Advanced Engineering Informatics 54 (2022) 101793

Contents lists available at ScienceDirect

Advanced Engineering Informatics

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

Full length article

Implications of data-driven product design: From information age towards

intelligence age
Zuoxu Wang a, Pai Zheng b, *, Xinyu Li c, Chun-Hsien Chen d
a
Department of Industrial and Manufacturing Systems Engineering, Beihang University, People’s Republic of China
b
Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong Special Administrative Region
c
College of Mechanical Engineering, Donghua University, Shanghai, People’s Republic of China
d
School of Mechanical and Aerospace Engineering, Nanyang Technological University, 639798, Singapore

A R T I C L E I N F O A B S T R A C T

Keywords: Data-driven design (D3 ), a new design paradigm benefited from advanced data analytics and computational
Data-driven design intelligence, has gradually promoted the research of data-driven product design (DDPD) ever since 2000 s. In
Product design and development today’s Intelligence Age, some theoretical and practical studies have tried to achieve the advanced intelligence
Digital transformation
capabilities in DDPD. However, to the authors’ best knowledge, there is still a lack of a holistic summary of DDPD
Computer intelligence
Cognitive computing
with chronological concern in the intelligence age. To bridge the gap, this research undertakes a literature review
of DDPD publications from 2000 to date (19/09/2022), of which 172 relevant papers are discussed via biblio­
metric analysis and state-of-the-art analysis. The results shown that DDPD has vitality in the Intelligence Age by
combining the cutting-edge digital technologies, such as AI, additive manufacturing, digital twin, and so on.
Moreover, current DDPD studies could outperform the classical design methods on the well-defined tasks, but it
still cannot master the creative/innovative design tasks which require the cognitive capability. This survey
further highlights several future research potentials including cognitive intelligence-enabled design, end-to-end
design integration, advanced design knowledge support, design for additive manufacturing, and sustainable
smart product-service systems. It is hoped that this work can be regarded as a reference to understand the
roadmap of DDPD and offer insights for the design practitioners to complete relevant tasks in today’s intelligence
age.

and intelligence approaches (e.g., AI algorithms), thus boosting the ef­

ficiency on product design.
1. Introduction Nowadays, DDPD research themes have become more profound and
much broader. From the technological aspect, lots of well-defined design
Digitalization, which opens the Information Age and the third in­ tasks can be excellently conducted. However, different techniques and
dustrial revolution [1], has radically changed the world for decades, algorithms were adapted into various well-defined design tasks, causing
including how people communicate and collaborate [2], how manu­ delicate DDPD niches [9]. From the business aspect, the academics and
facturers produce [1,3], and how companies create business values practitioners in DDPD have concerned with an increasing number of
[4,5]. In this context, data-driven design (D3 ), a new design paradigm issues, such as the time to market, quick response to the dynamic re­
owing to big data and computational intelligence approaches, has quirements, human cognition workloads in human–machine interac­
gradually promoted the research of data-driven product design (DDPD). tion, and user interface layout, thus making the product design broader
DDPD mainly refers to the process of design analysis and computational as an ecosystem [10].
design synthesis activities by leveraging massive product through- All those dramatic transformations facilitated the product develop­
lifecycle data [6]. Compared with conventional product design para­ ment field but also made the design concerns miscellaneous and without
digms dominated by engineers, DDPD shifted many design activities as a mainstream. In other words, with plenty of disparate design concerns,
semi-automated or automated activities, such as decision-makings in the students, researchers, and practitioners can hardly understand the
product design and computational design synthesis [7,8] with the existing achievements in DDPD and further explore the next research
application of the advanced digital technologies (e.g., IoT, CAD, PLM)

* Corresponding author.
E-mail address: [email protected] (P. Zheng).

https://doi.org/10.1016/j.aei.2022.101793
Received 23 May 2022; Received in revised form 19 September 2022; Accepted 18 October 2022
Available online 31 October 2022
1474-0346/© 2022 Elsevier Ltd. All rights reserved.
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Nomenclature GD Generative design

GD&T Geometric dimensioning and tolerancing
AI Artificial intelligence IoT Internet of Things
AD Affective design KBS Knowledge-based system
AM Additive manufacturing KE Kansei engineering
AR Augmented reality KG Knowledge graph
BEM Boundary-element method MCDM Multi-criteria decision-making
CAD Computer-aided design ML Machine learning
CBR Case-based reasoning MR Mixed reality
CFD Computational fluid dynamics NLP Natural language processing
CI Computational intelligence PLM Product lifecycle management
CNN Convolutional neural networks QFD Quality function deployment
D3 Data-driven design SCPs Smart and connected products
DDPD Data-driven product design SVM Support vector machine
DEMATEL Decision making trial and evaluation laboratory TRIZ Theory of inventive problem solving
DL Deep learning UCD User-centered design
FBS-ontology Function-behavior-structure ontology UGC User-generated contents
FEA Finite element analysis UX User experience
FEM Finite element method VAE Variational autoencoder
GAN Generative adversarial networks VR Virtual reality

directions of DDPD. To the authors’ best knowledge, although some computer-assisted design tools, such as solid modelling, numerical
studies have summarized the data-driven methodologies in the engi­ analysis, and computational design synthesis (CDS). In contrast, data-
neering field [6,8,11–13], there is still a lack of a survey to illustrate the driven product design intends to apply large-scale data for design.
evolution process of DDPD with chronological concern from the infor­
mation age to the intelligence age. To bridge this gap, this paper relooks
2.2. DDPD vs Data-enabled, data-informed, and data-centric product
at the status of DDPD and further addresses the challenges and future
design
perspectives of the existing DDPD approaches. Through the systematic
survey, it is hoped to clarify the following research questions:
Focusing on product design based on data, there are still several
similar design terms: data-enabled product design, data-informed product
(1) What is the essence of DDPD? How to distinguish it from other
design, data-driven product design, and data-centric product design. They
similar design terms?
can be clarified according to the extent of the data involved in design
(2) What are the driving forces that trigger the development of
processes [14].
DDPD? Will they continuously influence DDPD?
(3) What can the current DDPD methodologies achieve?
• Data-enabled product design indicates that the design activities can be
(4) What are the challenges and future perspectives of DDPD?
aware of the data but are not required to leverage it.
• Data-informed product design means that the design activities can
The rest of this paper is organized as follows. Section 2 compares the
analyze the data for decision-making.
essential meanings of DDPD with other similar terms. Section 3 shows
• DDPD refers to the activities that use data as the primary enabler for
the literature selection process and demonstrates an overall statistical
generating values, including design modelling and design reasoning.
analysis of DDPD. A comprehensive review was conducted based on the
• Data-centric product design is the paradigm in which the steps/pro­
selected literature from the perspective of driving forces behind the
cesses of design activities, including design modelling and design
DDPD (Section 4) and the evolution of DDPD (Section 5). The state-of-
reasoning, are predominated by the data; human’s contributions
the-art DDPD methods is summarized in Section 6. Main challenges
become subtle.
and future perspectives were also addressed in Section 7 and Section 8,
respectively. Finally, research contributions and main findings are
summarized in Section 9. 3. Literature search methodology

2. Clarification of data-driven product design The systematic literature selection about DDPD in this survey follows
the bibliometric process [15], as depicted in Fig. 1. Firstly, Web of Sci­
Before discussing the state-of-the-art of DDPD, the first step is to ence is selected as the database due to its broad scope of high-quality
clarify the meaning of DDPD. Since DDPD can be regarded as a discrete articles and reviews. The literature collection is conducted with a set
field of data-driven design in product design and development, this of expert-predefined keywords that covers the scope of “data-driven”
section aims to distinguish DDPD from other similar design terms to and “product design”. The search string is:
avoid misunderstandings or confusions. Topics =
(data-driven design OR artificial intelligence OR machine learning OR
2.1. DDPD vs Computer-aided product design deep learning OR neural network OR data science OR data analysis OR
data mining OR big data OR text mining) AND
Computer-aided product design and data-driven product design are the (product design OR product development OR product planning OR con­
two design paradigms that are easily mixed. Since the large-scale data­ ceptual design OR embodiment design OR detail design OR engineering
sets must be manipulated on computers, Computer-aided product design design OR industrial design OR computer-aided design OR CAD OR
is a broader paradigm that emphasizes the computational power of requirement analysis OR user experience)
digital computers. It refers to the design paradigm that applies AND Language = English

2
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Fig. 1. Literature selection procedure.

