry 4.0, cyber-physical manufacturing
much slower pace than the growth of the body of manufac- technologies grew more complex and more knowledge-
turing knowledge. In other words, some keywords remained intensive. Cyber technologies, over the last five years, have
active over the two-decade period. The new keywords in the expanded and enriched the body of the literature, which
literature emerged from research articles on intelligent and was principally composed by the physical technologies in
cyber manufacturing, which experienced a rapid expansion, the early twentieth century. Methods with growing popu-
and the research in this field experience a period of innova- larity can be grouped into three categories: data science
tion. While the keywords grew by two-thirds, the connections (machine learning, deep learning, and big data), cloud-based
between the keywords have tripled, implying that manufac- technologies (cloud computing and internet of things), and
turing research is moving toward the convergence of different cyber-physical technologies (augmented reality and digital
topics. twins). As the research in physical manufacturing dwindled,
Although the size of KCNs increased steadily over four the research focus shifted to the emerging cyber technologies
periods, the average weight changed only slightly. This that enhance the performance and efficiency of manufac-
pattern indicates that the topology of the manufacturing- turing machines, equipment, and systems; and this trend is
keyword co-occurrence network is a small world. The skewed expected to continue in the future.
weight distribution and clustering coefficients illustrate that From the topological characteristics of KCNs, it is
most keywords have a low degree, and most links have a observed that the connectivity among keywords has enhanced
low weight. In other words, the networks have a small num- with time, from 2000 to 2019. The network hubs represent
ber of keywords with a high degree and only a few links versatile and perennial technologies such as genetic algo-
with large weights. This node degree and link weight pattern rithm, optimization, fault diagnosis, and neural networks.
might be due to the tendency of researchers to pick often- While some keywords have more cohesiveness in the first
used keywords in their papers relative to less frequently used decade, that is not the case in the second decade during which
keywords. This could also be due to the possibility that only keywords tend to connect with a variety of emerging topics.
a small select set of keywords tend to be widely discussed in This indicates that researchers, in the last decade, were try-
literature. ing to explore the synergies among various topics instead of
In the manufacturing literature, we observe only a few focusing on only a run-of-the-mill topic. The convergence
keywords (e.g., genetic algorithm, optimization, neural net- of different areas has become a trend with time. Overall,
works, industry 4.0) co-occur with many other keywords, but the community of cyber-physical manufacturing has prolif-
a large number of keywords co-occur with only a small set of erated, and research is moving toward greater convergence.
other keywords. High-degree keywords (e.g., genetic algo- One of the important goals of this paper is to identify the
rithm, optimization, fault diagnosis, heuristics) tend to link specific domain knowledge associated with emerging intelli-
up with not only other high-degree keywords but also with gent and cyber manufacturing technologies. The nine pillars
low-degree keywords. The less frequently (e.g., mobile robot, of Industry 4.0 (Vaidya et al., 2018) are utilized as a reference
predictive maintenance, k-means clustering) occurring key- to categorize and organize the keywords, as shown in Fig. 8.
words tend to appear as part of only a few keyword groups, The top 50 most frequently used keywords (ranked based
but most frequently occurring words (e.g., genetic algorithm, on their strength) in 2015–2019 on the left side are subjec-
optimization, heuristics, fault diagnosis, data mining) co- tively mapped to one or multiple pillars of Industry 4.0 on the
occur with keywords positioned in a wide variety of keyword right side. Based on the authors’ subjective evaluations, “Au-
groups. This observation implies that the most frequently tonomous Robots”, “Internet of Things”, and “Big data” are
used manufacturing keywords are associated with a diverse the top three pillars with the highest number of connections
set of research tracks in manufacturing. That is to say, the with research keywords, which indicates their popularity in
keyword hubs (popular keywords such as genetic algorithm, the research area and their role in informing the technology
optimization) occur in articles covering diverse topics. Five trends in intelligent manufacturing. The mapping identifies
keywords—genetic algorithm, optimization, heuristics, fault the specific domain knowledge associated with emerging
diagnosis, data mining—are the most frequently used key- intelligent and cyber manufacturing technologies and can
words across all four time windows in two decades. Machine inform the design and development of new curricula for the
learning, Industry 4.0, big data, cloud manufacturing, and future manufacturing workforce. The number of links to the
deep learning are the emerging keywords indicating the dawn nine Industry 4.0 pillars indicate the relevance of the top-
of cyber manufacturing. These observations inform future ics (high-requency keywords) in their respective areas. The
manufacturing research, technology trends. This could also keywords linked to the pillars suggest possible topics to be
covered in those courses.

