International Journal of Production Economics xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

International Journal of Production Economics

journal homepage: www.elsevier.com/locate/ijpe

The smart factory as a key construct of industry 4.0: A systematic literature

review
Philipp Osterrieder∗, Lukas Budde, Thomas Friedli
University of St.Gallen, Institute for Technology Management, Switzerland

A R T I C LE I N FO A B S T R A C T

Keywords: Industry 4.0 is a ubiquitous term throughout general newspapers, on company websites or in scientific journals.
Smart factory One of its key constructs is the smart factory, envisioned as a future state of a fully connected manufacturing
Smart manufacturing system, mainly operating without human force by generating, transferring, receiving and processing necessary
Industry 4.0 data to conduct all required tasks for producing all kinds of goods. Although the understanding of smart factory
Internet-of-Things
concepts has been sharpened in the last years, it is still difficult for industrial companies to establish a concrete
Big data
Literature review
strategy roadmap within the jungle of different terminologies, ideas and concepts. To generate further clarity
and to consolidate the previous findings around smart factory for researchers as well as for practitioner, we
conducted a systematic literature review. For this purpose, we chose a five steps approach: Scope definition,
topic conceptualisation, literature search, literature analysis and synthesis, and synthesis of future research
questions. During our review, we found that research within each perspective is fragmented and unequally
advanced. Most publications treat single use cases with low generalizability, often rely upon machine data,
typically have a technical nature and seldom incorporate impact estimations. In this paper, our key academic
and practical contribution lies in the categorisation of the selected publications into eight thematic distinct
perspectives within the sphere of smart factory: Decision making, cyber-physical systems, data handling, IT
infrastructure, digital transformation, human machine interaction, IoT, and cloud manufacturing and services.
These are further developed into the smart factory research model, stating a foundation for future research
endeavors.

1. Introduction Despite the increasing devotion of academia and practitioners, the

research field around Industry 4.0 remains fragmented and spotty (Lee
Industry 4.0, the Internet-of-Things (IoT), big data, artificial in- and Lee, 2015). This fact becomes apparent when we think of the
telligence (AI) and cloud computing. These are few among other terms various perspectives from which the topic is approached. These may
in vogue that are ubiquitous throughout general newspapers, on com- have a rather technical peculiarity, such as the decrease of failure rates
pany websites or in scientific journals. The terms represent a number of of manufacturing systems, or adopt a business process perspective, for
concepts, tools and methods with the ability to change the modern instance. One of the key concepts within Industry 4.0 is the so-called
society and industry radically. Therefore, numerous authors have al- Smart Factory, also termed digital or intelligent factory (Lucke et al.,
ready discussed the origin (Almada-Lobo, 2015), examples (Lasi et al., 2008; Stock and Seliger, 2016). The smart factory depicts a future state
2014), the impact (Schmidt et al., 2015) and future paths of Industry of a fully connected manufacturing system, mainly operating without
4.0 in many diverse publications (Qin et al., 2016). While the expres- human force by generating, transferring, receiving and processing the
sion Industry 4.0 has been coined in Germany, to account for the necessary data to conduct all required tasks for producing all kinds of
emerging digitalisation, or digital transformation, of both products and goods (cf. Lasi et al., 2014). Again, it is coherent that the smart factory
production systems in the industrial sector, the United States labelled it concept may be approached with altering focal points and therefore
Smart Manufacturing (Porter and Heppelmann, 2015; Thoben et al., cannot be covered by one single research stream.
2017). Both approaches encompass a plurality of core technologies, This phenomenon was previously acknowledged by recent articles,
methodologies and trends (Stock and Seliger, 2016), which gained at- which made an effort to review the existing knowledge on Industry 4.0
tention in the past few years by industry and the research community. (Hofmann and Rüsch, 2017; Kirazli and Hormann, 2015; Lasi et al.,

∗
Corresponding author. Dufourstrasse 40a, 9000, St.Gallen, Switzerland.
E-mail address: [email protected] (P. Osterrieder).

https://doi.org/10.1016/j.ijpe.2019.08.011
Received 2 November 2017; Received in revised form 31 July 2019; Accepted 26 August 2019
0925-5273/ © 2019 Published by Elsevier B.V.

Please cite this article as: Philipp Osterrieder, Lukas Budde and Thomas Friedli, International Journal of Production Economics,
https://doi.org/10.1016/j.ijpe.2019.08.011
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

2014; Lu, 2017), big data (Sheng et al., 2017) and smart manufacturing findings, discuss their meaning for academia and practitioners, and end
(Kang et al., 2016). Lu (2017) found 23 articles treating the topic of the with a structured research agenda.
smart factory and smart manufacturing, following the literature search
methodology by Webster and Watson (2002). However, this number 2. Research background
appears to be relatively low. Additionally, in most cases the research
discusses single use cases on operative level dealing with the integra- 2.1. The smart factory concept
tion of machine and other internal data (Lee et al., 2014; Rezaei et al.,
2017; Wuest et al., 2014). Only few relate the studied problem to a The importance of the smart factory for academia and future man-
specific big data analytics method (Arunachalam et al., 2017; ufacturing is reflected by a not negligible attention coming from
Chongwatpol, 2015), or to the needed data type (Bauer et al., 2016; scholars, whose impetus was to describe the smart factory concept by
Tang et al., 2016). Moreover, the integration into a management system setting boundaries and making it more tangible for practitioners. This
and especially the impact quantification is under-researched. Malakooti lead to the emergence of multiple conceptualisations, yet, demon-
and Raman (2000) are among the few to discuss these topics, while strating a clear trend and similar basis. Interestingly, recent articles
lacking a generalizable approach. Demonstrating that research in the (Lasi et al., 2014; Thoben et al., 2017) refer to the description of Lucke
sphere of the smart factory is heterogeneous and unequally advanced. et al. (2008), who coped with this topic years before it was declared a
Therefore, especially Thoben et al. (2017) and O'Donovan et al. key pillar of the Industry 4.0 initiative (acatech, 2013). However, early
(2015a) point out the need for progressing research in the field of In- descriptions of the smart factory concept share the fact of not reaching
dustry 4.0 and explicitly call for the urge to condense and structure the as far in the degree of automation and absence of human workforce as
emerging knowledge. This was the impetus to follow the subsequent publications that are more recent. For instance, Lucke et al. (2008) refer
research questions during our literature review: to the smart factory as a manufacturing environment, where humans
and production processes are supported by the presence of intelligent,
RQ1. How far advanced is the research around the smart factory
computer-based systems ensuring a seamless, continuous flow of pro-
concept?
duction for increased performance and quality. While newer concepts
RQ2. How can the diverse perspectives be clustered with regard to their of Stock and Seliger (2016) or Suginouchi et al. (2017) believe in the
thematic context? autonomous cooperation of all machines needed for the respective
production process by using data originating from numerous sensors.
RQ3. What work needs to be done?
Wang et al. (2016a) and Qin et al. (2016) enhance this vision by of-
First, we hereby aim to provide an exhaustive overview onto this fering a more distinct view on the layers of and vertical integration
research topic and depict the current progress. Second, by relating the central for the smart factory.
advancements in this research field with future promising pathways, we Stripped down to the essential elements a smart factory is con-
strive to offer a foundation for upcoming studies. Third, we propose taining according to the authors treating this topic, we found con-
categories within the field of the smart factory that may and should be sistency among the following aspects. To some extent, a smart factory
used to increase the clarity about the perspective a researcher is always possesses machines equipped with sensors and actors, able to
adopting throughout her or his studies. collect, send, receive, process data and act accordingly (Stock and
To reach these targets we followed the systematic literature search Seliger, 2016; Wang et al., 2016a). These machines communicate with
methodology introduced by Vom Brocke et al. (2009) to increase the each other to fulfil predefined tasks and the system of machines orga-
accuracy and profound nature of our study. While analysing the sample nizes, as well as configures itself purposefully to reach the same ob-
of articles, we applied qualitative content analysis (Mayring, 2000) and jective (Suginouchi et al., 2017; Tang et al., 2016). The system is thus
adopted a coding style in accordance with the approach introduced by monitored by a higher entity, i.e. a master software model, a human or
Gioia et al. (2013) and Corbin and Strauss (2009). This method was a combination of both, and follows the directions of a sophisticated
transferred from qualitative research, where it is mainly applied to computer program (Lucke et al., 2008; Zuehlke, 2008).
interview data, for narrowing down the bulk of literature into distinct Consequently, as other scholars already discovered, the production
categories. plant cannot longer be seen as a plane of activities, instead it fans out
The adopted strategy led to the emergence of the subsequent eight into four distinct layers (Wang et al., 2016a). These are the physical,
classes, derived from a final number of 106 investigated publications: data, cloud & intelligence and control layer as depicted in Fig. 1.
(1) AI & Manufacturing; (2) Data Handling; (3) Supply Chain Adapted from Wang et al. (2016a) and Heidel et al. (2017), we assign
Management; (4) Decision-making; (5) IoT; (6) Digital Transformation; all machines, the shopfloor as a whole and those activities actual taking
(7) IT Infrastructure; and (8) Theoretical Contributions. Related to the place to the physical layer. The data layer incorporates the process of
suggested smart factory research model, we furthermore found that data transfer from the machines (sensors) to the cloud and vice versa,
among other proposed future research directions, exploring use cases while software is controlling what data is sent/received (data type and
for intelligent software tools in manufacturing and data generation, variety) and at which rate (velocity and volume). The data is then
data processing, as well as data mining frameworks, are very promi- stored, at least temporarily, in the cloud, where it can be processed by
nent. sophisticated analytics. On the top layer, the supervision is taking place.
These results demonstrate our main contribution to advance the Here, the master program controlling the smart factory can be adapted
research field of smart factory. Meaning, we offer a structured overview by human intervention whenever necessary.
on the relevant bulk of literature within this domain and a research Understanding the current view on the smart factory concept is vital
agenda that may motivate others to follow the proposed questions. Yet, to connect with the academic progress and future promising research
above all, by adopting a novel way of analysing the found literature streams in sub-areas of the smart factory presented in Chapter 4 and 5.
with acknowledged means inherited from qualitative research, we Hence, we introduce our applied research methodology in the sub-
contribute with eight distinct categories researchers may use to differ- sequent paragraphs to demonstrate the approach on how we con-
entiate their studied problem. structed the smart factory research model.
The remainder of the literature review at hand is structured as
follows. First, we explain the research background in the subsequent 2.2. Research methodology
chapter, introduce the smart factory concept, how we proceeded in
finding and analysing the literature in detail, before presenting the 2.2.1. Literature review method
investigated literature systematically. Afterwards, we elaborate on the Literature reviews are an important and qualified mean (Baker,

2
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Table 1
Cooper (1988) taxonomy.
Characteristics Categories

(a) Focus Research Research Theories Applications

Outcomes Methods
(b) Goal Integration Criticism Central Issues
(c) Organisation Historical Conceptual Methodological
(d) Perspective Neutral Representation Espousal of Position
(e) Audience Specialised General Practitioners/ General Public
Scholars Scholars Politicians
(f) Coverage Exhaustive Exhaustive Representative Central/Pivotal
& Selective