AND Year Published = (2000–2022) listed. Industrial Management & Data Systems (IMDS), Journal of Me­
AND Document Type = (article OR review). chanical Design (JMD), Computers & industrial Engineering, Advanced En­
gineering Informatics and Journal of Intelligent Manufacturing outnumber
The search results are restricted by Web of Science Categories = (En­
the other journals. Most of the articles from IMDS underline the cross-
gineering Multidisciplinary OR Engineering Industrial OR Engineering Me­
disciplinary research in the field of operation management and infor­
chanical OR Engineering Manufacturing OR Ergonomics OR Art)’ The
mation systems, and a small portion of the IMDS articles are related to
irrelevant literature in other fields such as chemistry, education, etc. are
DDPD. JMD addresses nearly all the engineering design activities.
excluded. A total of 3662 research articles are gathered. Secondly, the
Notably, JMD probed into the potentials and advances of DDPD by
publications from the irrelevant journals are removed, resulting in 696
announcing a special issue in 2016 [16].
research items. The DDPD-related journals include but is not limited to
Journal of Mechanical Design, Computers & Industrial Engineering,
4. Influential factors towards DDPD
Advanced Engineering Informatics, Journal of Intelligent Manufacturing,
Research in Engineering Design, CIRP Annals Manufacturing Technology,
Fig. 3 offers a chronological summary of typical product design
Journal of Engineering Design, Design Studies, and International Journal of
methodologies. Product design approaches are generally grouped into
Design. At last, the literature records are manually screened and
four types, namely descriptive product design paradigm, prescriptive product
restricted to the product design activities, such as product planning,
design paradigm, computer-aided product design, and DDPD. Descriptive
conceptual design, embodiment design, and detail design according to
product design approaches intends to enhance the designers’ creativity,
the title, abstract and keywords. As a result, 172 research journal articles
focusing on the creative design process/strategies/methods/tools, such
are refined for in-depth analysis to represent the extant DDPD and
as brainstorming and analogy [17]. Prescriptive product design covers
explore future research directions.
numerous systematic design methodologies, such as TRIZ, QFD, FBS-
A general statistical analysis was conducted based on the 1st-round
ontology, Axiomatic design, and so on. Computer-aided product design
literature search result. As shown in Fig. 2(a), research interests on
greatly enhance the product design process with higher efficiency and
DDPD proliferate since 2015. From 2000 to 2015, the related document
effectiveness by utilizing numerous mathematical algorithms and com­
counts stably increased from around 50 to 120. Thereafter, a significant
puter software. With the advance of the Internet, social network, and
increase occurred from 2015 to 2020, with 467 research studies pub­
intelligent algorithms, DDPD applies multimodal data, including user-
lished in 2021.
generated content, products/services data, and stakeholders’ interac­
Fig. 2(b) shows the universities or research institutes that account for
tion data/records for design activities.
the top 20 % of publications in DDPD. Shanghai Jiao Tong University
The product design methods are significantly affected by techno­
ranks the top with 60 publications in the past two decades, followed by
logical progress (i.e., the top part of Fig. 3) and marketing trans­
Georgia Institute of Technology (55), Hong Kong Polytechnic University
formations (i.e., the bottom part of Fig. 3). To understand the influential
(46), Massachusetts Institute of Technology (43), Pennsylvania State
factors that motivated DDPD, this section provides a broad discussion on
University-University Park (42), and Beihang University (37), to name a
the impacts of technological innovations (Section 4.1) and business
few. China and North America are the primary districts devoting to the
strategies (Section 4.2).
research of DDPD.
In Fig. 2(c), the Q1 or Q2 academic journals based on Journal Cita­
tion Reports (JCR) that contribute 80 % DDPD-related publications are

3
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Fig. 2. Statistical analysis results from Web of Science (Accessible Date: 19/09/2022).

4.1. Technological driving forces management systems, the value of data is revealed by product lifecycle
management (PLM) systems with a systematic view, which became the
The technological driving forces usually radically change product initial of the latter data-driven design.
design along with the industrial revolutions. As the top part of Fig. 3 With the development of Industry 4.0, the Intelligence Age begins with
shown, the product design has been upgraded through design with the application and integration of multiple technologies, such as cloud
versatile craftsman in Handicraft age, design with labor-division coop­ computing, Internet of Things (IoT), big data and analytics, additive
eration in Mechanical and Electronic age, design with computerization manufacturing (AM), autonomous robots, augmented reality (AR),
in the Information age, and AI-enabled design in the Intelligence age. artificial intelligence (AI), and so on [21–23]. Those technological ad­
Since the Information age, electronic digital computers, as the driving vances have changed the product design mainly from three aspects: (1)
force, have thoroughly changed the product design paradigm, as shown new design issues triggered by AM and AR [24]; (2) more decision-
in Table 1. Since the mid-1960 s, design with computerization has grad­ makings [25,26] with big data in the design stage; and (3) computa­
ually evolved the paradigm of CAD for higher design efficiency and tional design synthesis (CDS) requiring more design concepts generation
lower human errors in the engineering drawing process [18]. On the one with limited resources. For the first one, new design issues such as
hand, computer graphics technologies enable the geometric modelling standard data exchange files (e.g., Standard data exchange files (STL
of complex curved surfaces, e.g., the Bezier curves and surfaces [19], files)) are proposed because of AM. For the second one, computational
achieving digitalized design records from paper-based design record and intelligence, such as big data/analytics and AI, began to assist design
realizing fewer human errors. On the other hand, computational analytics engineers to make rational decisions [27]. Massive product-related data
[20], such as data mining, evolutionary computation, and CBR, enable are vertically integrated throughout the product lifecycle, and more
the simulation, evaluation, and validation of product models with user-generated content (UGC) becomes accessible owing to the wide­
higher efficiency and lower probability of quality failures during spread SCPs [28]. Meanwhile, the design decisions are made based on a
manufacturing. Besides, with the emergence of information comprehensive concern about user experience, including customer

4
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Fig. 3. Design paradigm shifts from the past to the present.

requirement [29,30], affective and cognitive design [31], personalized

Table 1
design [32], and co-design [33], in which those dimensions were
Product design spotlights along with the industrial revolutions.
emphasized as the value propositions. To achieve it, data analysis
Ages Driving forces Product design spotlights in the age techniques and AI techniques have boosted design informatics with their
Information • Digital computers, • Design with computerization impressive performances via pattern recognition among the multi-
Age • Internet • From paper-based records to digita­ sourced data and autonomic and reasonable decision-making. For the
lized records third one, generative design (GD) achieved by deep learning techniques
• Computational analytics in product
design
has also shown its strength in generating thousands of design alterna­
• Product design with systematic tives within limited resources [27,34,35]. AI techniques have become
concerns one of the primary enablers in the Intelligence Age.
Intelligence • ICTs, e.g., IoT, Cloud • New design issues triggered by new
Age computing, and CPS technologies, e.g., AM, AR
• AR • More decision-making with big data 4.2. Market transformations
• AM • Based on SCPs
• SCPs • User-experience-based design The market transformations towards DDPD also need to be clarified.
• Big data • Data-driven design
• Cognitive computing • AI-enabled design
The progress of product design caused by the market transformation can
• More design concept generation (i.e., be discussed from two perspectives, i.e., market strategy transformation
computational design synthesis) and digital servitization (originated from big data), as shown at the
within the limited time bottom of Fig. 4.
• AI-enabled design
Manufacturers have adopted various market strategies in different
ages, including mass marketing, multi-product mass marketing/market
segmentation/listen to the voice of customer, mass customization, and

5
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Fig. 4. Influential factors on DDPD.