123
436 Journal of Intelligent Manufacturing (2022) 33:425–439

Fig. 8 Subjective mapping between the top 50 keywords ranked based on their strength (left) and the nine pillars of Industry 4.0 (right). The number
next to the keyword indicate the number of keywords from the other side are connected to it

123
Journal of Intelligent Manufacturing (2022) 33:425–439 437

One of our recent studies analyzes the data science and cles to capture more in-depth insights into the scientific and
analytics skills gap in today’s manufacturing workforce (Li technological trends in intelligent and cyber manufacturing.
et al., 2021). The gap analysis was conducted on a com- This work identified the high-frequency keyworkds asso-
prehensive labor market analytics data, (i.e., Emsi) which ciated with the nine pillars of Industry 4.0, namely sim-
contains job posting and profile data providing insights into ulation, autonomous robots, system integration, additive
the trends and potential of high in-demand skills for future manufacturing, big data, augmented reality, IoT, cloud com-
manufacturing jobs. By connecting it with this work, it is puting, and cybersecurity. These keywords suggeste the
noticed that major domain knowledge gaps existing in data possible topics for future research and education in intel-
science-related jobs are machine learning, big data analyt- ligent manufacturing.
ics, data mining and optimization, which are aligned with
the emerging topics in manufacturing research. The trend of Acknowledgements This material is based upon work supported by the
National Science Foundation under Grant No. 1935646. Any opinions,
research focus shifted to the emerging cyber technologies findings, and conclusions, or recommendations expressed in this mate-
which enhance the performance and efficiency of manufac- rial are those of the author(s) and do not necessarily reflect the views
turing machines, equipment, and systems. It is also reflected of the National Science Foundation.
in the U.S. labor market which has experienced 8% growth in
Funding National Science Foundation (Grant No.1935646) funded by
the total number of job postings related to cyber technologies Sagar Kamarthi
in 2020..
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing, adap-
tation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
Conclusions source, provide a link to the Creative Commons licence, and indi-
cate if changes were made. The images or other third party material
This paper conducted a keyword co-occurrence network in this article are included in the article’s Creative Commons licence,
analysis of intelligent and cyber manufacturing-related key- unless indicated otherwise in a credit line to the material. If material
is not included in the article’s Creative Commons licence and your
words in the scientific 84,041 papers published in top-tier intended use is not permitted by statutory regulation or exceeds the
manufacturing journals during the period between 2000 and permitted use, you will need to obtain permission directly from the copy-
2019. The analysis reveals insight into knowledge struc- right holder. To view a copy of this licence, visit http://creativecomm
tures of the manufacturing field and the evolution of research ons.org/licenses/by/4.0/.
themes for the past two decades. To analyze keywords in all
publications, we built a network of keywords that are relevant
to the broader areas of intelligence and cyber manufacturing.
The statistical analysis of network characteristics effectively
analyzes the trends and patterns in research topics (e.g., cloud References
manufacturing, data mining, and machine learning as applied
Aggour, K. S., et al. (2019). Artificial intelligence/machine learning in
to manufacturing) over time. Traditional literature review
manufacturing and inspection: A GE perspective. MRS Bulletin,
methods discuss key features of technical methods, general 44(7), 545–558. https://doi.org/10.1557/mrs.2019.157
methodologies, and experimental findings. In contrast, the Alam, K. M., & El Saddik, A. (2017). C2PS: A digital twin architec-
KCN-based approach gives a macro-level understanding of ture reference model for the cloud-based cyber-physical systems.
IEEE Access, 5, 2050–2062. https://doi.org/10.1109/ACCESS.
research trends and the evolution of emerging technologies. It 2017.2657006
enables macro-level quantitative analysis and in-depth, com- Alcácer, V., & Cruz-Machado, V. (2019). Scanning the Industry 4.0:
prehensive reviews with less manual effort. This approach A Literature review on technologies for manufacturing systems.
can be readily applied to understand other scientific fields Engineering Science and Technology, an International Journal,
22(3), 899–919. https://doi.org/10.1016/j.jestch.2019.01.006
macroscopically. Barari, A., de Sales Guerra, M., Tsuzuki, Y. C., & Macchi, M. (2021).
The analysis of only keywords to observe the knowledge Editorial: intelligent manufacturing systems towards industry 4.0
structure and trends can lead to some biases. For exam- era. Journal of Intelligent Manufacturing. https://doi.org/10.1007/
ple, some authors may not include an important term in s10845-021-01769-0
Baroroh, D. K., Chu, C.-H., & Wang, L. (2020). Systematic literature
the keyword section, but it otherwise plays a significant review on augmented reality in smart manufacturing: Collabo-
role in the subject of the paper. Future work will extract ration between human and computational intelligence. Journal
words from the title and abstract to enrich the keywords of Manufacturing Systems. https://doi.org/10.1016/j.jmsy.2020.
database to better reveal the keywords correlation in the body 10.017
Barrat, A., Barthélemy, M., Pastor-Satorras, R., & Vespignani, A.
of knowledge. The keyword co-occurrence network-based (2004). The architecture of complex weighted networks. Proceed-
methodology can also potentially benefit from extensive text ings of the National Academy of Sciences, 101(11), 3747–3752.
mining and natural language processing of published arti- https://doi.org/10.1073/pnas.0400087101