Cooper (1988) modelencompassing six different aspects (Focus, Goal,

Organisation, Perspective, Audience and Coverage), in turn comprising
up to four classes, as it can be seen in Table 1. One peculiarity of this
model is that some of these classes can only be used exclusively within
one characteristic, while others are able to be paired up (Vom Brocke
et al., 2009).
Referring to Vom Brocke et al. (2009), the second phase treats the
Fig. 1. The 4 Layer Smart Factory Concept. Own illustration adapted from
conceptualisation of the topic with the target to narrow down potential
Wang et al. (2016a) and Heidel et al. (2017).
expressions, terms and search phrases to the most relevant key words.
Here, we need to find the appropriate research streams and filter
2000; Cooper, 1988) to summarise (Fischl et al., 2014) and progress the commonly accepted and used expressions to ensure an unequivocal
current state of knowledge in a specific field of interest (Denyer and conversation between academia and industry (Fischl et al., 2014). It is
Tranfield, 2009; Rowley and Slack, 2004; Tranfield et al., 2003). Good tremendous to reach a clearly defined set of key words (Rowley and
literature reviews make an endeavour to assess, structure and reveal the Slack, 2004), since these determine the completeness and quality of the
most relevant scientific publications within a certain domain (Denyer papers found later in the search process (Baker, 2000). To identify the
et al., 2008; Tranfield et al., 2003). They ultimately unfold research relevant set of search phrases, working definitions (Zorn and Campbell,
gaps and present the corresponding future research opportunities in an 2006), books, encyclopaedias (Baker, 2000) and concept mapping are
organised manner using narratives, charts and tables (Fischl et al., possible elements that may support the researcher (Rowley and Slack,
2014). Reliable literature reviews follow a systematic procedure to in- 2004).
crease rigour, validity and relevance (Tranfield et al., 2003; Vom After defining the set of search terms, Vom Brocke et al. (2009)
Brocke et al., 2009). This is because a systematic approach provides the suggest following the literature search process in the third phase by
opportunity for replicating the findings and demonstrates transparency, conducting the subsequent four steps. First, the researcher accumulates
also through thoroughness in documenting the complete literature relevant journals that cover the identified research streams. Here,
search and review procedure (Denyer and Tranfield, 2009; Vom Brocke Webster and Watson (2002) advise to rely on peer-reviewed journals
et al., 2009). Whereas a non-systematic literature review process cannot and conference proceedings to ensure high quality (Fischl et al., 2014).
be certain in providing a sufficient set of scientific articles, leading to a Second, databases containing these journals need to be chosen care-
weak assessment base for important decisions in academia and industry fully. Third, Vom Brocke et al. (2009) explain to conduct the keyword
(Fischl et al., 2014). Contrary, findings derived from literature reviews search by altering the identified search algorithms in each of the se-
based on evidence from a transparent and meticulous approach can lected databases. Finally, the derived articles are used as a starting
help guiding academia, industry and policymakers (Baker, 2000; point for executing a forward and backward search in the fourth step. A
Denyer and Tranfield, 2009). backward search depicts the mechanism of screening the cited refer-
ences by the found article, while the ultimate aim of a forward search is
2.2.2. Literature search proceeding to unfold those articles citing the identified artefact (Vom Brocke et al.,
With the impetus to answer the posed research questions with ri- 2009).
gour and due to the named advantages, we adopted a systematic lit- The fourth phase embraces the analysis of the final bulk of relevant
erature review procedure following the steps proposed by Vom Brocke articles and the synthesis of the main findings, similarities and insights
et al. (2009). In contrast to other suggested methodologies published in governed during the analysis, while setting the individual publications
the last decade (Colicchia and Strozzi, 2012; David and Han, 2004; in context of each other (Denyer and Tranfield, 2009; Vom Brocke et al.,
Newbert, 2007; Seuring and Gold, 2012) that mainly encompass ana- 2009).
lysis and synthesis steps (Fischl et al., 2014), we chose Vom Brocke In the final phase of the literature review process, a research agenda
et al. (2009) due to its recipe-alike structure covering all important is built by collecting and condensing future research questions des-
phases within a literature review. This approach consists of the fol- cended from the found articles and conclusions drawn by the researcher
lowing five steps (Vom Brocke et al., 2009): (Vom Brocke et al., 2009; Webster and Watson, 2002).

1) Scope definition
2) Topic conceptualisation 2.2.3. Data analysis approach
3) Literature search process The applied data, thus literature analysis methodology is split into
4) Literature analysis and synthesis two parts.
5) Synthesis of future research questions The review of the emerging literature was executed through quali-
tative content analysis (Mayring, 2000). Here, following the sugges-
During the first phase the range of the review is defined and Vom tions of Mayring (2000), the content of the articles is divided into units
Brocke et al. (2009) advise to refer to existing taxonomies that already of analysis and then investigated step by step. With the support of afore
proofed validity. Recommended by Vom Brocke et al. (2009), we apply created and meanwhile optimised tables, the analysis process was

3
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

structured and the important information extracted (Denyer and rather conceptual, since we strive to derive certain research clusters
Tranfield, 2009; Tranfield et al., 2003). Another possibility is to take within the field of smart factory.
advantage of concept frameworks (Salipante et al., 1982) and classical IV. The perspective of the researcher is treated in the fourth category.
storytelling tactics for simplifying the analysis procedure (Hammersley, Cooper (1988) proposes to either take a certain point of view to
2001). express a specific position, or to represent the bulk of literature
During the analysis of the present data, scholars have to rely on neutrally. For us, it seemed most suitable to present the literature
their “theoretical sensitivity” (Suddaby, 2006, p. 640) for discovering and our findings with a neutral position.
novel findings. Therefore, while analysing the found articles with an V. The fifth category reflects the audience a researcher is trying to
open mind (Edmonson and McManus, 2007) we realised that the reach. The audience affects the writing style the authors are ap-
journals publishing these, plus the covered topics, are very diverse. plying (Bem, 1995) and vice versa, the authors have to select the
Hence, we transferred coding techniques as an analysis method in- appropriate journal to reach the intended addressees. As indicated
herited from qualitative research for the development of constructs by the grey-shaded boxes, we address specialised, as well as general
(Corbin and Strauss, 2009; Corley and Gioia, 2004; Edmonson and scholars, and practitioners, because of the high relevance and clo-
McManus, 2007; Gioia et al., 2013; Miles and Huberman, 2007). This seness to the industry. Hence, we selected the International Journal
process followed suggestions by Gioia et al. (2013) to start with hor- of Production Economics, due to its distinct effort to connect man-
izontal or open coding. We derived first order topics by staying close to agement and engineering perspectives.
the original wording of each article's summarising phrase, before we VI. The last category of Cooper (1988) taxonomy represents the cov-
continued with vertical, respectively axial, coding to find higher order erage of the literature review and takes the responsibility for the
constructs embracing multiple first order themes (Corbin and Strauss, exhaustiveness of the review. Since we aim to condense a broad
2009; Gioia et al., 2013). After the third iteration of relating the range of literature within the field of smart factory into certain
emerging constructs to each other we observed saturation (Corbin and categories, we chose an exhaustive, but still selective coverage.
Strauss, 2009) and were able to connect all second order topics to the Selective, because we apply certain including criteria, such as the
afore mentioned eight aggregating artefacts. Two researchers con- quality of the journal and intentionally eliminated articles in the
ducted this process individually and the results were reviewed by a intersection of smart factory/smart manufacturing with supply
third to decrease bias and increase reliability and rigour (Berends and chain management (SCM), due the vast amount of SCM related
Johnston, 2009). publications. Nevertheless, even though we neglected SCM specific
A sound methodological foundation, thus to mention and follow terms in our literature search process, some articles that cover smart
scientifically acknowledged procedures in analysing literature or data manufacturing, or Industry 4.0, with a SCM lens were still found
in general is vital for achieving thorough results and a good theoretical and appear in the final list, due to their relevance. Besides that, a
contribution (Corbin and Strauss, 2009; Eisenhardt, 1989). Since we purely exhaustive, a representative, or a central/pivotal re-
just unfolded how we proceeded during our research in theory, we presentation of the literature is possible (Cooper, 1988).
present a structured overview onto the investigated literature in the
imminent chapter. 3.2. Topic conceptualisation

3. Structured presentation of literature The second phase of the literature review process, as it was in-
troduced by Vom Brocke et al. (2009), treats the conceptualisation of
3.1. Scope definition the topic. They hereby mean that a researcher needs to discover and
select the original search phrases commonly used in the field, in which
Following the recommendation of Vom Brocke et al. (2009), we the researcher is moving (Fischl et al., 2014). Selecting key words is a
used Cooper (1988) model to set the boundaries of our literature re- complicated process, since it has a strong impact on the completeness,
view. The highlighted boxes in Table 1 illustrate the categories we respectively on the quality of the review (Baker, 2000). Applying search
adopted. phrases leading to too many results can hinder the researcher to iden-
tify the important contributions, while search phrases setting too
I. The first category addresses the focus of the reviewed literature a narrow boundaries fear the risk of eliminating promising publications
researcher is interested in. In most cases, literature reviews assess out of the sample. In comparison to many other research streams, smart
the research outcomes, the research methods, the applied theories, factory and smart manufacturing emerged as research concepts only a
and/or the investigated applications (Bem, 1995; Torraco, 2005). few years ago (Lucke et al., 2008). Therefore, the delineation of the
With regard to our first research question, our focal points are to research field appeared to be relatively obvious. Additionally, the au-
gather and structure the research outcomes and the investigated thors already possess significant background knowledge in the treated
applications. This is because these categories promise to offer the thematic context giving the authors the possibility to build upon ex-
most relevant indications in what has already been researched and isting vocabulary. We thus proceeded in two steps. First, the focus lied
what has still to be done. on use cases in relation with smart factory, while we searched for other
II. The second category covers the goal of the literature review. Here, important constructs extracted from diverse studies in the second step.
Cooper (1988) differentiate between the endeavour of integrating, In the first case, we mapped expressions, such as smart factory and
or synthesising, the existing literature base, a critical examination smart manufacturing with more specific aspects, including application
of the relevant literature or the allocation of the major points of and use case, to account better for our research purpose. Our concept
interest within a research field. While a synthesis of the bulk of map (Rowley and Slack, 2004) is depicted in Fig. 2.
literature is the main objective for research question one, we want Besides the named expressions, we identified synonyms frequently
to examine the central issues in the second question. Therefore, used in closely related publications and sometimes applied inter-
these two classes depict our research goal and are highlighted in changeably. The digital factory, for instance, stands for the virtual re-
Table 1. presentation of a real factory to ease design and planning purposes
III. Cooper (1988) third category treats the organisation of the literature (Shariatzadeh et al., 2016). The term intelligent manufacturing (Burns,
review being historically, conceptually or methodologically. We 1997; Lu et al., 2017; Romero et al., 2017; Ruiz et al., 2011) or in-
investigated the historical evolution of the publication year of re- telligent factory (Boillot and Begin, 1993; Lee, 1996) account for more
levant articles and used methodological aspects to structure and information technology and software based perspectives on production
support our analysis. Nevertheless, the organisation of our review is systems. Here, the scholars primarily focus on software modelling and

4
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Keywords: Search algorithm

3a 4a
use case
1 2 example AND issue
smart factory application
digital industr*
intelligent AND production
AND
cloud manufactur* 3b
auto* progress*
cyber-physical development
evolution
advance*

Fig. 2. Search phrase mapping. Own illustration.