mass personalization. As summarized in Table 2, mass customization business strategies under the context of technology advances. As for the
went further from the market-of-few to the market-of-one, namely mass DDPD paradigm itself, besides the new design tasks issued by new
personalization [36]. The realization of mass personalization requires technologies, the main design tasks can be grouped into design infor­
multiple resources and technologies, including personal data, Internet, matics and design synthesis. Both are facing the challenges of more
digital platform, and configurable products [36]. The evolution of decision-making and more concept generation under limited resources.
product development along with the marketing strategies is a way of
pursuing the exact fulfilment of individual requirements with flexibility 5. Evolution of DDPD
and quick responses.
Another progress in the market transformation is digital servitization This section reviews the total trend of DDPD and the evolution of
initialized by the use of big data. As the arrows in Fig. 3 shown, since the each stage in DDPD, including product planning, conceptual design,
mid-2000 s, social networks and digital platforms have connected in­ embodiment design, and detail design [9].
dividuals, allowing big data collection and analysis. Since then, big data
analytics has become a significant role in the current digital trans­ 5.1. Total trend of DDPD
formation strategy. Digital transformation is a broad strategy that
comprehensively integrates big data analysis and other technologies, Fig. 5 shows the keywords about DDPD based on the 667 research
and further profoundly alters the business. It is worth mentioning that items in the 2nd literature search round. It is visualized by a widespread
the digital integration of upstream and downstream, as a digital scientific literature visualization tool called CiteSpace [39]. The key­
ecosystem, also significantly affect the organizations of product design words are labelled from the circle centres, and the circle radius refers to
engineering from the system level. the term frequency. The larger the radius is, the higher the frequency of
the keyword appears. Meanwhile, the horizontal axis represents the
4.3. Summary of the influential factors on data-driven design timeline, and the vertical axis indicates no meaning. The circles of the
keywords are vertically arranged to avoid the overlapping.
As shown in Fig. 4, advanced ICTs and digitalization technologies, Since technologies usually grow from infancy to mature adoptions,
big data, and cognitive computing support the upgrades of product we can explore the evolution of DDPD according to the changes of circle
design and development from the technical perspective. Mass person­ radius and the keyword meanings. Fig. 5 shows that DDPD can be
alization, values in big data, and the shift towards downstream also generally grouped into three phases, i.e., DDPD with IT, DDPD with
upgrade the conventional product design to DDPD from the business Computational intelligence (CI), and DDPD with digital transformation.
perspective. Digital transformation, as an umbrella, wrapped up the Phase I mainly focuses on the optimization and the modelling of
product design, which generally indicates the design modelling/synthesis
Table 2 tasks such as design optimization exploration. Besides, the management
Product design spotlights along with the marketing strategy transformation. of product design tasks based on the Internet was also attracted research
Marketing Time Design spotlights Examples interests to a certain extent.
strategies The emphasis of Phase II is big data analytics with data mining/neural
Mass 1990 s- • Promoted by Ford’s assembly line and
networks/machine learning/text mining techniques, which is enabled by
customization 2000 s Internet Budd’s all-steel vehicle the computational power and the learning capability of intelligent al­
[37,38] • Design for [18] gorithms. More product design tasks, such as quality control, diagnosis,
fragmented decision making, appear in this phase. Those design tasks require
demands
advanced data analytics and decision makings on the performance of
Mass 2000 s- • Design for Amazon’s personalized
personalization present individual recommendation [31] product design.
[36] demands Furthermore, beginning from 2018, it can be observed in Fig. 5 that
• Requires multiple many new technologies and innovations, e.g., additive design (3D printing)
data and and Cyber-physical systems, appeared under the environment of digital
technologies
transformation in Phase III. Till the day that created Fig. 5 (19/09/

6
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Fig. 5. Research keywords evolution of DDPD.

2022), no outstanding technologies have dominated the product design big data was firstly deployed in concept design, followed by the other
field. Meanwhile, new concerns also emerged, including product-service stages around 2017 or 2018. This big data trend is consistent with the
systems, uncertainty, and sustainability, besides the continuation of Phase timestamp of big data of Phase II in Fig. 5. Third, since 2019, the
II’s analytical and decision-making tasks. It is worthwhile noting that keyword of deep learning has stood out from other keywords in product
although some terms are not newly created in their own fields, we still planning. Meanwhile, it also appeared in concept design and detail
cannot overlook their potentials of inspiring new possibilities/oppor­ design in 2021 and 2020, respectively. Although it appeared together
tunities by integrating them into the DDPD. with other keywords such as digital twin and AM, the trend of deep
learning to achieve higher intelligence capabilities in DDPD is still
apparent. By laterally comparing the three keywords’ evolution lines in
5.2. Evolution of each stage in DDPD Fig. 6, we can also find that the gaps between different design stages on
prevalence have been gradually decreased that deep learning technique
The literature in the 2nd search round is further grouped into the has been deployed in to DDPD at about the same time.
stages of product planning, concept design, embodiment design, and
detail design. The evolution of research keywords in each stage of DDPD 6. Status of DDPD approaches in design stages
is illustrated in Fig. 6. Similar to Fig. 5, the research keywords are
extracted from the title, abstract, and keywords in the literature in the This section investigates whether the data-driven methods can effi­
2nd literature search round and visualized by CiteSpace [39]. The x-axis ciently and effectively achieve the design tasks in different stages,
refers to the timeline of the appearance of the keywords. The y-axis including product planning, concept design, embodiment design, and
indicates no meaning. For the ease of visualization, the circles are detail design. Fig. 7 provides an overview of the summary of DDPD
vertically arranged to prevent overlapping texts. methods in each stage.
Broadly speaking, detail design has the most active research interests
in DDPD, illustrated by the significantly larger radius on the great mass
of research keywords. 6.1. Product planning
From a specific point of view, as highlighted in Fig. 6, some common
keywords appeared in all stages but with different timestamps, namely, Product planning, covering marketing analysis and preliminary as­
neural network, big data, and deep learning. First, since 2013, the sessments on technical and manufacturing features of the products,
studies of detail design are the first batch to apply neural networks in should be done initially before constructing the product concepts. This
product design, followed by product planning and concept design (both stage could take several months for personnel from marketing, design,
ascend to 2006). With huge radius on neural networks and product manufacturing, finance, etc., to build up a business case, including
quality, it is clear that Phase I of DDPD pay the most attention to detail market segmentation, product positioning, product family construction,
design, aligning with the result of Phase I in Fig. 5. Second, since 2015, etc.

7
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Fig. 6. Research keywords evolution of each stage in product design.

Requirement management is the core step of product planning, con­ [61], have been continuously attempted to complete the design tasks in
taining three subtasks: requirement elicitation, requirement evaluation and the uncertain and heterogeneous product design environment. The data
prioritization, and dynamic requirement management [40], as shown in majorly applied in this stage is the textual UGC, including customer
Table 3. reviews and comments, whereas user behavioural data, such as eye
Conventional requirement elicitation has applied many business movements [58] and historical browsing/purchase data [49], were also
analysis approaches can be found in this stage, including questionnaire, integrated into the research as well.
focus group, interview, brainstorming, etc. [41]. They highly rely on the
intuition of the market analysts and the interactions between end-users 6.2. Conceptual design
and the market teams, which cannot guarantee effectiveness because of
the high reliance on professional domain knowledge and the efficiency Conceptual design contains a series of design activities to identify the
due to the long time spent. However, the trend of data-driven ap­ functions and structures of products. It usually needs to align the func­
proaches, such as data mining techniques [47], NLP techniques [43,44], tional requirements with the customer requirements (i.e., function
and knowledge mining techniques [45], began to be applied in specification), generate concepts (i.e., design concept generation), and
requirement elicitation. evaluate and select proper design concepts (i.e., design concept evalua­
To evaluate and prioritize the requirements, classical decision-making tion), as shown in Table 4.
methods (e.g., AHP/ANP [70,71], DEMATEL[71,72]) and mathemat­ To quickly produce products that meet customers’ requirements,
ical algorithms (e.g., fuzzy theory and rough set theory [73]) still function specification methods have already been maturely developed
dominate the research. Nevertheless, they are the non-data-driven ap­ with data-driven approaches. To reuse the design knowledge from the
proaches because the evaluation indicators are manually predefined. previous design cases, case-based reasoning (CBR) has been applied to
Moreover, the rating scores for analysis are relatively small-scale. In the solve the new problems by associating similar cases and modifying
intelligence age, machine learning approaches (e.g., artificial bee colony based on them. CBR can overcome the bottlenecks in the knowledge
algorithm [57] and Naïve Bayes [59]) and deep learning models such as acquisition process [6,126]. However, it will show reduced effectiveness
hierarchical attention network [60] have already been attempted based while previous design cases are inaccessible or limited; meanwhile, it
on the accessible UGCs. They surpass the classical MCDM methods by cannot outperform in the radical product design. Besides CBR, other
their self-learning capabilities to reveal hidden patterns among the data-driven methods, including neural network (NN) [48,80], genetic
UGCs. algorithm (GA) [79], Support vector machine (SVM) [74], text mining
The data-driven studies for dynamic requirement management have techniques [78], deep learning algorithms (e.g., convolutional neural
begun since the 2010 s [63]. The intelligent approaches driven by AI, networks (CNN) [127], transfer learning [76], hierarchical attention
such as artificial immune system [63], neural network system, and NLP networks [60]) have also been attempted in the function specification