123
438 Journal of Intelligent Manufacturing (2022) 33:425–439

Chen, T. C. T. (2017). Cloud intelligence in manufacturing. Journal of Majstorović, V. D. et al. (2018) Cyber-physical manufacturing in
Intelligent Manufacturing, 28(5), 1057–1059. https://doi.org/10. context of industry 4.0 model. In Lecture Notes in Mechanical
1007/s10845-015-1122-9 Engineering, vol. 0, no. 9783319895628, Pleiades Publishing,
Choudhary, A. K., Harding, J. A., & Tiwari, M. K. (2009). Data min- pp. 227–238.
ing in manufacturing: A review based on the kind of knowledge. Miškuf, M., & Zolotová, I. (2016). Comparison between multi-class
Journal of Intelligent Manufacturing, 20(5), 501–521. https://doi. classifiers and deep learning with focus on industry 4.0. In 2016
org/10.1007/s10845-008-0145-x Cybernetics & Informatics (K&I) (pp. 1–5). IEEE.
Cui, Y., Kara, S., & Chan, K. C. (2020). Manufacturing big data ecosys- Moghaddam, M., Cadavid, M. N., Kenley, C. R., & Deshmukh, A. V.
tem: A systematic literature review. Robotics and Computer- (2018). Reference architectures for smart manufacturing: A criti-
Integrated Manufacturing, 62, 101861. https://doi.org/10.1016/j. cal review. Journal of Manufacturing Systems. https://doi.org/10.
rcim.2019.101861 1016/j.jmsy.2018.10.006
Da Xu, L., He, W., & Li, S. (2014). Internet of things in industries: Moghaddam, M., Cadavid, M. N., Kenley, C. R., & Deshmukh, A. V.
A survey. IEEE Transactions on Industrial Informatics, 10(4), (2018). Reference architectures for smart manufacturing: A critical
2233–2243. https://doi.org/10.1109/TII.2014.2300753 review. Journal of Manufacturing Systems, 49, 215–225. https://
De Felice, F., Petrillo, A., & Zomparelli, F. (2018). A bibliometric doi.org/10.1016/j.jmsy.2018.10.006
multicriteria model on smart manufacturing from 2011 to 2018. Moghaddam, M., & Nof, S. Y. (2017). The collaborative factory of
IFAC-PapersOnLine, 51(11), 1643–1648. https://doi.org/10.1016/ the future. International Journal of Computer Integrated Manu-
j.ifacol.2018.08.221 facturing, 30(1), 23–43. https://doi.org/10.1080/0951192X.2015.
Duvvuru, A., Kamarthi, S., & Sultornsanee, S. (2012). Undercovering 1066034
research trends: Network analysis of keywords in scholarly arti- Monostori, L., et al. (2016). Cyber-physical systems in manufactur-
cles. In 2012 Ninth International Conference on Computer Science ing. CIRP Annals—Manufacturing Technology, 65(2), 621–641.
and Software Engineering (JCSSE) (pp. 265–270). IEEE. https://doi.org/10.1016/j.cirp.2016.06.005
Escobar, C. A., McGovern, M. E., & Morales-Menendez, R. (2021). Mourtzis, D., Vlachou, E., & Milas, N. (2016). Industrial big data as
Quality 4.0: a review of big data challenges in manufacturing. a result of IoT adoption in manufacturing. Procedia CIRP, 55,
Journal of Intelligent Manufacturing, 1–16. 290–295. https://doi.org/10.1016/j.procir.2016.07.038
Franco, D., Miller Devós Ganga, G., de Santa-Eulalia, L. A., & God- Nguyen Ngoc, H., Lasa, G., & Iriarte, I. (2021). Human-centred design
inho Filho, M. (2020). Consolidated and inconclusive effects of in industry 4.0: case study review and opportunities for future
additive manufacturing adoption: A systematic literature review. research. Journal of Intelligent Manufacturing. https://doi.org/10.
Computers & Industrial Engineering, 148, 106713. https://doi.org/ 1007/s10845-021-01796-x
10.1016/j.cie.2020.106713 Nti, I. K., Adekoya, A. F., Weyori, B. A., & Nyarko-Boateng, O. (2021).
Henzel, R., & Herzwurm, G. (2018). Cloud Manufacturing: A state- Applications of artificial intelligence in engineering and manufac-
of-the-art survey of current issues. Procedia CIRP, 72, 947–952. turing: A systematic review. Journal of Intelligent Manufacturing.
https://doi.org/10.1016/j.procir.2018.03.055 https://doi.org/10.1007/s10845-021-01771-6