programming solutions, reflected by the character of the publishing Plattform Industrie 4.0, 2016). These reports are for instance the ori-
journals. Likewise, when talking about cloud manufacturing, the lit- ginal study introducing the term Industry 4.0 (acatech, 2013) and the
erature typically has a very technical nature (Ren et al., 2013; Ren publication describing the reference architecture model RAMI 4.0
et al., 2016; Yu et al., 2015), whereas approaches exist that try to set (Heidel et al., 2017). RAMI 4.0 represents a three-dimensional land-
this concept into a management environment (Cao et al., 2017; scape to structure and define the frame of Industry 4.0 with the smart
Klosowski, 2012). Further, automatic or autonomous production factory concept being a central element of the framework (Heidel et al.,
(Menascé et al., 2015) is another term describing the envisioned future 2017). These reports supported in finding appropriate terms needed to
state of a smart factory. However, compared to the other expressions, fully enlighten and grasp all relevant publications onto the topic of
such as cyber-physical systems (CPS), it delivered less relevant results. smart factory. This list of important expressions comprises inter alia
CPS have their origin in computer science. Yet, they are more and more manufacturing execution system (MES), enterprise resource planning
transferred to other subjects, since the central meaning illustrates a (ERP), administration cell and decentral production. While decentral
combined system, respectively an interaction, of physical products production and other similar terms describe a certain mode of produ-
(hardware) with digital products (software), leading to a recently in- cing goods, MES and ERP are rather process and software oriented
creased attention by a myriad of scholars (Babiceanu and Seker, 2016; elements within a manufacturing company. Fig. 4 illustrates the com-
Majstorovic et al., 2015; Monostori, 2014; Monostori et al., 2016; Wang bination of these expressions with smart factory in the search process to
et al., 2015). CPS stand for one of the key concepts within Industry 4.0 reach distinct contributions to our focal topic.
to achieve manufacturing systems connectivity (Jeschke et al., 2017; Hence, to the best of the author's knowledge, these sets of keywords
Wang et al., 2015). It combines the physical components with the vir- in combination with not setting time constraints (Fischl et al., 2014),
tual elements of the factory (Wang, Wan, Zhang, Di Li and Zhang, were sufficient to cover the biggest portion of published articles in the
2016b). Therefore, CPS integrate different technologies such as IoT, context of this literature review.
information and communication technology (ICT), cloud services, big
data, AI and machines on connectivity platforms to enable further au-
3.3. Literature search process
tomation and autonomy in production (Kagermann et al., 2013; Lasi
et al., 2014). CPS can be partly associated and rooted with the ideas of
With the sound preparation of relevant keywords in the previous
computer-integrated manufacturing (CIM) in the 90s. Compared to the
phase, the main activity of the third phase is to conduct the literature
concept of CIM, CPS goes beyond. CIM is associated with the integra-
search process in the imminent four steps, following the advices of Vom
tion of different computerized activities such as product design and
Brocke et al. (2009).
engineering, production planning and manufacturing processes, sales
First, we conducted a journal search. To account for the diverse
ordering, finance and human resource management (Hill, 1991).
perspectives the smart factory concept can be approached by, we in-
Therefore, different software systems and data communication streams
cluded journals from Production Management, Industrial and
need to be integrated and connected, such as for example CAD, CAQ
Manufacturing Engineering, and Information Systems. We further in-
and CAM to intelligently control the whole manufacturing cycle
tegrated journals out of lists from other authors in a related field (Fischl
(Gunasekaran et al., 1994). Although processes are controlled and de-
et al., 2014) and constrained the sample with regard to their quality.
cisions are supported by computers, the factory is not completely run-
The basic quality criteria were for articles to be peer-reviewed, i.e. to be
ning independently (Hitomi, 1991; Rowlinson et al., 1994). That is the
published in scientific journals and on conferences (Rowley and Slack,
major advancement and difference compared to CPS logic, where the
2004). In addition, we assessed their ranking according to the
intelligent connection and data usage are used to build autonomous and
VHB-JOURQUAL 3 Ranking and the Scimago impact factor. This pro-
self-controlling structures in manufacturing to achieve the next in-
cedure led to a final list of 124 journals reflecting the exhaustive and
dustrial age (Wang et al., 2016b).
selective character referring to Cooper (1988).
Besides the enumerated keywords, few others were discovered in
After listing the initial set of journals, we compared their occurrence
early phases of the research. For example Shaw (1983, p. 10) early
within the prevalent databases and selected the four major databases
examination of an “unmanned factory”, or the concept of “computer-
EBSCOhost, Emerald, ProQuest and Science Direct for the subsequent
integrated manufacturing”, as we introduced it above (Boaden and
keyword search. Another advantage of these databases is the con-
Dale, 1986, p. 30). Yet, these results were rare and still captured by the
tinuous access throughout the whole research process and their di-
search algorithm of the queried databases, making the task of opti-
mension.
mising the search phrases obsolete.
Following Vom Brocke et al.’s (2009) proposition in the third step,
In the latter case, we extracted a list of 13 constructs often men-
we performed the literature search process in a structured and trans-
tioned in relation with smart factory from a number of important
parent manner, as it is illustrated in Figs. 3 and 4. We iterated the
publications (acatech, 2013; Adolphs et al., 2015; Heidel et al., 2017;
search algorithm by going from left to right and up to down one

5
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Databases: Search Results

ProQuest Emerald EBSCOhost Science Direct

decentral production 0 0 0 1(0)
modular production 2(1) 0 0 21(5)
versatile production 0 0 0 0
adaptable production 0 0 0 5(0)
value networks 3(1) 0 0 19(4)
end-to-end engineering 1(0) 0 1(0) 4(2)
end-to-end digital integration AND smart factory 1(0) 0 2(1) 11(5)
enterprise resource planning 7(3) 3(1) 2(2) 76(12)
manufacturing execution system 8(3) 3(0) 1(0) 76(12)
embedded manufacturing system 0 0 3(0) 1(0)
networked manufacturing system 1(1) 1(1) 0 15(8)
cyber-physical system 8(2) 2(0) 1(1) 250(19)
administration cell 0 0 0 0
net 2 0 3 15
sub-total I 20
selection procedure
sub-total II 6
forward-backward search
Total 6

Fig. 3. Literature search process – Part 2. Own illustration.

database after the other. The representation of the results on the right example, leading to 56 hits, whereof eleven paper of higher interest
side in Figs. 3 and 4 can be understood as follows: “overall number of were identified. Here, it has to be noted that terms ending with an
results (amount of relevant and listed articles)”. This means, ex- asterisk (e.g. industr*) reflect all words that have the first part up to the
emplarily stated for the search process depicted by Fig. 3, we started asterisk in common (e.g. industry, industries, industrial, etc.). Ac-
with inquiring in ProQuest with Boolean search terms smart factory AND cording to Cooper (1988), we followed this exhaustive search process
use case, resulting in four hits, whereof one is of higher relevance and with rigour and applied keywords, such as use case and application, with
therefore registered. The second inquiry was smart factory AND regard to the focus of the literature review. We further documented all

1 2 3
Databases: Search Results
application
example
use case

ProQuest Emerald EBSCOhost Science Direct

1 2 3 1 2 3 1 2 3 1 2 3
factory 4(1) 56(11) 51(14) 0 9(2) 8(2) 0 0 0 25(5) 137(11) 156(18)
industr* 0 28(0) 23(0) 0 2(0) 1(0) 0 0 1(0) 1(0) 20(1) 21(1)
smart
production 0 24(5) 21(5) 0 8(1) 5(0) 0 0 1(0) 15(0) 54(5) 55(5)
manufactur* 0 68(8) 66(7) 0 17(4) 15(3) 5(0) 2(1) 6(2) 34(3) 212(10) 232(12)
factory 2(0) 45(4) 42(3) 1(0) 21(2) 20(2) 1(0) 2(0) 12(0) 42(1) 55(1) 61(1)
industr* 2(0) 93(3) 53(2) 0 8(0) 5(0) 0 2(0) 0 0 35(0) 23(0)
digital
production 1(1) 123(2) 75(2) 2(0) 69(1) 41(1) 8(0) 1(0) 2(0) 10(0) 24(0) 13(0)
manufactur* 2(0) 234(15) 233(11) 3(0) 73(5) 81(6) 0 13(2) 28(2) 9(1) 139(2) 105(3)
factory 0 10(1) 7(1) 0 4(3) 6(4) 1(0) 1(0) 37(1) 3(0) 20(0) 48(2)
industr* 1(0) 42(0) 29(0) 0 2(0) 1(0) 0 1(0) 3(1) 0 6(0) 6(0)
intelligent
production 0 67(3) 68(4) 0 21(1) 20(1) 2(0) 136(3) 2(0) 15(1) 31(2) 37(3)
manufactur* 37(2) 271(10) 283(12) 2(0) 86(4) 100(3) 1(0) 103(2) 147(5) 14(1) 54(2) 56(2)
AND
factory 0 1(0) 0 0 0 0 1(0) 1(0) 10(0) 0 3(0) 2(0)
industr* 0 24(0) 22(0) 0 1(0) 2(0) 0 0 1(1) 3(0) 25(0) 27(0)
cloud
production 0 4(1) 3(1) 0 3(1) 3(1) 3(0) 27(1) 64(5) 2(0) 4(1) 5(1)
manufactur* 2(0) 60(11) 63(10) 0 11(2) 10(2) 1(0) 2(1) 11(4) 19(1) 66(5) 71(8)
factory 1(0) 69(2) 45(2) 0 83(1) 79(0) 0 2(0) 1(0) 1(0) 42(0) 33(0)
industr* 0 42(3) 71(1) 0 12(0) 10(0) 0 14(0) 32(0) 1(0) 34(1) 32(1)
auto*
production 3(0) 154(2) 91(2) 0 138(0) 94(0) 0 23(0) 29(0) 18(1) 143(2) 120(3)
manufactur* 2(0) 160(2) 137(3) 3(0) 110(1) 83(0) 0 44(0) 81(1) 16(1) 68(2) 42(2)
factory 0 0 0 0 0 0 3(0) 42(0) 26(3) 0 0 0
cyber- industr* 0 0 0 0 0 0 0 3(0) 0 0 0 0
physical production 0 6(3) 7(2) 0 1(0) 1(0) 3(0) 50(0) 2(0) 17(2) 16(4) 18(3)
manufactur* 0 2(1) 1(1) 0 0 0 0 10(0) 18(0) 2(0) 14(1) 9(1)
net 66 14 21 37
sub-total I 138
selection procedure
sub-total II 80
forward - backward search
Total 100

Fig. 4. Literature search process – Part 1. Own illustration.

6
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Table 2
Overview on an excerpt of articles and their characteristics. Own illustration.
# Author Approach Management Impact Quantification Levela Data Relation BDA Method – Relation Data (Type) -
Integration Sourcesb Problem Problem
Focused Broad

1 Chongwatpol (2015) X X X 1 1|2 X X

2 Meziane et al. (2000) X 1|2 X
3 Rezaei et al. (2017) X X 1|3 1|2|3 X X
4 Zhong et al. (2014) X X 1 1|2 X
5 Krumeich et al. (2016) X 1|2 1
6 Lee et al. (2014) X 1 1 X X
7 Schuh et al. (2014) X X 1|2 1|2 X
8 Zheng & Wu (2017) X X 1|2 1|2
9 Arunachalam et al. (2017) X 1|2|3 X X
10 Tang et al. (2016) X 1 1|2 X
11 Bauer et al. (2016) X 1 1|2 X
12 Qin et al. (2016) X X 2|3
13 Wang, Zhang, Duan, & Gao X X 1|2 1|2
(2017a)
14 Kobbacy & Vadera (2011) X 1 1|2|3
15 Ilie-Zudor et al. (2015) X 1 1|2
16 Ruiz et al. (2011) X 1 1|2
17 Sanders & Gegov (2013) X 1 1|2|3
18 Wuest et al. (2014) X X 1 1|2
19 Malakooti & Raman (2000) X X 1 1|2

BDA = Big Data Analytics.

a
1 = operative; 2 = tactical; 3 = strategic.
b
1 = product/machine data; 2 = internal data; 3 = external data.