8
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Fig. 7. State-of-the-art of different design stages of DDPD.

tasks. Technically, NN can quickly extract patterns from large-scale

Table 3
data; hence, it can handle the ever-evolving and incremental data
Data-driven approaches in requirement management.
such as customer comments. SVM can outperform on relatively small
Design tasks Pain points of Typical data- Refs datasets. GA has strength in ease of use and extensibility. Furthermore,
conventional driven methods
text mining, computer vision [127], and knowledge graph-based models
design tasks
[75] can analyze end-user’s sentiments or semantics in terms of their
Requirement 1. Difficulties in • Data mining/ [30,41,42–56] expected products. Briefly, the essence of function specification is the
elicitation understanding and text mining
prioritizing user • Knowledge
mapping process between customer requirements and product functions
needs mining [74,128], numerous data-driven approaches suit this task only if they
2. Hard to manage approaches have the ability of pattern extraction.
dynamic • NN Design concept generation intends to create conceptual product models
requirements • NLP
with functions, design structures/features but without the dimensions
Requirement 1. Need experts to • AI algorithms, [57–60]
evaluation & manually identify e.g., artificial and tolerances in order to distinguish it with the design synthesis task in
prioritization the evaluation bee colony detail design [129]. Many design heuristics approaches have been
indicators algorithm developed for concept generation, such as brainstorming [95], TRIZ
2. Apply [94], general sorting technique [96], morphological chart [97], etc.
algorithms
without learning
However, they depend on the domain knowledge/experience of the
capability product designers. To tackle the drawback of intensive labor demand,
Dynamic 1. Not timely • Data mining/ [49,55,56,58,61–69] three kinds of data-driven approaches have been developed: (1)
requirement update dynamic text mining crowdsourcing, (2) case-based/knowledge-based approaches, and (3)
management requirements • Knowledge
generative design. Crowdsourcing [93,99], the process of soliciting
2. No dynamic mining
requirement approaches large-scale inputs from Internet, has become faddish in the 2010 s. Be­
forecast • AI algorithms sides, similar to the function specification, historical design cases inte­
grated with CBR/KBR can also be reused for concept generation. For
instance, Chang and Chen [98] proposed a graphics-based concept
generation approach that integrates knowledge-based system (KBS) and

9
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Table 4
Data-driven approaches in conceptual design.
Design tasks Pain points of conventional design tasks Typical data-driven methods Refs

Function 1. Highly rely on the experience/domain knowledge/inspirations of • CBR [48,51,60,68,69,74–92]

specification engineers • Neural network (NN)
• Support vector machine (SVM)
• Genetic algorithm (GA)
• Text mining
• Knowledge graph models
Design concept 1. Highly rely on the experience/domain knowledge/inspirations of • Crowdsourcing [92–110]
generation engineers • Generative design with ML/DL approaches
• Knowledge-based/case-based
Design concept 1. Need experts to predefine the evaluation indicators • Data mining/text mining in a crowdsourcing [89–92,102,104–106,111–125]
evaluation 2. Restricted by merely textual data environment
3. Hard to quickly respond due to strict laboratory environment • Sensory user experiment approaches, e.g., eye
settings, long experiment duration, proper subject recruitment, etc. movement, sensory data analysis
• Deep learning approaches, e.g., LSTM, RNN

3D printing, verified by generating a large number of graphic golf club critical tasks in the embodiment design and detail design [9,133].
concepts. Furthermore, generative design [9,130], a creative design To achieve the mass customization, researchers and engineers have
strategy that generates many feasible design alternatives via iterative spurred a concentration on product family design and platform-based
algorithms, has raised widespread research interests from conceptual product development. One endeavour is configurational product fam­
design to detail design. Khan and Awan [100] proposed a generative ily design, which develops a modular product platform to generate
design (GD) technique called space-filling GD technique for innovative product varieties by adding, substituting, and/or removing one or more
design creation, in which the criterion of space-filling is applied for the functional modules [134]. The configurational product family design
aim of uniformly explore designs in the design space. Machine learning/ usually cluster product elements in terms of several predefined metrics,
deep learning algorithms have recently capacitated the generative such as modularity, commonality, variety, cost, profit, and so on. For
design strategy with learning capability and generative capability. instance, Lin et al. formalized the modularity analysis of a product ar­
However, major relevant studies lay in embodiment design rather than chitecture design as a multi-objective optimization problem and solved
concept generation [34,131]. it with the consideration of commonality, economic efficiency and
The goal of design concept evaluation is to choose several satisfactory performance loss [135]. Baylis et al. proposed a product family platform
design concepts for further detail design. The evaluation can be con­ selection method using a Pareto front of maximum commonality and
ducted from multiple dimensions for companies’ benefits, including strategic modularity [136]. Martin & Ishii proposed a “design-for-vari­
functional dimensions (e.g., customer satisfaction), physical dimensions ety” approach that incorporates standardization and modularization to
(e.g., product quality, reliability, and durability), and economic di­ reduce future design costs and efforts [137]. Alizon et al. integrated
mensions (e.g., cost, investment appraisal) [132]. In this context, clas­ Design Structure Matrix, Value Analysis, Commonality versus Diversity
sical multi-criteria decision-making (MCDM) approaches are qualified Index to improve and existing product family [138]. Peng et al. proposed
[6]. In contrast, if the design team concentrates more on user experi­ open-architecture product to achieve the modular product design for
ences, such as customers’ satisfaction, affection, and cognition, then mass customization [139]. Emmatty and Sarmah integrated DFMA into
axiomatic design [116,117], text mining techniques [115], and sensory the platform-based design to optimize the cost of individual products
user experiments [113,114] will play their roles. Nevertheless, all the [140]. Another way to stimulate ideas for alternative configuration
above approaches are still facing challenges. Both the classical MCDM generation is the 40 Inventive Principles of TRIZ. Many studies inte­
methods and axiomatic design require predefined evaluation metrics, grated TRIZ with other design methods, including axiom design, grey
they do not have the adaptability to adjust the evaluation indicators relational analysis, and system dynamics [94,141]. Although the above
once the design/market environment has changed. Fortunately, the methods have been proved effective in many studies and practical cases,
exploration of introducing intelligent algorithms for either automatic they are either empirical principles or mathematical programming
evaluation metric identification or automatic decision-makings have methods, rather than data-driven methods. For data-driven product
been noticed and attempted [112]. Additionally, the text mining tech­ family design, Agard and Kusiak proposed a three-step methodology in
niques are constraint by the decorated textual data actively generated by which data mining algorithm, association rule mining algorithm, and a
users but omitting the other multi-sensory data which reflects users’ true product structure were applied for customer segmentation, customer
behavior or sentiments. The sensory user experiments, however, cannot requirement-functional requirement mapping, and product architecture
quickly response with robust results due to the limitation of strict lab­ design based on product variability [142]. Moon et al. used association
oratory environment settings, long experiment duration, proper subject rule mining technique to cluster the functional features and applied
recruitment, and so on. fuzzy c-means clustering to determine the initial clusters of the modules
[143]. Ma et al. developed a k-means clustering method to group the
6.3. Embodiment design and detail design customer requirements and product architectures [102]. For data-driven
product family evolution, Le et al. developed a generative network for
We discuss embodiment design and detail design together since their product evolution [144].Li et al. used a Bayesian network to analyse the
boundary has become blurred. The integration happened by emphasis­ dynamic relationships between the customer requirements and product
ing concurrent design that contributes to moving decisions forward as architectures [145].
soon as possible, thus shortening product development life cycle time, The other endeavour is to design product families in a “stretched” or
which is enabled by computer-aided engineering [9]. The embodiment “scaled” way, i.e., parametric product family design. For instance,
design and detail design refers to the processes of determining the Simpson et al. proposed a scalable platform called Product Platform
complete specifications of all the parts (e.g., geometry, materials, tol­ Concept Exploration Method (PPCEM) to generate product variants that
erances, etc.) and all the other decisions (e.g., make-or-buy decisions, have same functions with varying capacities [146]. Messac et al. applied
bill of material, manufacturing tools, cost estimation, etc.) [9]. Product physical programming with a product family penalty function to select
family design as well as design exploration and optimization are a few of the the common and scaling parameters for product varieties[147].In the