Hofmann, E., & Rüsch, M. (2017). Industry 4.0 and the current status NWB Team. (2006). Network Workbench Tool. Indiana University,
as well as future prospects on logistics. Computers in Industry, 89, Northeastern University, and University of Michigan. http://nwb.
23–34. https://doi.org/10.1016/j.compind.2017.04.002 slis.indiana.edu.
Järvenpää, E., Siltala, N., Hylli, O., & Lanz, M. (2019). The devel- Onel, S., Zeid, A., & Kamarthi, S. (2011). The structure and analysis of
opment of an ontology for describing the capabilities of manu- nanotechnology co-author and citation networks. Scientometrics.
facturing resources. Journal of Intelligent Manufacturing, 30(2), https://doi.org/10.1007/s11192-011-0434-6
959–978. https://doi.org/10.1007/s10845-018-1427-6 Ooi, K. B., Lee, V. H., Tan, G. W. H., Hew, T. S., & Hew, J. J. (2018).
Kusiak, A. (2018). Smart manufacturing. International Journal of Cloud computing in manufacturing: The next industrial revolu-
Production Research, 56(1–2), 508–517. https://doi.org/10.1080/ tion in Malaysia? Expert Systems with Applications, 93, 376–394.
00207543.2017.1351644 https://doi.org/10.1016/j.eswa.2017.10.009
Kusiak, A. (2019). Editorial: Intelligent manufacturing: Bridging two Oztemel, E., & Gursev, S. (2020). Literature review of industry 4.0 and
centuries. Journal of Intelligent Manufacturing, 30(1), 1–2. https:// related technologies. Journal of Intelligent Manufacturing, 31(1),
doi.org/10.1007/s10845-018-1455-2 127–182. https://doi.org/10.1007/s10845-018-1433-8
Kutschenreiter-Praszkiewicz, I. (2008). Application of artificial neural Radhakrishnan, S., Erbis, S., Isaacs, J. A., & Kamarthi, S. (2017). Novel
network for determination of standard time in machining. Journal keyword co-occurrence network-based methods to foster system-
of Intelligent Manufacturing, 19(2), 233–240. https://doi.org/10. atic reviews of scientific literature. PLoS ONE, 12(3), 1–16. https://
1007/s10845-008-0076-6 doi.org/10.1371/journal.pone.0172778
Li, G., Yuan, C., Kamarthi, S., Moghaddam, M., & Jin, X. (2021). Rajkumar, R., Lee, I., Sha, L., Stankovic, J. (2010) Cyber-physical
Data science skills and domain knowledge requirements in the systems: The next computing revolution. In Proceedings—design
manufacturing industry: A gap analysis. Journal of Manufacturing automation conference, pp. 731–736, doi: https://doi.org/10.1145/
Systems. https://doi.org/10.1016/j.jmsy.2021.07.007 1837274.1837461.
Li, H., An, H., Wang, Y., Huang, J., & Gao, X. (2016). Evolu- Recommendations for implementing the strategic initiative INDUS-
tionary features of academic articles co-keyword network and TRIE 4.0 April 2013 Securing the future of German manufacturing
keywords co-occurrence network: Based on two-mode affiliation industry Final report of the Industrie 4.0 Working Group.
network. Physica A: Statistical Mechanics and its Applications, Schuh, G., Anderl, R., Gausemeier, J., ten Hompel, M., & Wahlster, W.
450, 657–669. https://doi.org/10.1016/j.physa.2016.01.017 (2017). Industrie 4.0 maturity index. Munich: Herbert Utz.
Lozano, S., Calzada-Infante, L., Adenso-Díaz, B., & García, S. (2019). Shibata, N., Kajikawa, Y., Takeda, Y., Sakata, I., & Matsushima,
Complex network analysis of keywords co-occurrence in the recent K. (2011). Detecting emerging research fronts in regenerative
efficiency analysis literature. Scientometrics, 120(2), 609–629. medicine by the citation network analysis of scientific publications.
https://doi.org/10.1007/s11192-019-03132-w Technological Forecasting and Social Change, 78(2), 274–282.
Lu,Y., Morris, K. C., & Frechette, S. (2016). Current standards land- https://doi.org/10.1016/j.techfore.2010.07.006
scape for smart manufacturing systems, National Institute of Shrouf, F., Ordieres, J. Miragliotta, G. (2014) Smart factories in
Standards and Technology, NISTIR, Vol. 8107, p. 39. Industry 4.0: A review of the concept and of energy manage-