steps thoroughly, to increase replicability and validity (Tranfield et al., process of examining the specific set of articles. These findings are
2003; Vom Brocke et al., 2009). The results encompass in most cases all threefold. First, we applied a qualitative content analysis (Mayring,
years up to December 2017, where we performed the database search. 2000) leading to a distinct overview on the main characteristics of the
Yet, in few cases, the amount of hits went beyond a reasonable number publications. Second, we transferred the qualitative coding technique
and has therefore been filtered carefully. Typically, the timeframe was (Corbin and Strauss, 2009; Gioia et al., 2013) to our sample and iden-
set to all publications after the year 2007 and obvious not-fitting tified eight categories with a specific thematic context each. Third, we
journals were neglected. This seemed reasonable, since the constant created a research model out of the gathered insights to provide a
overlap with the analysis demonstrated that more than 80% of the tangible overview for future scientific purposes.
potentially promising articles were published in the last decade. In the subsequent paragraphs, we first describe the results of the
In the process of selecting relevant articles out of the list of results content analysis, before heading to the presentation of the eight cate-
after an inquiry, we screened the title, journal and abstract. This was gories and the research model in the second part.
done subjectively by intentionally profiting from the authors expertise
(Yin, 2009). Afterwards, we again applied including criteria on the
4.1. Qualitative content analysis
intermediate list of relevant articles by focusing on the impact factor of
the publishing journal (quality) and the accessibility. Concerning the
Following the explanations of Mayring (2000) on the procedure of
quality criteria, we set the threshold to an impact factor of 0.5 (Sci-
qualitative content analysis, we first defined the main aspects we are
mago), depicting the line of where about 20 percent of the articles were
interested in. These are the applied research method, the core finding,
excluded from the sample.
covered topics, subjective relevance, a summarising phrase and the
The use of certain excluding and including criteria is thereby in line
proposed future research questions of the article. To increase the
with recommendations and the design of other literature reviews
quality of our analysis and limit predisposition, we overlapped emer-
(Brereton et al., 2007; Crossan and Apaydin, 2010).
ging information with afore defined topics in the covered topic column
Following the imminent step of Vom Brocke et al.’s (2009) ap-
and hence revised and enlarged this list, if necessary (Eisenhardt, 1989;
proach, we used the previously selected sample of articles (80, re-
Eisenhardt and Graebner, 2007; Miles and Huberman, 2007). Ad-
spectively 6) as the basis for the forward and backward search. Thus,
ditionally, the subjective relevance served as an indicator during the
we investigated the reference list of the found articles and those arti-
analysis, if a specific artefact is thought of being particularly important.
cles, citing the object of interest. We received the amount of articles
Thus, we marked a number between zero (not at all important) and ten
citing the investigated artefact either from the according database, al-
(highly important) in our table. To help in structuring and visualising
ready delivered with the download of the article, or through the asso-
the emerging information, we prepared and filled tables (Eisenhardt
ciated function offered by Google Scholar. We identified 30 articles,
and Graebner, 2007). We continued with reading and evaluating the
whereof 20 are accessible and suffice the quality criteria. Finally, we
sample with regard to the defined categories. In the meantime, certain
reduced the initial set of 158 publications with respect to their quality,
patterns began to emerge. Finding patterns in the process of data ana-
accessibility and the goal, focus and coverage of the literature review
lysis is the ultimate target for creating theory and deriving key con-
(Cooper, 1988) to 106. These were analysed, synthesised and we pre-
clusions (Edmonson and McManus, 2007), which is why we strive to
sent the resulting insights subsequently.
outline these patterns in tabular format.
In coherence with other publications, we reveal the result table
4. Findings before elaborating on the insights that can be drawn from it (Breugst
et al., 2015). Therefore, Table 2 illustrates an overview on an excerpt of
While we presented the foundation of our analysis in the previous 19 particularly suitable articles to demonstrate the key findings of our
section, we now want to introduce the results we gathered in the analysis (Sutton and Callahan, 1987).

7
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

We observed early that a majority of articles treat idiosyncratic use identify competencies a company needs to have to understand and
cases. Only few aimed at delivering a broader approach onto aspects of apply big data analytics. Along the above-mentioned causal chain, these
the smart factory and Industry 4.0, neglecting literature reviews in this are data generation, data processing, data analysing and data visua-
sense. These single use cases are commonly based on specific problems lising competencies.
of industrial companies. For instance, Chongwatpol (2015) applies data Thus, we investigated these characteristic within each article of our
analysis techniques, such as data mining, on bundles of parts to unfold set and poured the information into an enlarged version of Table 2. This
root causes of failures and thus decrease failure rates and increase the approach gave us the possibility to identify patterns and correlations
manufacturing performance and quality. He thereby base his research between these characteristics. It can be stated that the major part of the
on the case of a company in Thailand producing optical receiver- investigated literature treat single use cases with low generalizability.
transceiver devices. Zhong et al. (2014) use shopfloor data derived from They most often rely upon machine data, being sometimes supported by
radio frequency identification devices (RFID) of a company called internal data of the case company. The nature of the publications is
Huaiji Dengyun Auto-parts (Holding) Co., Ltd to determine actual rather technical, on an operative level, not related to the management
standard operating times (SOT). During their research, they develop an system of the enterprise and consequently no monetary impact quan-
overview of SOTs for typical operations by using similar to tification is executed in most cases. Further, the causal chain between
Chongwatpol (2015) data analysis techniques, including data mining problem and solution is typically only partly displayed. Ultimately, we
approaches. A third example is the approach of Wuest et al. (2014), extracted conclusions and propositions serving as reference points for
who intend to offer a possibility to improve production quality by the development of a research agenda that is depicted in the fifth sec-
gathering data about the produced good over the complete manu- tion.
facturing process and analysing this data by combining AI tools.
Obviously all of the treated cases have in common that they use
4.2. Qualitative coding approach
manufacturing data to improve quality, increase efficiency or develop a
framework for example. While examining our sample, we noticed in
In the process of analysing the content of the literature with the help
most issues the researcher solved the problem by using data being close
of the predefined tables, as we outlined in the previous section, we
to the original source. This may seem evident, yet, including external or
found no dominant perspective the authors are adopting. We realised
environmental data into the data basis, beside product/machine and
no significant thematic context being prevalent compared to the other
internal data of the manufacturing firm could lead to a more general
treated topics. Table 3 underlines this fact.
perspective. Thus, in our set of articles we see a strong correlation of
The six journals with the most articles in our sample only account
research outcomes that have a broader applicability and the use of
for about half of the total set of articles. The residual articles are pub-
external data, despite not being limited to it (Arunachalam et al., 2017).
lished in another 37 different journals.
For example Rezaei et al. (2017) relate internal as well as external data
However, the articles have in common that over 80 percent of the
to the target of quantifying, assessing and increasing supply chain
artefacts in our sample were written in the last decade. In addition,
performance measurement, while all stakeholders can access the gen-
Fig. 5 highlights a clear trend and the emergence of a plurality of ar-
erated data. Being in line with the insights we discovered before, the
ticles after the publication of the Industry 4.0 working group's final
majority of core findings play on an operative level, very often with
report in 2013 (acatech, 2013).
regard to the shopfloor in industrial companies. Sometimes they are
{{}}The blue columns illustrate the initially found and selected bulk
enhanced to a tactical horizon, but seldom rose to a long-term strategic
of literature, whereas the grey columns are those papers originating
view. Again, Rezaei et al. (2017) explain that by using different data
from the backward and forward search. It is not surprising that espe-
sources, they were able to develop a framework that supports in solving
cially in the last four years, many scholars dedicated their work on
high-level strategic decisions to improve supply chain operations. Yet,
Industry 4.0, smart factory, smart manufacturing and related topics.
the same framework is also able to cope with low-level operative de-
Yet, it is fascinating to see that already in 1983, Shaw envisioned a
cision through profiting from machine intelligence. In contrast, Bauer
future state of a completely autonomous factory by relying on advanced
et al. (2016) propose using smart wearables for data acquisition by blue
robotics. Further, Meziane et al. (2000) by this time achieved to break
collars in the production line, stating a rather operative horizon of the
down the advantages and disadvantages of AI methodologies with re-
research outcome.
spect to their field of application within manufacturing operations.
When we related the decision level with the question, whether or
Their seminal article describes for instance, why neural networks are
not an author provides statements on the integration of, or the possi-
advantageous in solving process-planning problems and what examples
bility to integrate the findings into a management framework, we found
exist for the use of genetic algorithms for scheduling tasks. While these
a significant correlation between these two characteristics. Thus, the
are two cases that are not explicitly linked to a smart factory, especially
probability of an author connecting her or his findings with a higher
the latter one describes tools, seen as tremendous to reach the state of a
order management framework increases with the decisional level the
fully connected and interacting production system. These focal points of
researcher is coping with. One interesting exception is the research
the authors demonstrate again the heterogeneity of the bulk of litera-
outcome of Chongwatpol (2015) we introduced before. He uses ma-
ture related to the smart factory. Thus, this anomaly gave us the im-
chine data to find causes of failures and can predict possible future
petus to cluster the identified literature into specific thematic dimen-
failures by identifying thresholds in the data at hand. The results, such
sions. We therefore used a qualitative coding procedure inherited from
as less scrap rate and decrease of cost, are rather operative. Never-
theless, Chongwatpol (2015) elaborate on the consequences that can
Table 3
follow by evaluating the reasons for the appearance of the defect causes Amount of Articles per Journal. Own illustration.
and implementing actions to increase the overall production system
performance in long-term. Additionally, the author underpins his Journal Amount of Articles

findings with a distinct impact quantification in contrast to most other Journal of Intelligent Manufacturing 18
scholars in our sample. Procedia CIRP 14
Slightly apart from the previous conclusions, we found that barely Industrial Management & Data Systems 8
any researcher clearly illustrates a complete causal chain, starting from International Journal of Production Research 7
Assembly Automation 5
the examined problem, to the type of data, which and why it was used,
International Journal of Production Economics 3
the reasons for or against the applied (big) data analysis method and Total 55
the final solution. Though, for instance, Arunachalam et al. (2017)

8
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

20
18
16
14
12
10
8
6
4
2
0
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
Initial Set of Articles FW-BW Search Articles

Fig. 5. Number of Publications per Year. Own illustration.

Fig. 6. Qualitative coding approach (Corbin and Strauss, 2009; Gioia et al., 2013; Rooks, 2000; Saygin and Sarangapani, 2006).

qualitative research, in accordance with Gioia et al. (2013) and Corbin Besides them, Malakooti and Zhou (1994) propose the application of
and Strauss (2009), mainly used on interview data to streamline our set neural networks to decision-making problems under certainty with
of articles. Fig. 6 illustrates an excerpt of the approach, explained in multiple variables, such as production scheduling purposes, for deriving
Chapter 2.2.3 above, to enhance the replicability, transparency (Corley the best option. This section represents an area with a small share of
and Gioia, 2004; Pratt, 2007; Tracy, 2010) and credibility onto our articles we were able to address to, most probably because AI and its
material (Lincoln and Guba, 1985). tools were not represented in the keywords we applied during our lit-
While conducting the horizontal and vertical coding, we experi- erature search. This is yet suitable, since on the one hand AI depicts a
enced saturation after the third iteration and identified the following self-standing research stream, but on the other hand has a significant
eight categories with their respective occurrence, shown in Fig. 7. It has correlation to the smart factory and should therefore not be neglected
to be noted that despite the fact most articles were easy to position as a potential research area in this sense.
within the categories, some do comprise focal points or core findings
that lie in the intersection between two or more groups. In this case we
judged subjectively the higher coherence with one category or another. 4.2.1. Data Handling
Furthermore, we have to state that since all articles are selected due to Part of the second group are articles with their focus on data
their correlation to the smart factory construct, they share core char- handling procedures in the broader field of (intelligent) manufacturing
acteristics, such as the use of data in various ways, leading to the fact systems. Data handling stands hereby for all partial aspects, including
that the perspectives may blur between each other. data generation, acquisition, mining and analysis, not explicitly men-
The first group encompass articles that explicitly mention the use of tioning AI, SCM or IoT. The second order topics are dominated by data-
AI methodologies within the sphere of manufacturing systems. Beside based production process optimisation, data analysis possibilities
the work of Meziane et al. (2000) we already outlined above, Munguía within manufacturing and big data related subjects. For example,
et al. (2009), for instance, describe a possible application for AI tools to Segura Velandia et al. (2016) elaborate on a case study where RFID
help an engineer in deciding upon ideal product and production vari- elements are used to monitor and improve production and assembly
ables by referring to overall requirements during the design phase. procedures. This is affiliated to Zhong et al. (2016) proposition for the
visualization of shopfloor logistics data retrieved from RFID devices to

9
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Fig. 7. The eight coded Categories and their Share of the Articles. Own illustration.