10
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

scalable product family design, determining optimal values of common convincing explanations will increasing credibility from the product
and distinctive variables by satisfying performance and economic re­ designers/engineers’ point of view. In brief, it is still a far way to develop
quirements is a critical problem. The frequently used methods contain DDPD analysis approaches with humanlike cognitive capabilities.
the linear and non-linear programming algorithms, such as successive
linear programming (SLP), sequential quadratic programming (SQP), 7.2. Large quantity of design concept generation/synthesis within a
and generalized reduced gradient (GRG), as well as the derivative-free limited time
methods, such as genetic algorithm [148], ant colony algorithm
[149,150], and particle swarm optimization algorithm [151]. For The product development process has been accelerated to generate
selecting the proper components from available alternative components, more design concepts within a limited lead time, at lower cost, and
Ilhami et al. applied a non-linear programming model mathematical higher quality [158]. A new wave of intelligent design synthesis has
model to trade off among quality, production capacity, and production become realizable based on AI techniques with learning capability and
manufacturing profit [152]. even generative capability, in which engineers dedicate their creative
Design exploration and optimization is a critical task in configuration efforts to what they are building rather than how to follow the workflow
design and parametric design. Several typical mathematical model- [159]. In particular, KBS that mines and reuses historical design data/
based methods have been deployed for decades, including finite information/knowledge to automatically generate design alternatives
element analysis (FEA), boundary-element method (BEM), and compu­ has played an essential role in fostering quick-response and user-centric
tational fluid dynamics (CFD). All FEA, CFD, and BEM are usually con­ design [160,161]. Besides, generative design strategy combining
ducted within the CAD environment, which requires intensive generative models such as VAE or GAN can significantly improve the
computational demands due to the iterative FEA [34]. To tackle this generative capability meanwhile with high efficiency via iterative al­
drawback, deep learning approaches have been integrated into design gorithms [34,100,130]. But importantly, the supports on the generative
optimization and exploration. For example, Oh et al. [34] integrated design algorithms/approaches onto the conceptual design is still limited.
generative adversarial networks (GAN) with topological optimization, Most studies explore the potential of GD in embodiment design and
generating a large number of 2D design alternatives with limited his­ detail design. Moreover, the generated alternatives are sometimes
torical design data. Yoo et al. extended the 2D generative design method restricted by historical cases with low creativity [25]. To extend the GD
to 3D wheel design using the deep learning approaches [153]. McComb applications onto conceptual design, the core issue is to enrich and
utilized a variational autoencoder (VAE) to accelerate design synthesis manage various product data/knowledge from multidisciplinary do­
and analysis [154]. Similarly, Guo et al. applied convolution neural mains, not only the design but also the other open-sourced domains to
networks (CNN) for the iterative CFD process [155]. In fact, all those explore design stimulus [6,160].
studies follow the generative design paradigm and combined with deep
learning algorithms for performance enhancement, especially GAN. 7.3. Design based on multimodal data
Deep learning algorithms are located at the intersections of many dis­
ciplines, including the detail design in the engineering field, and serve as In the big data era, explosive product-related data, such as customer
the promising enabler for the design automation system. requirements, historical design features and specifications, design con­
straints, production process, and other data, have been vertically inte­
7. Challenges of DDPD grated into systematic product design support tools. However, nearly 80
% of data is unstructured data in the database [162]. Furthermore, they
The crux of DDPD can be found as either the design data/informa­ are usually multimodal data with different representation manner,
tion/knowledge or the algorithmic intelligence. Nevertheless, DDPD still which can be generally divided into five categories, namely (1) pictorial
faces several practical challenges. data such as sketches and engineering drawings, (2) symbolic data such
as assembly trees, (3) linguistic data, for example, customer re­
7.1. Design analysis paradox between effective data analysis and limited quirements, (4) virtual data, for instance, CAD models, and (5) algo­
cognitive capability rithmic data such as parametrizations [163]. Lack of a unified knowledge
representation protocol is the primary concern of the interoperability of
Nowadays, the product design decision-makings need to be accurate the knowledge-based design support tools, especially in the product
and efficient with the least number of resources, including time, money, design stage [163]. Additionally, developments in knowledge discovery
and effort [156]. A product design concept to be practicable must pass techniques to automatically extract valuable information among the
through the tedious, rigid, and iterative product development process, in massive, unstructured, and multi-sourced data will lead to better inte­
which a vast number of design decisions are required to be made. By gration and digitalization of product-related data [163].
exploiting the explosive product through-life data, the design engineers
can harness their design activities with rational analysis by uncovering 7.4. Design based on tacit knowledge
the hidden patterns, exploring new insights, and reusing design
knowledge. However, current prevalent design analysis approaches are Considering that SCPs have become even more intricate, thus making
facing the paradox between effective data analysis capability and product development knowledge-intensive, there is a significant de­
limited cognitive capability. mand for an automatic knowledge-based support tool to reduce product
Specifically, the limited cognitive capability of DDPD analysis development duration and improve design efficiency. Existing DDPD, to
approach is reflected in two aspects, i.e., limited generalization and a large extent, relies on explicit design knowledge [164], such as
limited explainability. On the one hand, the current DDPD analysis ap­ prioritized customer requirements, uncoupled function specifications,
proaches need prescriptive codes and historical design data/informa­ well-organized design constraints, etc. Nevertheless, a large proportion
tion/knowledge for the particular design tasks, which restricts the of design knowledge is still embraced in the personalized and contex­
design outputs within the design variants of the historical designs. On tualized tacit knowledge. Tacit knowledge extraction is still a challenge
the other hand, extant design analysis approaches, such as machine for the next generation of DDPD.
learning approaches or text mining techniques, can hardly offer Contrary to the formal knowledge embedded in product-related
convincing explanations for design engineers. Considering that the early documents, design repositories, decision support tools, and other re­
product design stage decisions will commit 60 %-80 % of the total sources, tacit knowledge refers to the knowledge tied to personal ex­
product’s life cycle cost [157], rational decisions with convincing ex­ periences, intuition, and justification [165]. As argued by Nonaka [165]
planations are promising and necessary. Moreover, the decisions with and verified by plentiful academic experiments [166–168], tacit