123
Journal of Intelligent Manufacturing (2022) 33:425–439 439

ment approached in production based on the Internet of Things Xu, X. (2012). From cloud computing to cloud manufacturing. Robotics
paradigm. In IEEE international conference on industrial engi- and Computer-Integrated Manufacturing, 28(1), 75–86. https://
neering and engineering management, vol. 2015, pp. 697–701, doi.org/10.1016/J.RCIM.2011.07.002
doi: https://doi.org/10.1109/IEEM.2014.7058728. Yao, X., Zhou, J., Lin, Y., Li, Y., Yu, H., & Liu, Y. (2019). Smart manu-
Su, H. N., & Lee, P. C. (2010). Mapping knowledge structure by key- facturing based on cyber-physical systems and beyond. Journal of
word co-occurrence: A first look at journal papers in Technology Intelligent Manufacturing, 30(8), 2805–2817. https://doi.org/10.
Foresight. Scientometrics, 85(1), 65–79. https://doi.org/10.1007/ 1007/s10845-017-1384-5
s11192-010-0259-8 Zhong, W., An, H., Gao, X., & Sun, X. (2014). The evolution of commu-
Thoben, K.-D., Wiesner, S., & Wuest, T. (2017). ‘Industrie 4.0’ and nities in the international oil trade network. Physica A: Statistical
smart manufacturing—A review of research issues and application Mechanics and its Applications, 413, 42–52. https://doi.org/10.
examples. International Journal of Automation Technology, 11(1), 1016/j.physa.2014.06.055
4–16. https://doi.org/10.20965/ijat.2017.p0004
Tian, G. Y., Yin, G., & Taylor, D. (2002). Internet-based manufacturing:
A review and a new infrastructure for distributed intelligent man- Publisher’s Note Springer Nature remains neutral with regard to juris-
ufacturing. Journal of Intelligent Manufacturing, 13(5), 323–338. dictional claims in published maps and institutional affiliations.
https://doi.org/10.1023/A:1019907906158
Usuga Cadavid, J. P., Lamouri, S., Grabot, B., Pellerin, R., & Fortin,
A. (2020). Machine learning applied in production planning and
control: a state-of-the-art in the era of industry 4.0. Journal of
Intelligent Manufacturing, 31(6), 1531–1558. https://doi.org/10.
1007/s10845-019-01531-7
Vaidya, S., Ambad, P., & Bhosle, S. (2018). Industry 4.0—A glimpse.
Procedia Manufacturing, 20, 233–238. https://doi.org/10.1016/j.
promfg.2018.02.034

123