improve decision-making. Going in the same direction of the idea, to thematic cell with the fewest artefacts in our set. Nevertheless, one
use new sensor data from the shopfloor and data analysing capabilities exponent is Malakooti (2011, 2012), who proposes decision-making
to provide prioritised action items for the production manager in case of procedures for multi-variable problems by following distinct phases.
the need for quick responses by Schuh et al. (2014), or to use this data This is similar to Kobbacy et al., (1995) introduction of a system in-
for monitoring energy efficiency (Tan et al., 2017). Concerning data tended to support decision-making problems intelligently with the
analysis possibilities, Hsu (2014), for instance, applies data analysis focus on maintaining complex industrial devices.
methods to forecast the performance of manufacturing systems and
derive needed actions for anticipating performance decline by allo- 4.2.4. IoT
cating according resources in advance. Further, LaValle et al. (2011) The fifth group is a special category devoted to artefacts explicitly
describe their findings of a case study where they investigated how treating IoT in the context of manufacturing and the smart factory
pioneer organisations make use of data analysis methods to transform construct. The articles range from a case study description of how
the raw data into statements on which it is possible to base decisions for General Electric is coping with the possibilities of IoT (Winnig, 2016) to
future actions. Finally, a representative of the third group is O'Donovan the assessment of IoT tools for creating value within a manufacturing
et al. (2015b), who explains an architecture for a big data application system that can have a broader impact on the way of doing business by
intended for manufacturing analytics in large scale production systems. the organisation (Rymaszewska et al., 2017). In addition, Lee and Lee
Processing data, either in the acquisition or in the analysis, with the (2015) elaborate on IoT with respect to new products and differentiates
idea to connect, automate and optimise production is the central aspect between three areas, where a production company may use IoT appli-
of a smart factory. Therefore, it is obvious that this category possesses cations, including monitoring, cooperation and analytics. Another
the largest share of articles in our set. contribution is the proposition of an architecture for measuring devia-
tions from the ideal production process in combination with data-based
4.2.2. Supply Chain Management aids to increase the overall performance (Zhang et al., 2014; Zhang
The third category is a descendent of the SCM research stream and et al., 2014). Similar to the first category, IoT is related to the smart
summarises, as already mentioned above, publications closely related factory and Industry 4.0 leading to overlaps between both streams. For
to characteristics of the smart factory and to the other seven groups, this reason, IoT was not explicitly listed in the set of keywords, yet,
while explicitly naming SCM. Here, assessing the potential, impact and implicitly derived out of the results and we therefore accumulated the
describing the use of cloud computing is a recurring phenomenon (Cao according artefacts in group number five. These obviously present only
et al., 2017; Durowoju et al., 2011; Gowda and Subramanya, 2015; Jede a very small share of publications on IoT, but constitute a position to
and Teuteberg, 2016). Next to it strive Gunasekaran et al. (2017) to demonstrate the inability to differentiate clearly between aspects of
describe the effect of management dedication on the assimilation of big Industry 4.0 and the heterogeneity of aspects within the smart factory
data analytics and the consecutive impact of predictive analytics on the construct.
supply chain and overall company performance. Related to it, Taticchi
et al. (2015) present a survey on tools that support decision-making and 4.2.5. Digital Transformation
measuring performance. The sixth category embraces another rather small set of articles,
treating the transformation of industrial products or production systems
4.2.3. Decision-making to intelligent and connected systems, including the topic of CPS. Bloss
Decision-making problems and solutions are already discussed in (2005) and Huang et al. (2008), for example, describe the smart device
depth. They also appear in other categories than this one. Nevertheless, transformation by implementing sensors, actuators, or RFID chips. Be-
we chose to dedicate one small group on more general decision-making sides them, Monostori (2014) explains the transition of regular manu-
aspects while staying close to the topic of the paper at hand, due to their facturing sites to CPS and the high promises descending from them.
importance and the perpetual interest of scholars. Therefore, the arti- While the articles allocated in this group do not provide an exhaustive
cles in this group differ by not setting the focus on data usage and ex- overview in our perception, the publication dates suppose that scholars
ploitation to increase decision-making performance, leading to the concentrated on this field in the late 20th and early 21st century and

10
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

now moved on to trending topics, such as IoT and Industry 4.0. Fur- as well as detailed in the following.
thermore, we have to note that literature coping with CPS are some- Decision-making encompasses research activities around data based
what underrepresented in our results. The concept of CPS can be reg- decision-making in manufacturing using different technologies, in-
ularly crossed in the sphere of the smart factory and Industry 4.0, which cluding visualization techniques, machine learning and AI. All kind of
is why we dedicated a future research avenue in Chapter 4.3 below. decisions in manufacturing, for instance design, scheduling, process
planning and control are part of this research stream.
4.2.6. IT Infrastructure The second research field CPS is concerned with developing con-
Within the second largest accumulation of articles, the main simi- cepts, models for assistant systems for operators, self-steering manu-
larity lies in the nature of the core findings, or key outcomes. These all facturing systems and CPS. This research stream has been included into
have in common that the respective author designs, models or presents the Digital Transformation category before, but due to its under-
a software application. For instance, Chen and Muraki (1997) propose a representation in the results and its importance in the realm of the
model to improve online planning and controlling systems by selecting smart factory, we dedicated one pillar to it. This avenue focuses on the
the optimal parameters. Azadegan et al. (2011), in turn, list use cases of development of theories, concepts and models for semi-autonomous
applying fuzzy logic elements on a variety of production problems, (interaction between physical and digital world) up to autonomous
while Yeh (1995) developed a software tool to condense important factories. It is an interdisciplinary research stream containing elements
production data and improve planning and execution of job items. of AI, IoT, decision-making and data handling elements. Community
Additionally, Otto et al. (2014) investigate self-programming software building and cross-functional research should be fostered in this re-
intended to serve as the basis for an autonomous plug-and-play man- search field to enhance self-steering systems and CPS research.
ufacturing system. Here, the authors typically aim to automate certain The third pillar, data handling, is built on research activities deriving
tasks by transferring the needed work in software models (Shariatzadeh models and theories on how to exploit the potential of data with a focus
et al., 2016) and optimise production processes through the intelligent on data generation, acquisition, mining and analysis. Scholars in this
combination of generated data and software (Candido, 1998). They field contribute to the smart factory with the objective to provide data
tend to use software to describe a bigger picture of digital manu- models towards a single source of truth and intelligent data exchange
facturing (Nylund and Andersson, 2011) or a smart factory architecture models.
likewise (Tang et al., 2016). Research field IT infrastructure represents in our view the basis for
all smart factory research areas focusing on the greater software and
4.2.7. Theoretical Contributions hardware concepts of company networks. It is a technology-focused
Group number eight summarises theoretical contributions on prin- discussion on how to organize data generation and exchange between
cipal topics, including IoT, big data and Industry 4.0, in turn containing different layers in the company. It is about how to organize the hor-
smart factory and smart manufacturing related artefacts. This can be a izontal and vertical data integration and steering in a manufacturing
set of definitions of CIM (Boaden and Dale, 1986), or a more general environment. Cloud technologies, blockchain and traditional informa-
view on Industry 4.0 and CPS (Lasi et al., 2014; Lu, 2017; Stock and tion system technologies such as ERP and MES are part of this research
Seliger, 2016; Thoben et al., 2017). Moreover, Magruk (2016) presents stream. Thus, many interfaces to the data handling research field exist,
potential threats and opportunities regarding Industry 4.0 similar to a requiring an interdisciplinary research approach.
SWOT analysis. Besides them, Holmström and Partanen (2014) in- Digital transformation copes with the transformational path of fac-
troduce a conceptual view on the combination of smart manufacturing tories towards smart factories. It treats how companies should for-
with other domains, such as logistics and equipment. Finally, Kang mulate visions, define suitable future states, and organize transforma-
et al. (2016) describe the five prevalent core technologies in the field of tional changes within the factory. This part rather concentrates on the
smart manufacturing, being after them CPS, cloud computing, IoT, big social science research field of a transformation towards smart factories
data analytics and intelligent sensors. by including and focusing on the human perspective that comes along
The evolved perspectives may seem rather similar, though it is with this revolution.
possible to see group number two and eight as the basis of the literature The sixth pillar, human machine interaction encompasses research
related to smart factory. These describe the core concepts, views, focal activities creating solutions for the co-automation, physical and digital
points and thematic boundaries, while the residual perspectives ex- assistant systems. Beside technological developments, the human per-
plicitly relate this foundation with specific peculiarities or research spective and role in autonomous smart factories is central within this
streams, namely SCM, IoT, AI, IT infrastructure, decision-making and stream. This pillar is solidly connected to decision-making and CPS,
digitalisation. wherefore we recommend cross-field research as well.
By setting the boundaries of our systematic literature review, we The research field of IoT accounts for the connectivity of elements
focus on use cases or applications within the field of smart factory and sensor technologies, Wang et al. (2016a, p. 6) describe as “com-
leading to the eight coded categories presented above. However, it prehensive connections”. To enable self-steering systems in intra lo-
became apparent that few more streams exist, which should not be gistics or in manufacturing processes it is mandatory to have system
neglected, rather mentioned to derive a complete picture of the smart elements in place providing the relevant information to connect it with
factory research context. Hence, we propose a smart factory research intelligent decision algorithms, automatically deriving the right actions
model in the subsequent paragraph. for the actors. Traceability and transparency (real-time data informa-
tion) of system elements and states is the main research contribution by
4.3. The smart factory research model this research stream.
Cloud manufacturing and services, the last pillar of our research
Based on the results of our systematic literature review and en- model, is split into a technology and business stream. The technology
hanced with the authors’ experiences, we identified research activities, stream discusses cloud manufacturing architectures models and the-
core findings and intersections related to smart manufacturing and ories. The business stream highlights the business development per-
conceptually derived a smart factory research model, potentially ser- spective for smart factories. For instance, new operating models are
ving as a basis for smart factory research activities. Our model consists possible, enabled by the digitalisation of factory capabilities. Moreover,
of eight main pillars that address criteria and elements such as adapt- cloud manufacturing services such as design, production or quality
ability, predictability, decision-support, physical and digital assistants, management services can be exploited with increasing process auto-
co-automation and learning. The pillars are visualised and defined to- mation and digitalisation of implicit knowledge.
gether with specific research targets and existing key concepts in Fig. 8, Linking the eight proposed research fields back to the conceptual

11
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Fig. 8. The Smart Factory Research Model. Own illustration (Košturiak and Gregor, 1999; Lee, 2015; Qu et al., 2012; Wu et al., 2013; Xu, 2012; Zancul et al., 2016;
Zhang et al., 2016).

setup of a smart factory, we are able to state two things. First, the four Chapter 4.3. We hereby strived to connect and aggregate similar sug-
layers of the smart factory concept are not translated one-to-one to the gestions, while it has to be noted that many articles did not offer an
research fields in our model. Second, the eight pillars labelled with outlook on future research opportunities. The research gaps mirror the
rather trending headings do take the four layers as a boundary condi- heterogeneity of the publications, while certain clusters emerge. For
tion and do not contradict them. instance, some authors request the identification and exploration of
Summarising chapter four, we presented a more general analysis more use cases for the application of software models within production
and insights on the nature and characteristics of the selected literature systems (Azadegan et al., 2011; Desforges et al., 2004;
for our review in the first part. The second half highlighted the het- Phaithoonbuathong et al., 2010; Rudtsch et al., 2014). This is related to
erogeneity of the treated topics related to the broader term smart factory the suggestion of enlarging the knowledge on and exploring use case of
and ended with the presentation of our proposed research model. The digital applications throughout the smart factory (Huang et al., 2008;
consequent enumeration and description of a choice of articles under- Monostori, 2014; Syberfeldt et al., 2016; Wu et al., 2016). To name a
lines the characteristics (single use cases, operational/tactical level, last example, certain scholars promote a deeper involvement with data
technical nature, mainly uses machine data) discovered before. mining approaches in manufacturing (Chongwatpol, 2015; Choudhary
The aim is to achieve a neutral representation of the sample of lit- et al., 2009).
erature for scholars and practitioners (Cooper, 1988), therefore we
discuss the implications in the last. 5.2. Other avenues for research