11
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

knowledge is deeply rooted in personal actions and minds under specific understand the ontological concepts and logic, and able to reason. And
contexts. In product development, tacit knowledge can be frequently generative capability is regarded as the system’s capability to create or
generated by specialists or skilled technicians or occurred in co-working generate things. With the affective computing and cognitive computing
activities [163]. At present, DDPD based on tacit knowledge is still technologies, the cognitive intelligence-enabled product design tasks
facing two challenges. Initially, although plentiful tacit knowledge contain but not are not limited to:
extraction approaches have been developed in engineering design and
manufacturing [168], major tacit knowledge extraction methods are still (1) Emotional and cognitive communication system for CIPD. Emo­
confined to restricted efficiency and effectiveness by face-to-face in­ tions of human beings gradually become a direct reference index
teractions, long-duration and in-depth interviews, and repeated verifi­ for spiritual world. Emotion cognition will become an important
cation processes occupy a long time for the tacit knowledge acquisition application of the CIPD. The emotion cognition of the current
process. Secondly, few research studies investigate the means of DDPD tasks could achieve the emotion detection or perception
extracting experts’ hidden intents, whereas the intents, indeed, are an evaluation based on the large-scale dataset.
indispensable part of the tacit knowledge [169]. (2) In-depth mining of the implicit customer requirements. By
deploying the affective computing technologies and natural lan­
8. Future perspectives of DDPD in intelligence age guage processing technologies, it is feasible to mine the implicit
customer requirements from the multimodal user-generated data,
This section highlights a couple of potential research directions of such as facial expressions and user comments.
DDPD. (3) Multimodal design inspiration capture. Design inspirations are
usually implicit and fleeting. In the future, it is possible to auto­
8.1. Cognitive intelligence-enabled product design matically capture and record a designer’s inspiration by moni­
toring the status of the design system and designer. Specifically,
Following the data-information-knowledge-wisdom (DIKW) model the design system could recognize the pattern changes in the
[170], the DDPD tasks will be further enhanced by integrating the designer’s physiological states, such as eye movements and EEG
multimodal data and product lifecycle data, exploiting multidisciplinary data.
domain knowledge, and incorporating the cutting-edge interactive (4) Intelligent design based on domain knowledge. By integrating the
technologies such as AR/VR/MR. As a result, a cognitive intelligence- domain knowledge base, inference engine, and query mechanism
enabled product design paradigm could be foreseeable [159,171,172]. into the CAD system, the intelligent design system is empowered
It intends to achieve humanlike capabilities, such as listening and to make decisions based on knowledge and coordinate the data­
speaking, natural language understanding, understanding emotions, and bases/graphic libraries/knowledge bases to complete design
image recognition. Fig. 8 shows the layered intelligent capabilities of tasks with the designers. The intelligent design system is expected
cognitive intelligence. to support the product scheme selection, product architecture
The learning capability should be maintained in the intelligence age design, design exploration and optimization.
since it is the basis of the other advanced intelligent capabilities. With (5) Intelligent design automation: generative design. Generative
learning capability, intelligent algorithms can explore the hidden pat­ design systems have transformed from using evolution algo­
terns on their own [13]. Moreover, advanced algorithmic intelligence rithms or generative grammars [130] to applying deep learning
gradually levels up to an autonomy design paradigm to achieve product algorithms [131,153]. With the support of deep generative
design tasks, including self-aware capability, adaptive capability, models such as GANs and autoencoders, design synthesis is no
cognitive capability, and generative capability. Self-aware capability longer solely based on the changes in technical parameters but
refers to the system’s ability to know the status of itself and the envi­ becomes a multi-objective optimization problem that maximizes
ronment. Adaptive capability indicates the capability to adapt the sys­ the similarity to the market reference designs on a large scale. To
tem’s results/performance when inputs or environments change. From keep the diversity of the generated models, reinforcement
the cognitive capability upwards, the algorithm intelligence promotes learning models could be operational.
an essential change. Cognitive capability means that the system can

Fig. 8. Cognitive intelligence capabilities.

12
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

The key technologies to achieve the cognitive intelligence-enabled autonomous, and continuous evolution [178]. Boschert et al. claimed
product design may focus on human–machine interaction, natural lan­ that the next generation of digital twins should be connected with the
guage processing, sentiment analysis, and computer vision. To further knowledge graphs and product lifecycle management system,
achieve the generative capabilities, generative models such as GAN, will composing a semantically linked digital artifact [179]. Lu and Zheng
be expected for automatic design generation. proposed a formal definition of the cognitive digital twin, which are
“DTs with enhanced semantic capabilities for recognizing the dynamics
8.2. End-to-end design integration of the virtual model evolutions, facilitating understanding of in­
terrelationships between the virtual models and enhancing the decision-
On a system level, a thorough end-to-end design integration should makings” [180]. In a recent study, Al Faruque et al. proposed the
be achieved among humans, machines, SCPs, and other associated ob­ cognitive digital twin framework with cognitive capabilities including
jects [13,173,174]. Two potential research directions are highlighted, perception, attention, memory, reasoning, problem-solving, and
namely the horizontal integrations of design activities based on product learning [181]. An ever-evolving and cognitive product design system
through-lifecycle data to establish a value co-creation design ecosystem could be expected by incorporating the cognitive digital twin with se­
[13,28] and the vertical integration of multimodal and cyber-physical mantic technologies such as ontology modelling, knowledge graph,
data to achieve design contextualization. model-based system engineering, and product lifecycle management
system.
(1) Horizontal end-to-end design integration: End-to-end value chain
(4) AR/VR/MR as the interaction tool for cyber-physical design
The horizontal end-to-end design integration connects the accessible integration
product through-lifecycle data to enable the iterative design process
with higher data transparency and traceability [33,175], composing an Augmented reality, virtual reality, and mixed reality not only merge
end-to-end value chain. By integrating and analyzing the product the 3D digital models into the environment, but also allow the users to
through-lifecycle data, complexity (i.e., more data sources, entities, and interact with the digital content in real-time. Shen et al. demonstrated
stakeholders with complex interactions) and uncertainty (i.e., dynamics, that AR could facilitate the communication and design efficiency among
randomness, and roughness) of the design ecosystem will increase. A a multidisciplinary design team because AR tools can visualize the
basic research question is how to manage complex and uncertain in­ product’s structure and information [182]. Tang et al. developed an
formation in the product design system so that the possibility of experiment to prove that MR tools could improve the users’ under­
designing successful product varieties in the market could be expected. standing of geometric relationships and creativity [183]. Ong and Shen
One possible solution would be monitoring the dynamics in the product developed a system that realizes the bi-directional communication be­
design system, for example, analyzing the dynamics of the customer tween the MR environment and the CAD system [184]. The visualization
requirements, the evolutions of the product family design, the changes and co-modelling of the product models could be achieved among a
in the product function-behaviour-structure frameworks, and the collaborative design team. Uva et al. presented an AR-based framework
transformation of the product modelings [13]. Another possible solution that integrates augmented technical drawings, interactive FEM simula­
is to achieve the adaptability of a product design system so that the tion, multimodal annotation and chat tools, web content integration and
product family design could be easily reconfigurable [176]. Several collaborative client/server architecture [185]. Their proposed frame­
design activities with advanced abilities are enhanced, such as dynamic work could support the technical design review with a tangible inter­
requirement management and rapid (even real-time) reconfigurable face. It is evident that AR/VR/MR is powerful in 3D model visualization
design. and real-time interaction. The product design system can be endowed
with higher-level intelligence by combining AR/VR/MR with other
(2) Vertical end-to-end design integration: Multimodal and cyber- technologies, such as combining with mathematical models, empirical
physical data integration design principles, or knowledge learning models.

The vertical cyber-physical design integration applies multimodal 8.3. Advanced design knowledge support
data collected from multiple sources to describe the same design activ­
ities, hence enriching the design contexts [13]. With rich multimodal As a supporting tool, design knowledge has already become a vital
data, attention and fusion mechanisms can be introduced into the and fundamental resource for design practitioners to smoothly complete
multimodal customer perception analysis, explicit and implicit human design tasks. However, it is still not a worry-free and effortless way to
intention recognition, AR/VR/MR-enabled prototype validation, multi­ manage design knowledge that relies on explosive data.
modal knowledge fusion, and so on. A promising research area would be The approaches to extracting design knowledge from the unstruc­
biometric-based user experience design [177], in which the eye-tracking tured and multi-sourced data will be expected. To the authors’ knowl­
data, behavioural data, brain data, and speeches could be integrated edge, the approaches of design knowledge extraction from textual data
with the design task descriptions and environmental data. Technically, [43,54,160], pictural data [127], and behavioral data [114] have been
the multimodal data integration could be achieved by multimodal ma­ separately attempted in product design. Typical approaches contain text
chine learning techniques. mining [43,54,160], CNNs [127], and human factor experiments [114].
It is significant progress that they have been proved effective in
(3) Cognitive digital twin as the medium of cyber-physical design customer requirement elicitation, sentiment assessment and prediction,
integration conceptual design, etc. Nevertheless, they mainly devote themselves to
the well-developed design tasks that can be derived or adapted from
Both the horizontal design integration and the vertical cyber- historical paradigms/cases. The knowledge management supporting
physical data integration could be achieved via a cognitive digital radical product innovation is still in the infancy stage that is believed
twin. Compared with the digital twin that links the physical data with with further research interests.
the virtual data of a component/product/system, the cognitive digital Moreover, design knowledge support is also expected on massive
twin enhances the computational intelligence to capture and understand tacit knowledge extraction, which is necessary for DDPD [160,186]. KG
the status of a system, as well as how the system interacts with other is a potential and blooming technique that can support design knowl­
components in a complex system [178]. The cognitive digital twin edge management in both well-developed and radical design environ­
presents the characteristics of DT-based, cognitive, full lifecycle-based, ments [186]. It has been successfully applied in personalization