5. Conclusion Beside the research suggestions derived from the set of literature of
the review at hand, we discovered further research avenues as a logical
During the last four chapters, we elaborated on our impetus for consequence of the findings in chapter four. We highlighted that most
conducting a systematic literature review on the smart factory, as a key articles focus on single case studies. Especially research on job sche-
construct of Industry 4.0 (Lucke et al., 2008; Stock and Seliger, 2016), duling, planning procedures, general optimisation problems of pro-
by underlining the expressed need for such a structured representation duction systems and software applications for this purpose, is wide-
of the relevant literature (O'Donovan et al., 2015a; Thoben et al., 2017). spread. Nevertheless, we see that there still exist niches that need to be
Thus, we outlined the applied research methodology in detail, before enlighten in future research. These niches include, for example, busi-
presenting our findings, emerged from the qualitative content analysis ness processes occurring in the smart factory and prescriptive main-
and the qualitative coding approach. We hereby derived contributions tenance for involved machines, products and residual equipment.
to the research community as well as for practitioners, we want to in- Additionally, there is a paucity of research adopting a more general
troduce in the following. Afterwards, we summarise proposed research view, describing the complete causal chain of smart factory in detail and
avenues by the authors in our sample, stating the last phase of the evaluating the business or monetary impact of the implementation of
approach by Vom Brocke et al. (2009) and end with according im- Industry 4.0 technologies.
plications and limitations of our literature review.
5.3. Academic and practical contributions
5.1. Research agenda
Our academic contributions are threefold. First, according to the
Table 4 comprises the proposed research gaps of the majority of ultimate aim of a literature review, we achieved to condense the ex-
scholars in our set of literature, related to the eight research fields in isting knowledge on a vast variety of aspects referring to the concept of

12
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Table 4
Research Agenda. Own illustration.
Author Category Aggregated Future Research Opportunities

Gunasekaran et al. (2017) I Enriching the knowledge of big data analytics

Schuh et al. (2014) I Investigation of data based performance forecasting
Celik et al. (2012) I Optimisation and further development of solving real world decision problems
Kobbacy et al. (1995) I
Malakooti (2012) I
Malakooti & Zhou (1994) I, II Further improvement and incorporation of multiple modules into AI aided manufacturing tools
Meziane et al. (2000) I, II
Munguía et al. (2009) I, II
Lachenmaier et al. (2017) II Further extension of simulation models in the sphere of cyber-physical systems
Arunachalam et al. (2017) III Investigating data generation/collection approaches and the integration of multiple forms of data
Bauer et al. (2016) III
Zhong et al. (2016) III Investigation of data based software tools for the organisation of knowledge in production systems
Wang et al., 2017 III
Chongwatpol (2015) III Enriching the knowledge with regard to data mining approaches in manufacturing
Choudhary et al. (2009) III
Zhong et al. (2014) III
Zhong et al. (2015) III
Azadegan et al. (2011) IV Investigation of a plurality of use cases for software tools in manufacturing systems
Desforges et al. (2004) IV
Phaithoonbuathong et al. (2010) IV
Rudtsch et al. (2014) IV
Chen & Muraki (1997) IV Further improvement and development of (existing) software tools in manufacturing
Labib & Yuniarto (2005) IV
Shariatzadeh et al. (2016) IV Investigation of data security issues within manufacturing
Kang et al. (2016) VIII Investigation of application strategies for key technologies within (smart) manufacturing
Syberfeldt et al. (2016) I, II, V, VI, VII Enriching the knowledge on and investigation of the use of digital applications throughout the smart factory
Wu et al. (2016)
Huang et al. (2008)
Monostori (2014)
Lee et al. (2015) VII Investigation of different perspectives on IoT within manufacturing, including behavioural, economic and managerial aspects
Liu et al. (2017) VII
Rymaszewska et al. (2017) VII
Mishra et al. (2016) VII
Thoben et al. (2017) VII
Caputo et al. (2016) VII Investigation of variables that influence the adoption of IoT applications in manufacturing
Ren et al. (2013) VIII Investigation of future cloud manufacturing applications and business models for cloud manufacturing
Cao et al. (2017) VIII Investigation of the operational impact and benefits of cloud computing
Durowoju et al. (2011) VIII
Gowda & Subramanya, 2015 VIII
Jede and Teuteberg, 2016 VIII

smart factory in a systematic, transparent and thorough manner. By 5.4. Limitations

adopting a structured literature review approach (cf. Vom Brocke et al.,
2009), we were able to unfold the main artefacts in the treated thematic In addition to the implications, there exist limitations that accom-
context and presented the approach and findings in a neutral, unbiased pany a systematic literature review. Primarily, the literature search
way. Second, we synthesised the findings, structured them and visua- process is based upon the definition of a set of keywords. Despite the
lised an excerpt in tabular format (cf. Table 2). This served to demon- intended exhaustive character of the review (Cooper, 1988), it is pos-
strate the derived main characteristics of the artefacts and as a basis for sible we did not determine and incorporate all keywords required to
showing the inter-correlations in narrative style. We found that most find the relevant publications hereafter. Thus, we cannot guarantee
publications under the roof of smart factory are single use cases with completeness, while we believe we were able to integrate the majority
low generalizability, often rely upon machine data, are typically very of the vital artefacts. Moreover, the literature selection process is sub-
technical and seldom incorporate an impact estimation. Third, we ap- jected to bias and a subjective view of the researcher. We strived to
plied a meticulously conducted qualitative coding procedure, leading to limit predisposition by adopting a best possible objective view in the
the definition of eight distinct categories followed by a transfer into selection process by referring to neutral criteria and by implementing
eight distinct research fields. These serve as a frame for thematically redundancy.
similar articles and aid in setting a certain perspective by scholars up- To conclude, good systematic literature reviews are an important
front. The idea is to help authors and readers to better differentiate the mean to advance research and in setting new baselines (Denyer and
adopted point-of-view. Especially, when following recommendations Tranfield, 2009; Rowley and Slack, 2004; Tranfield et al., 2003).
for future research opportunities described below. Therefore, we hope to inspire scholars to build their consecutive re-
Furthermore, our findings may be applicable for practitioners. search on herein treated topics and to use the derived categories to
Principally, managers within manufacturing companies are advised to frame and set their focal points.
start with selected use cases when striving to make their production
system “smart”. Consequently, the results of this paper serve as an input
to those practitioners searching for dominant use cases in the academic Conflicts of interest
discourse, in turn, leading to purposeful starting points of the smart
factory journey. None.

13
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Acknowledgements research synthesis. Organ. Stud. 29 (3), 393–413.