13
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

recommendation [187], biometric and medical field [188] but lacks critical to comprehensively understand the status of DDPD and further
plentiful publications in the product design field. highlight future research potentials in the new Intelligence Age. This
survey investigates this research topic from a holistic view, including
8.4. Design for additive manufacturing bibliometric literature analysis, critical concept clarification, driving
forces summary, current research status analysis, and future research
Additive manufacturing, as a promising production technology, en­ potentials. The main findings of this study are summarized below.
ables to produce more customized and personalized complex products
with lower production costs and shorter product development cycles (1) Clarified the different design concepts in the current design field. This
[189–191]. Design for additive manufacturing that is different with the study compared the concepts of DDPD with three data-related
traditional design for manufacturing and assembly principles has been design terms to clarify the essence of DDPD. Different with
widely studied in much literature [24,190,192–195], including new data-enabled, data-informed, and data-centric product design,
approaches to explore large design spaces [191], meanwhile to incor­ DDPD emphasizes that data is applied as the primary enabler in
porate hybrid design concerns on material, mesostructure [195,196], product design activities.
and multi-scale design. Existing studies mainly focus on three levels, (2) Conducted a bibliometric literature analysis on DDPD. The statistical
namely the part level with macro-scale complexity, the material level analysis results of DDPD-related publications show that the DDPD
with micro-scale complexity, and the product level with multi-scale field is vibrant with a significantly increasing research interest
complexity [189]. At the part level, promising research points lie in since 2015. The top contributing research institutes and journals
the combination of multiple materials and colors, free-form geometric are also listed. China and North America are the primary districts
design, topology optimization, and personalized product design. At the devoting to the DDPD.
material level, the research about design for additive manufacturing (3) Provided a holistic review on DDPD about its driving forces, evolution,
focuses on custom microstructure, custom material composition, custom and status. Through reviewing 172 relevant papers, this study
surface/texture/porosity, and custom lattice/truss/cellular structures outlines three technological driving forces and three business
for better functionality. At the product level, researchers and practi­ driving forces that trigger the development of DDPD. From the
tioners consider component integration, embedded objects/electronics, technology aspect, it is (1) advanced ICTs and digitalization
and direct production of components during the design phase. More technologies, (2) big data, and (3) cognitive computing that
detailed reviews and discussions could be found as [24,189,197]. energize the fundamentals of DDPD. From the business perspec­
tive, it is (1) the exact pursuit of mass personalization, (2) the
8.5. Sustainable smart product-service systems values in big data, and (3) digital integration that motivates the
movements towards DDPD. Furthermore, among the DDPD
Higher sustainability and extended product lifecycle are the remit­ publications, two research interest fluctuations and one recent
ting pursuits of product design in manufacturing companies. To realize a research interest increase were found via domain keywords
sustainable product development process, many research potentials analysis, which are separated as (1) DDPD with IT, (2) DDPD with
have been studied in both academics and practices, including sustain­ CI, and (3) DDPD with digital transformation. In the recent DDPD
able strategies [198], circular systems [199], green design strategies with digital transformation stage, DDPD is evolved with the
[101,200], and product lifecycle management systems [33]. These emergence of new technologies such as additive manufacturing
strategies and approaches aim to effectively reduce non-renewable and cyber-physical system and new design concerns such as
resource consumption and mitigate environmental impacts. To extend product-service systems, uncertainty, and sustainability.
the product lifecycle, product-service system (PSS) appeared as a busi­ (4) Summarized the state-of-the-art of DDPD approaches in the stage
ness paradigm to design the product-service bundles from the product of product planning, concept design, embodiment design and
planning phase. PSS could be further enhanced as a smart PSS paradigm detail design. The widely used data-driven approaches are ma­
based on the widespread ICT infrastructure, digitalization technologies, chine learning and deep learning algorithms in each design
and intelligent algorithms. Product-sensed data and user-generated data phase. Natural language processing and case-based reasoning
could be collected and traceable via smart products [201], which en­ have also been applied for ill-defined problems in product plan­
ables the traceability of the materials/products in the circulation and ning and concept design stages. In the embodiment design,
supports the rapid reconfiguration of future product architectures evolutionary algorithms have been deployed for product family
[104,202]. With hybrid concerns on sustainability, smartness, and PSS, design.
sustainable smart PSS becomes a potential research direction of the (5) Proposed the challenges and future perspectives of DDPD. To fully
DDPD. explore DDPD’s potentials, the authors indicated four challenges
A basic research question of the sustainable smart PSS is what sus­ of current DDPD, namely design analysis paradox between
tainable strategies should be applied in each product design phase to effective data analysis and limited cognitive capability, large
reallocate the cyber and physical resources for higher sustainability quantity of design concept generation/synthesis within a limited
[203]. Another critical research core is how to perceive the multimodal, time, design based on multimodal data, and design based on tacit
large-scale product-sensed data and user-generated data, then encodes knowledge. Towards an advanced DDPD in the Intelligence Age,
them with corresponding design contextual features, achieving the it is recommended that product design practitioners can further
context-awareness of the system [204–207]. As a result, the data anal­ notice the following five possible directions, i.e., cognitive
ysis technologies and context-aware technologies would incorporate intelligence-enabled design, end-to-end design integration,
sustainable strategies such as design for reconfiguration, design for advanced design knowledge support, design for additive
remanufacturing, design for redistribution, design for reuse/recycling, manufacturing, and sustainable smart product-service systems.
and so on [33].
The authors hope this study can be regarded as a reference for design
9. Conclusion practitioners to clarify the entitative meanings of DDPD, understand the
evolvement and status of DDPD, and finally offer some insights for the
DDPD, a powerful design paradigm to conduct design tasks auto­ design practitioners to conduct relevant design tasks in the Intelligence
matically based on big data analysis and computational intelligence, has Age.
been ever-evolvingly developed among academics and industry. With
advanced technological innovations and marketing transformation, it is

14
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

Declaration of Competing Interest Prod. Res. 55 (12) (2017) 3609–3629, https://doi.org/10.1080/