Desforges, X., Habbadi, A., Geneste, L., Soler, F., 2004. Distributed machining control and
monitoring using smart sensors/actuators. J. Intell. Manuf. 15 (1), 39–53.
The authors want to thank the editors and the anonymous reviewers Durowoju, O.A., Chan, H.K., Wang, X., 2011. The impact of security and scalability of
for the precious advices and input on important aspects of the smart cloud service on supply chain performance. J. Electron. Commer. Res. 12 (4),
factory construct. Furthermore, we acknowledge the support of the 243–256.
Edmonson, A.C., McManus, S.E., 2007. Methodological fit in management field research.
Swiss Commission for Technology and Innovation under the research Acad. Manag. Rev. 32 (4), 1155–1179.
grant number 27399.1 PFES-ES. Eisenhardt, K.M., 1989. Building theories from case study research. Acad. Manag. Rev. 14
(4), 532–550.
Eisenhardt, K.M., Graebner, M.E., 2007. Theory building from cases: opportunities and
References challenges. Acad. Manag. J. 50 (1), 25–32.
Fischl, M., Scherrer-Rathje, M., Friedli, T., 2014. Digging deeper into supply risk: a sys-
acatech, 2013. Recommendations for Implementing the Strategic Initiative Industrie 4.0: tematic literature review on price risks. Supply Chain Manag.: Int. J. 19 (5/6),
Final Report of the Industrie 4.0 Working Group. Retrieved September 08, 2017, 480–503.
from. http://www.acatech.de/fileadmin/user_upload/Baumstruktur_nach_Website/ Gioia, D.A., Corley, K.G., Hamilton, A.L., 2013. Seeking qualitative rigor in inductive
Acatech/root/de/Material_fuer_Sonderseiten/Industrie_4.0/Final_report__Industrie_4. research: notes on the Gioia methodology. Organ. Res. Methods 16 (1), 15–31.
0_accessible.pdf. Gowda, A.B., Subramanya, K.N., 2015. The influence of variables on designing a cloud
Adolphs, P., Bedenbender, H., Dirzus, D., Ehlich, M., Epple, U., Hankel, M., et al., 2015. supply chain network: a factor Analysis approach. IUPC J. Supply Chain Manag. 12
Statusreport: Referenzarchitekturmodell Industrie 4.0 (RAMI4.0). Retrieved January (3), 35–49.
17, 2018, from. https://www.zvei.org/fileadmin/user_upload/Themen/Industrie_4. Gunasekaran, A., Virtanen, I., Martikainen, T., Yli-Olli, P., 1994. The design of computer-
0/Das_Referenzarchitekturmodell_RAMI_4.0_und_die_Industrie_4.0-Komponente/ integrated manufacturing systems. Int. J. Prod. Econ. 34 (3), 313–327. from. http://
pdf/Statusreport-Referenzmodelle-2015-v10.pdf. www.sciencedirect.com/science/article/pii/0925527394901619.
Almada-Lobo, F., 2015. The Industry 4.0 revolution and the future of manufacturing Gunasekaran, A., Papadopoulos, T., Dubey, R., Wamba, S.F., Childe, S.J., Hazen, B.,
execution systems (MES). J. Innov. Manag. 3 (4), 16–21. Akter, S., 2017. Big data and predictive analytics for supply chain and organizational
Arunachalam, D., Kumar, N., Kawalek, J.P., 2017. Understanding big data analytics performance. J. Bus. Res. 70, 308–317.
capabilities in supply chain management: unravelling the issues, challenges and Hammersley, M., 2001. On 'systematic' reviews of research literatures: a 'narrative' re-
implications for practice. In: Transportation Research Part E: Logistics and sponse to Evans & Benefield. Br. Educ. Res. J. 27 (5), 543–554.
Transportation Review. Heidel, R., Hoffmeister, M., Hankel, M., Döbrich, U., 2017. Industrie4.0 Basiswissen
Azadegan, A., Porobic, L., Ghazinoory, S., Samouei, P., Saman Kheirkhah, A., 2011. Fuzzy RAMI4.0: Referenzarchitekturmodell mit Industrie4.0-Komponente (1. Auflage).
logic in manufacturing: a review of literature and a specialized application. Int. J. Berlin, Berlin, Wien, Zürich: VDE Verlag GmbH; Beuth Verlag GmbH.
Prod. Econ. 132 (2), 258–270. Hill, T., 1991. Production/operations Management: Text and Cases, 2. ed. Prentice Hall,
Babiceanu, R.F., Seker, R., 2016. Big Data and virtualization for manufacturing cyber- New York.
physical systems: a survey of the current status and future outlook. Comput. Ind. 81, Hitomi, K., 1991. Strategic integrated manufacturing systems: the concept and structures.
128–137. Int. J. Prod. Econ. 25 (1), 5–12. from. http://www.sciencedirect.com/science/
Baker, M.J., 2000. Writing a literature review. Mark. Rev. 1 (2), 219–247. article/pii/092552739190125D.
Bauer, D., Wutzke, R., Bauernhansl, T., 2016. Wear@Work – a new approach for data Hofmann, E., Rüsch, M., 2017. Industry 4.0 and the current status as well as future
acquisition using wearables. Procedia CIRP 50, 529–534. prospects on logistics. Comput. Ind. 89, 23–34.
Bem, D.J., 1995. Writing a review article for psychological bulletin. Psychol. Bull. 118 Holmström, J., Partanen, J., 2014. Digital manufacturing-driven transformations of ser-
(2), 172–177. vice supply chains for complex products. Supply Chain Manag. 19 (4), 421–430.
Berends, L., Johnston, J., 2009. Using multiple coders to enhance qualitative analysis: the Hsu, C.-y., 2014. Integrated data envelopment analysis and neural network model for
case of interviews with consumers of drug treatment. Addict. Res. Theor. 13 (4), forecasting performance of wafer fabrication operations. J. Intell. Manuf. 25 (5),
373–381. 945–960.
Bloss, R., 2005. Pallet conveyor now smarter and slimmer. Assemb. Autom. 25 (2), Huang, G.Q., Zhang, Y.F., Chen, X., Newman, S.T., 2008. RFID-enabled real-time wireless
106–107. manufacturing for adaptive assembly planning and control. J. Intell. Manuf. 19 (6),
Boaden, R.J., Dale, B.G., 1986. What is computer-integrated manufacturing? Int. J. Oper. 701–713.
Prod. Manag. 6 (3), 30–37. Ilie-Zudor, E., Ekárt, A., Kemeny, Z., Buckingham, C., Welch, P., Monostori, L., 2015.
Boillot, J.P., Begin, G., 1993. Intelligent factory automation with machine vision systems. Advanced predictive-analysis-based decision support for collaborative logistics net-
Ind. Robot: Int. J. 20 (6), 29–32. works. Supply Chain Manag. 20 (4), 369–388.
Brereton, P., Kitchenham, B.A., Budgen, D., Turner, M., Khalil, M., 2007. Lessons from Jede, A., Teuteberg, F., 2016. Towards cloud-based supply chain processes. Int. J. Logist.
applying the systematic literature review process within the software engineering Manag. 27 (2), 438–462.
domain. J. Syst. Softw. 80 (4), 571–583. Jeschke, S., Brecher, C., Meisen, T., Özdemir, D., Eschert, T., 2017. Industrial internet of
Breugst, N., Patzelt, H., Rathgeber, P., 2015. How should we divide the pie?: equity things and cyber manufacturing systems. In: Jeschke, S., Brecher, C., Song, H., Rawat,
distribution and its impact on entrepreneurial teams. J. Bus. Ventur. 30 (1), 66–94. D.B. (Eds.), Industrial Internet of Things: Cybermanufacturing Systems. Springer
Burns, R., 1997. Intelligent manufacturing. Aircraft Eng. Aero. Technol. 69 (5), 440–446. International Publishing, Cham, pp. 3–19.
Candido, M.A.B., 1998. A genetic algorithm based procedure for more realistic job shop Kagermann, H., Wahlster, W., Helbig, J., 2013. Securing the future of German manu-
scheduling problems. Int. J. Prod. Res. 36 (12), 3437–3457. facturing industry. Recom. Imple. Strateg. Initiative INDUSTRIE 4.
Cao, Q., Schniederjans, D.G., Schniederjans, M., 2017. Establishing the use of cloud Kang, H.S., Lee, J.Y., Choi, S., Kim, H., Park, J.H., Son, J.Y., et al., 2016. Smart manu-
computing in supply chain management. Oper. Manag. Sci. 10 (1–2), 47–63. facturing: past research, present findings, and future directions. Int. J. Precis. Eng.
Caputo, A., Marzi, G., Pellegrini, M.M., 2016. The Internet of Things in manufacturing Manuf-Green. Technol. 3 (1), 111–128.
innovation processes. Bus. Process Manag. J. 22 (2), 383–402. Kirazli, A., Hormann, R., 2015. A conceptual approach for identifying Industrie 4.0 ap-
Celik, N., Nageshwaraniyer, S.S., Son, Y.-J., 2012. Impact of information sharing in plication scenarios. In: IIE Annual Conference.Proceedings, pp. 862–871.
hierarchical decision-making framework in manufacturing supply chains. J. Intell. Klosowski, G., 2012. Cloud manufacturing concept as a tool of multimodal manufacturing
Manuf. 23 (4), 1083–1101. systems integration. Found. Manag. 4 (1), 17–42.
Chen, W., Muraki, M., 1997. An action strategy generation framework for an on-line Kobbacy, K.A.H., Proudlove, N.C., Harper, M.A., 1995. Towards an intelligent main-
scheduling and control system in batch processes with neural networks. Int. J. Prod. tenance optimization system. J. Oper. Res. Soc. 46 (7), 831.
Res. 35 (12), 3483–3508. Kobbacy, K.A.H., Vadera, S., 2011. A survey of AI in operations management from 2005
Chongwatpol, J., 2015. Prognostic analysis of defects in manufacturing. Ind. Manag. Data to 2009. J. Manuf. Technol. Manag. 22 (6), 706–733.
Syst. 115 (1), 64–87. Košturiak, J., Gregor, M., 1999. Simulation in production system life cycle. Comput. Ind.
Choudhary, A.K., Harding, J.A., Tiwari, M.K., 2009. Data mining in manufacturing: a 38 (2), 159–172.
review based on the kind of knowledge. J. Intell. Manuf. 20 (5), 501–521. Krumeich, J., Werth, D., Loos, P., 2016. Prescriptive control of business processes. Bus. &
Colicchia, C., Strozzi, F., 2012. Supply chain risk management: a new methodology for a Inf. Syst. Eng. 58 (4), 261–280.
systematic literature review. Supply Chain Manag.: Int. J. 17 (4), 403–418. Labib, A.W., Yuniarto, M.N., 2005. Intelligent real time control of disturbances in man-
Cooper, H.M., 1988. Organizing knowledge syntheses: a taxonomy of literature reviews. ufacturing systems. J. Manuf. Technol. Manag. 16 (7), 864–889.
Knowl.Soc. 1 (1), 104–126. Lachenmaier, J.F., Lasi, H., Kemper, H.-g., 2017. Simulation of production processes in-
Corbin, J.M., Strauss, A., 2009. Basics of Qualitative Research, third ed. SAGE, Los volving cyber-physical systems. Procedia CIRP 62, 577–582.
Angeles [u.a.]. Lasi, H., Fettke, P., Kemper, H.-g., Feld, T., Hoffmann, M., 2014. Industry 4.0. Bus. & Inf.
Corley, K.G., Gioia, D.A., 2004. Identity ambiguity and change in the wake of a corporate Syst. Eng. 6 (4), 239–242.
spin-off. Adm. Sci. Q. 49 (2), 173–208. LaValle, S., Lesser, E., Shockley, R., Hopkins, M.S., Kruschwitz, N., 2011. Big data, ana-
Crossan, M.M., Apaydin, M., 2010. A multi-dimensional framework of organizational lytics and the path from insights to value. MIT Sloan Manag. Rev. 52 (2), 21–32.
innovation: a systematic review of the literature. J. Manag. Stud. 47 (6), 1154–1191. Lee, E.T., 1996. Intelligent factories using fuzzy expert systems. Kybernetes 25 (3), 51–55.
David, R.J., Han, S.-K., 2004. A systematic assessment of the empirical support for Lee, I., Lee, K., 2015. The internet of things (IoT): applications, investments, and chal-
transaction cost economics. Strateg. Manag. J. 25 (1), 39–58. lenges for enterprises. Bus. Horiz. 58 (4), 431–440.
Denyer, D., Tranfield, D., 2009. Producing a Systematic Review (The Sage Handbook of Lee, J., 2015. Smart factory systems. Informatik-Spektrum 38 (3), 230–235.
Organizational Research Methods). SAGE Publications, London, pp. 671–689. Lee, J., Bagheri, B., Kao, H.-A., 2015. A Cyber-Physical Systems architecture for Industry
Denyer, D., Tranfield, D., van Aken, J.E., 2008. Developing design propositions through 4.0-based manufacturing systems. Manuf. Lett. 3, 18–23.

14
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Lee, J., Kao, H.-A., Yang, S., 2014. Service innovation and smart analytics for industry 4.0 213–217.
and big data environment. Procedia CIRP 16, 3–8. Rowley, J., Slack, F., 2004. Conducting a literature review. Manag. Res. News 27 (6),
Lincoln, Y.S., Guba, E.G., 1985. Naturalistic Inquiry. SAGE Publications, Beverly Hills, 31–39.
Calif. Rowlinson, M., Procter, S., Hassard, J., 1994. CIM and the process of innovation: in-
Liu, M., Ma, J., Lin, L., Ge, M., Wang, Q., Liu, C., 2017. Intelligent assembly system for tegrating the organization of production. Int. J. Prod. Econ. 34 (3), 359–369. from.
mechanical products and key technology based on internet of things. J. Intell. Manuf. http://www.sciencedirect.com/science/article/pii/0925527394901643.
28 (2), 271–299. Rudtsch, V., Gausemeier, J., Gesing, J., Mittag, T., Peter, S., 2014. Pattern-based business
Lu, J., Liu, S., Cui, L., Pan, L., Wu, L., 2017. Crowd wisdom drives intelligent manu- model development for cyber-physical production systems. Procedia CIRP 25,
facturing. Int. J. Crowd Sci. 1 (1), 39–47. 313–319.
Lu, Y., 2017. Industry 4.0: a survey on technologies, applications and open research is- Ruiz, N., Giret, A., Botti, V., Feria, V., 2011. Agent-supported simulation environment for
sues. J. Ind. Inf. Integrat. 6, 1–10. intelligent manufacturing and warehouse management systems. Int. J. Prod. Res. 49
Lucke, D., Constantinescu, C., Westkämper, E., 2008. Smart factory - a step towards the (5), 1469–1482.
next generation of manufacturing. In: Mitsuishi, M., Ueda, K., Kimura, F. (Eds.), Rymaszewska, A., Helo, P., Gunasekaran, A., 2017. IoT powered servitization of manu-
Manufacturing Systems and Technologies for the New Frontier. The 41st CIRP facturing – an exploratory case study. Int. J. Prod. Econ. 192, 92–105.
Conference on Manufacturing Systems May 26-28, 2008, first ed. Springer Verlag Salipante, P., Notz, W., Bigelow, J., 1982. A matrix approach to literature reviews. Res.
London Limited, Tokyo, Japan, pp. 115–118 s.l. Organ. Behav. 4, 321–348.
Magruk, A., 2016. Uncertainty in the sphere of the industry 4.0 - potential areas to re- Sanders, D., Gegov, A., 2013. AI tools for use in assembly automation and some examples
search. Bus. Manag. Educ. 14 (2), 275–291. of recent applications. Assemb. Autom. 33 (2), 184–194.
Majstorovic, V., Macuzic, J., Sibalija, T., Zivkovic, S., 2015. Cyber-physical manu- Saygin, C., Sarangapani, J., 2006. RFID on the manufacturing shop floor: applications and
facturing systems - manufacturing metrology aspects. In: Proceedings in challenges. In: IIE annual conference.Proceedings, pp. 1–6.
Manufacturing Systems. 10. pp. 9–14 1. Schmidt, R., Möhring, M., Härting, R.-C., Reichstein, C., Neumaier, P., Jozinović, P.,
Malakooti, B., 2011. Systematic decision process for intelligent decision making. J. Intell. 2015. Industry 4.0 - potentials for creating smart products: empirical research results.
Manuf. 22 (4), 627–642. In: Abramowicz, W. (Ed.), Business Information Systems: 18th International
Malakooti, B., 2012. Decision making process: typology, intelligence, and optimization. J. Conference, BIS 2015, Poznań, Poland, June 24-26, 2015, Proceedings. Springer
Intell. Manuf. 23 (3), 733–746. International Publishing, Cham, pp. 16–27.
Malakooti, B., Raman, V., 2000. An interactive multi-objective artificial neural network Schuh, G., Potente, T., Thomas, C., Hempel, T., 2014. Short-term cyber-physical pro-
approach for machine setup optimization. J. Intell. Manuf. 11 (1), 41. duction management. Procedia CIRP 25, 154–160.
Malakooti, B., Zhou, Y.Q., 1994. Feedforward artificial neural networks for solving dis- Segura Velandia, D.M., Kaur, N., Whittow, W.G., Conway, P.P., West, A.A., 2016.
crete multiple criteria decision making problems. Manag. Sci. 40 (11), 1542–1561. Towards industrial internet of things: crankshaft monitoring, traceability and
Mayring, P., 2000. Qualitative content analysis. Forum Qual. Soc. Res. 1 (2), 105–114. tracking using RFID. Robot. Comput. Integr. Manuf. 41, 66–77.
Menascé, D., Krishnamoorthy, M., Brodsky, A., 2015. Autonomic smart manufacturing. J. Seuring, S., Gold, S., 2012. Conducting content‐analysis based literature reviews in supply
Decis. Syst. 24 (2), 206–224. chain management. Supply Chain Manag.: Int. J. 17 (5), 544–555.
Meziane, F., Vadera, S., Kobbacy, K., Proudlove, N., 2000. Intelligent systems in manu- Shariatzadeh, N., Lundholm, T., Lindberg, L., Sivard, G., 2016. Integration of digital
facturing: current developments and future prospects. Integr. Manuf. Syst. 11 (4), factory with smart factory based on internet of things. Procedia CIRP 50, 512–517.
218–238. Shaw, R., 1983. Towards the unmanned factory. Ind. Manag. Data Syst. 83 (3/4), 10–12.
Miles, M.B., Huberman, M.A., 2007. Qualitative Data Analysis, second ed. SAGE Sheng, J., Amankwah-Amoah, J., Wang, X., 2017. A multidisciplinary perspective of big
Thousand Oaks [u.a.]. data in management research. Int. J. Prod. Econ. 191, 97–112.
Mishra, D., Gunasekaran, A., Childe, S.J., Papadopoulos, T., Dubey, R., Wamba, S., 2016. Stock, T., Seliger, G., 2016. Opportunities of sustainable manufacturing in industry 4.0.
Vision, applications and future challenges of Internet of Things. Ind. Manag. Data Procedia CIRP 40, 536–541.
Syst. 116 (7), 1331–1355. Suddaby, R., 2006. From the editors: what grounded theory is not. Acad. Manag. J. 49 (4),
Monostori, L., Kádár, B., Bauernhansl, T., Kondoh, S., Kumara, S., Reinhart, G., et al., 633–642.
2016. Cyber-physical systems in manufacturing. CIRP Ann. - Manuf. Technol. 65 (2), Suginouchi, S., Kokuryo, D., Kaihara, T., 2017. Value Co-creative manufacturing system
621–641. for mass customization: concept of smart factory and operation method using au-
Monostori, L., 2014. Cyber-physical production systems: roots, expectations and R&D tonomous negotiation mechanism. Procedia CIRP 63, 727–732.
challenges. Procedia CIRP 17, 9–13. Sutton, R.I., Callahan, A.L., 1987. The stigma of bankruptcy: spoiled organizational image
Munguía, J., Lloveras, J., Llorens, S., Laoui, T., 2009. Development of an AI-based rapid and its management. Acad. Manag. J. 30 (3), 405–436.
manufacturing advice system. Int. J. Prod. Res. 48 (8), 2261–2278. Syberfeldt, A., Holm, M., Danielsson, O., Wang, L., Brewster, R.L., 2016. Support systems
Newbert, S.L., 2007. Empirical research on the resource-based view of the firm: an as- on the industrial shop-floors of the future – operators' perspective on augmented
sessment and suggestions for future research. Strateg. Manag. J. 28 (2), 121–146. reality. Procedia CIRP 44, 108–113.
Nylund, H., Andersson, P.H., 2011. Framework for extended digital manufacturing sys- Tan, Y.S., Ng, Y.T., Low, J.S.C., 2017. Internet-of-Things enabled real-time monitoring of
tems. Int. J. Comput. Integr. Manuf. 24 (5), 446–456. energy efficiency on manufacturing shop floors. Procedia CIRP 61, 376–381.
O'Donovan, P., Leahy, K., Bruton, K., O'Sullivan, D.T.J., 2015a. Big data in manu- Tang, D., Zheng, K., Zhang, H., Sang, Z., Zhang, Z., Xu, C., et al., 2016. Using autonomous
facturing: a systematic mapping study. J. Big Data 2 (1), 145. intelligence to build a smart shop floor. Procedia CIRP 56, 354–359.
O'Donovan, P., Leahy, K., Bruton, K., O'Sullivan, D.T.J., 2015b. An industrial big data Taticchi, P., Garengo, P., Nudurupati, S.S., Tonelli, F., Pasqualino, R., 2015. A review of
pipeline for data-driven analytics maintenance applications in large-scale smart decision-support tools and performance measurement and sustainable supply chain
manufacturing facilities. J. Big Data 2 (1), 117. management. Int. J. Prod. Res. 53 (21), 6473–6494.
Otto, J., Henning, S., Niggemann, O., 2014. Why cyber-physical production systems need Thoben, K.-D., Wiesner, S.A., Wuest, T., 2017. “Industrie 4.0” and smart manufacturing –
a descriptive engineering approach – a case study in plug & produce. Procedia. a review of research issues and application examples. Int. J. Autom. Technol. 11 (1),
Technol. 15, 295–302. 4–16.
Phaithoonbuathong, P., Monfared, R., Kirkham, T., Harrison, R., West, A., 2010. Web Torraco, R.J., 2005. Writing integrative literature reviews: guidelines and examples.
services-based automation for the control and monitoring of production systems. Int. Hum. Resour. Dev. Rev. 4 (3), 356–367.
J. Comput. Integr. Manuf. 23 (2), 126–145. Tracy, S.J., 2010. Qualitative quality: eight "Big-Tent" criteria for excellent qualitative
Plattform Industrie 4.0, 2016. Progress Report April 2016: Digitization of Industrie - research. Qual. Inq. 16 (10), 837–851.
Plattform Industrie 4.0. Retrieved January 17, 2018, from. https://www.plattform- Tranfield, D., Denyer, D., Smart, P., 2003. Towards a methodology for developing evi-
i40.de/I40/Redaktion/EN/Downloads/Publikation/digitization-of-industrie- dence-informed management knowledge by means of systematic review. Br. J.
plattform-i40.pdf?__blob=publicationFile&v=4. Manag. 14 (3), 207–222.
Porter, M.E., Heppelmann, J.E., 2015. How smart, connected products are transforming Vom Brocke, J., Simons, A., Niehaves, B., Niehaves, B., Reimer, K., Plattfaut, R., Cleven,
companies. Harv. Bus. Rev. 93 (10), 96–114. A., 2009. Reconstructing the giant: on the importance of rigour in documenting the
Pratt, M.G., 2007. Fitting oval pegs into round holes: tensions in evaluating and pub- literature search process. In: European Conference on Information Systems in 2009. ..
lishing qualitative research in top-tier North American journals. Organ. Res. Methods Wang, J., Zhang, L., Duan, L., Gao, R.X., 2017a. A new paradigm of cloud-based pre-
11 (3), 481–509. dictive maintenance for intelligent manufacturing. J. Intell. Manuf. 28 (5),
Qin, J., Liu, Y., Grosvenor, R., 2016. A categorical framework of manufacturing for in- 1125–1137.
dustry 4.0 and beyond. Procedia CIRP 52, 173–178. Wang, L., Törngren, M., Onori, M., 2015. Current status and advancement of cyber-
Qu, T., Yang, H.D., Huang, G.Q., Zhang, Y.F., Luo, H., Qin, W., 2012. A case of im- physical systems in manufacturing. J. Manuf. Syst. 37, 517–527.
plementing RFID-based real-time shop-floor material management for household Wang, S., Ouyang, J., Li, D., Liu, C., 2017b. An integrated industrial ethernet solution for
electrical appliance manufacturers. J. Intell. Manuf. 23 (6), 2343–2356. the implementation of smart factory. IEEE Access 5, 25455–25462.
Ren, L., Zhang, L., Tao, F., Zhao, C., Chai, X., Zhao, X., 2013. Cloud manufacturing: from Wang, S., Wan, J., Li, Di, Zhang, C., 2016a. Implementing smart factory of industrie 4.0:
concept to practice. Enterp Inf. Syst. 9 (2), 186–209. an outlook. Int. J. Distributed Sens. Netw. 12 (1), 3159805.
Ren, L., Zhang, L., Wang, L., Tao, F., Chai, X., 2016. Cloud manufacturing: key char- Wang, S., Wan, J., Zhang, D., Li, Di, Zhang, C., 2016b. Towards smart factory for industry
acteristics and applications. Int. J. Comput. Integr. Manuf. 30 (6), 501–515. 4.0: a self-organized multi-agent system with big data based feedback and co-
Rezaei, M., Akbarpour Shirazi, M., Karimi, B., 2017. IoT-based framework for perfor- ordination. Comput. Network. 101, 158–168.
mance measurement. Ind. Manag. Data Syst. 117 (4), 688–712. Webster, J., Watson, R.T., 2002. Analyzing the past to prepare for the future: writing a
Romero, D., Jardim-Goncalves, R., Grilo, A., 2017. Factories of the future: challenges and literature review. MIS Q. 26 (2), xiii–xxiii.
leading innovations in intelligent manufacturing. Int. J. Comput. Integr. Manuf. 30 Winnig, L.W., 2016. GE's big bet on data and analytics. MIT Sloan Manag. Rev. 57 (3).
(1), 1–3. Wu, D., Greer, M.J., Rosen, D.W., Schaefer, D., 2013. Cloud manufacturing: drivers,
Rooks, B., 2000. Tecnomatix weaves its e‐Manufacturing web. Assemb. Autom. 20 (3), current status, and future trends. In: Proceedings of the ASME 8th International

15
P. Osterrieder, et al. International Journal of Production Economics xxx (xxxx) xxx–xxx

Manufacturing Science and Engineering Conference - 2013. Presented at ASME 2013 Zhang, D., He, Z., Qian, Y., Wan, J., Li, D., Zhao, S., 2016. Revisiting unknown RFID tag
8th International Manufacturing Science and Engineering Conference, June 10 - 14, identification in large-scale internet of things. IEEE Wirel. Commun. 23 (5), 24–29.
2013, Madison, Wisconsin, USA (V002T02A003). ASME, New York, NY. Zhang, Y., Wang, W., Liu, S., Xie, G., 2014. Real-time shop-floor production performance
Wu, L., Yue, X., Jin, A., Yen, D.C., 2016. Smart supply chain management: a review and analysis method for the internet of manufacturing things. Adv. Mech. Eng. 6, 270749.
implications for future research. Int. J. Logist. Manag. 27 (2), 395–417. Zheng, M., Wu, K., 2017. Smart spare parts management systems in semiconductor
Wuest, T., Irgens, C., Thoben, K.-D., 2014. An approach to monitoring quality in manu- manufacturing. Ind. Manag. Data Syst. 117 (4), 754–763.
facturing using supervised machine learning on product state data. J. Intell. Manuf. Zhong, R.Y., Huang, G.Q., Dai, Q.Y., Zhang, T., 2014. Mining SOTs and dispatching rules
25 (5), 1167–1180. from RFID-enabled real-time shopfloor production data. J. Intell. Manuf. 25 (4),
Xu, X., 2012. From cloud computing to cloud manufacturing. Robot. Comput. Integr. 825–843.
Manuf. 28 (1), 75–86. Zhong, R.Y., Huang, G.Q., Lan, S., Dai, Q.Y., Chen, X., Zhang, T., 2015. A big data ap-
Yeh, C.‐H., 1995. Production data modelling: an integrated approach. Int. J. Oper. Prod. proach for logistics trajectory discovery from RFID-enabled production data. Int. J.
Manag. 15 (8), 52–62. Prod. Econ. 165, 260–272.
Yin, R.K., 2009. Case study research. In: fourth ed. Applied Social Research Methods Zhong, R.Y., Lan, S., Xu, C., Dai, Q., Huang, G.Q., 2016. Visualization of RFID-enabled
Series, vol. 5 SAGE, Los Angeles, Calif [u.a.]. shopfloor logistics big data in cloud manufacturing. Int. J. Adv. Manuf. Technol. 84
Yu, C., Xu, X., Lu, Y., 2015. Computer-integrated manufacturing, cyber-physical systems (1–4), 5–16.
and cloud manufacturing – concepts and relationships. Manuf. Lett. 6, 5–9. Zorn, T., Campbell, N., 2006. Improving the writing of literature reviews through a lit-
Zancul, E.d.S., Takey, S.M., Barquet, A.P.B., Kuwabara, L.H., Cauchick Miguel, P.A., erature integration exercise. Bus. Commun. Q. 69 (2), 172–183.
Rozenfeld, H., 2016. Business process support for IoT based product-service systems Zuehlke, D., 2008. SmartFactory – from vision to reality in factory technologies. In: IFAC
(PSS). Bus. Process Manag. J. 22 (2), 305–323. Proceedings Volumes 41. pp. 14101–14108 (2).

16