00207543.2017.1308576.
[24] I. Gibson, D. Rosen, B. Stucker, M. Khorasani, Additive manufacturing
The authors declare that they have no known competing financial technologies, Springer, 2014.
interests or personal relationships that could have appeared to influence [25] C. McMahon, Design informatics: supporting engineering design processes with
the work reported in this paper. information technology, J. Indian Inst. Sci. 95 (4) (2015) 365–378.
[26] C.-H. Lim, M.-J. Kim, J.-Y. Heo, K.-J. Kim, Design of informatics-based services in
manufacturing industries: case studies using large vehicle-related databases,
Data availability J. Intell. Manuf. 29 (3) (2018,) 497–508, https://doi.org/10.1007/s10845-015-
1123-8.
[27] Y. Chen, Y. Li. Computational intelligence assisted design: in industrial revolution
No data was used for the research described in the article. 4.0, CRC Press, 2018.
[28] P. Zheng, Z. Wang, C.-H. Chen, L.P. Khoo, A survey of smart product-service
Acknowledgement systems: Key aspects, challenges and future perspectives, Adv. Eng. Inf. 42
(2019), 100973, https://doi.org/10.1016/j.aei.2019.100973.
[29] A.M. Valencia Cardona, R. Mugge, J.P.L. Schoormans, H.N.J. Schifferstein,
The authors wish to acknowledge the funding support from the Na­ Challenges in the design of smart product-service systems (PSSs): Experiences
tional Natural Research Foundation of China (No. 52005424), and from practitioners, in: Proceedings of the 19th DMI: Academic Design
Management Conference, September 2-4 2014, Design Management Institute,
Jiangsu Provincial Policy Guidance Program (Hong Kong/Macau/ London, UK, 2014.
Taiwan Science and Technology Cooperation, BZ2020049). [30] Z. Wang, C.-H. Chen, P. Zheng, Xinyu Li, Li Pheng Khoo, A novel data-driven
graph-based requirement elicitation framework in the smart product-service
system context, Adv. Eng. Inf. 42 (2019), https://doi.org/10.1016/j.
References
aei.2019.100983.
[31] F. Zhou, Y. Ji, R.J. Jiao, Affective and cognitive design for mass personalization:
[1] M.L. Martín-Peña, E. Díaz-Garrido, J.M. Sánchez-López, The digitalization and status and prospect, J. Intell. Manuf. 24 (5) (2013) 1047–1069.
servitization of manufacturing: A review on digital business models, Strategic [32] J. Zhang, P. Gu, Q. Peng, S. Hu, Open interface design for product
Change 27 (2) (2018) 91–99, https://doi.org/10.1002/jsc.2184. personalization, CIRP Ann. 66 (1) (2017) 173–176, https://doi.org/10.1016/j.
[2] N. Urbach, F. Ahlemann, T. Böhmann, The Impact of Digitalization on the IT cirp.2017.04.036.
Department, Bus. Inform. Syst. Eng. 61 (1) (2019) 123–131, https://doi.org/ [33] X. Li, Z. Wang, C.-H. Chen, P. Zheng, A data-driven reversible framework for
10.1007/s12599-018-0570-0. achieving Sustainable Smart product-service systems, J. Cleaner Prod. 279
[3] J. Zhou, Digitalization and intelligentization of manufacturing industry, Adv. (2021), 123618, https://doi.org/10.1016/j.jclepro.2020.123618.
Manuf. 1 (1) (2013) 1–7. [34] S. Oh, Y. Jung, S. Kim, I. Lee, N. Kang, Deep generative design: Integration of
[4] J. Hagberg, M. Sundstrom, N. Egels-Zandén, The digitalization of retailing: an topology optimization and generative models, J. Mech. Des. 141 (11) (2019).
exploratory framework, Int. J. Retail Distrib. Manage. 44 (7) (2016) 694–712, [35] N.A. Kallioras, N.D. Lagaros, DzAIℕ: Deep learning based generative design,
https://doi.org/10.1108/IJRDM-09-2015-0140. Procedia Manuf. 44 (2020) 591–598.
[5] C. Legner, T. Eymann, T. Hess, C. Matt, T. Böhmann, P. Drews, A. Mädche, [36] A. Kumar, From mass customization to mass personalization: a strategic
N. Urbach, F. Ahlemann, Digitalization: Opportunity and Challenge for the transformation, Int. J. Flex. Manuf. Syst. 19 (4) (2007) 533–547.
Business and Information Systems Engineering Community, Bus. Inform. Syst. [37] B.J. Pine, S. Davis. Mass customization: the new frontier in business competition,
Eng. 59 (4) (2017) 301–308, https://doi.org/10.1007/s12599-017-0484-2. Harvard business school press: Boston, 1993.
[6] Y. Feng, Y. Zhao, H. Zheng, Z. Li, J. Tan, Data-driven product design toward [38] M.M. Tseng, J. Jiao, Mass customization, Handbook of industrial engineering 3
intelligent manufacturing: A review, Int. J. Adv. Robot. Syst. 17 (2) (2020), (2001) 684–709.
https://doi.org/10.1177/1729881420911257. [39] C. Chen, CiteSpace II: Detecting and visualizing emerging trends and transient
[7] J. Bacus. Digital Sketching: Computer-Aided Conceptual Design, John Wiley & patterns in scientific literature, J. Am. Soc. Inform. Sci. Technol. 57 (3) (2006)
Sons, 2020. 359–377, https://doi.org/10.1002/asi.20317.
[8] M. Cantamessa, F. Montagna, S. Altavilla, A. Casagrande-Seretti, Data-driven [40] W. Song, Requirement management for product-service systems: Status review
design: the new challenges of digitalization on product design and development, and future trends, Comput. Ind. 85 (2017) 11–22, https://doi.org/10.1016/j.
Des. Sci. 6 (2020) e27. compind.2016.11.005.
[9] G.E. Dieter, L.C. Schmidt, Engineering design, McGraw-Hill Higher Education [41] C. Palomares, X. Franch, C. Quer, P. Chatzipetrou, Lidia López, T. Gorschek, The
Boston, 2009. state-of-practice in requirements elicitation: an extended interview study at 12
[10] A. Sklyar, C. Kowalkowski, B. Tronvoll, D. Sörhammar, Organizing for digital companies, Requir. Eng. 26 (2) (2021), https://doi.org/10.1007/s00766-020-
servitization: A service ecosystem perspective, Journal of Business Research 104 00345-x.
(2019) 450–460, https://doi.org/10.1016/j.jbusres.2019.02.012. [42] A. Jarzębowicz, P. Weichbroth, A qualitative study on non-functional
[11] S. Zhang, J. Xu, H. Gou, J. Tan, A Research Review on the Key Technologies of requirements in agile software development, IEEE Access 9 (2021) 40458–40475,
Intelligent Design for Customized Products, Engineering 3 (5) (2017) 631–640, https://doi.org/10.1109/ACCESS.2021.3064424.
https://doi.org/10.1016/J.ENG.2017.04.005. [43] Z. Wang, C.-H. Chen, P. Zheng, X. Li, L.P. Khoo, A graph-based context-aware
[12] F. Chiarello, P. Belingheri, G. Fantoni, Data science for engineering design: State requirement elicitation approach in smart product-service systems, Int. J. Prod.
of the art and future directions, Comput. Ind. 129 (2021), 103447, https://doi. Res. (2019) 635–651, https://doi.org/10.1080/00207543.2019.1702227.
org/10.1016/j.compind.2021.103447. [44] A. Ferrari, A. Esuli, An NLP approach for cross-domain ambiguity detection in
[13] R. Jiao, S. Commuri, J. Panchal, J. Milisavljevic-Syed, J.K. Allen, F. Mistree, requirements engineering, Automated Softw. Eng. 26 (3) (2019) 559–598,
D. Schaefer, Design Engineering in the Age of Industry 4.0, J. Mech. Des. 143 (7) https://doi.org/10.1007/s10515-019-00261-7.
(2021), https://doi.org/10.1115/1.4051041. [45] Z. Tu, M. Lv, X. Xu, Z. Wang, Crowdsourcing service requirement oriented
[14] R. Deutsch, Data-driven design and construction: 25 strategies for capturing, requirement pattern elicitation method, Neural Comput. Appl. 32 (14) (2019)
analyzing and applying building data, John Wiley & Sons, 2015. 10109–10126, https://doi.org/10.1007/s00521-019-04542-w.
[15] M. Castillo-Vergara, A. Alvarez-Marin, D. Placencio-Hidalgo, A bibliometric [46] S. Bagheri, R.J. Kusters, J.J.M. Trienekens, P.W.P.J. Grefen, A reference model-
analysis of creativity in the field of business economics, Journal of Business based user requirements elicitation process: Toward operational business-IT
Research 85 (2018) 1–9, https://doi.org/10.1016/j.jbusres.2017.12.011. alignment in a co-creation value network, Inform. Softw. Technol. 111 (2019)
[16] H.H.M. Kim, Y. Liu, C.C. Wang, Y. Wang, Special Issue: Data-Driven Design (D3), 72–85, https://doi.org/10.1016/j.infsof.2019.03.012.
J. Mech. Des. 139 (11) (2017), https://doi.org/10.1115/1.4037943. [47] Irina Todoran Koitz, Martin Glinz, A Fuzzy Galois Lattices Approach to
[17] S. Finger, J.R. Dixon, A review of research in mechanical engineering design. Part Requirements Elicitation for Cloud Services, IEEE Trans. Serv. Comput. 11 (5)
I: Descriptive, prescriptive, and computer-based models of design processes, Res. (2015) 768–781, https://doi.org/10.1109/TSC.2015.2466538.
Eng. Des. 1 (1) (1989) 51–67, https://doi.org/10.1007/BF01580003. [48] C.-H. Chen, Li Pheng Khoo, Wei Yan, A strategy for acquiring customer
[18] E. Candelo, Marketing Innovations in the Automotive Industry: Meeting the requirement patterns using laddering technique and ART2 neural network, Adv.
Challenges of the Digital Age, Springer, 2019. Eng. Inf. 16 (3) (2002) 229–240.
[19] G. Farin, Curves and surfaces for computer-aided geometric design: a practical [49] L. Liu, Q. Zhou, J. Liu, Z. Cao, Requirements cybernetics: Elicitation based on user
guide, Elsevier, 2014. behavioral data, J. Syst. Softw. 124 (2017) 187–194.
[20] A. Kusiak, FA. Salustri, Computational Intelligence in Product Design [50] K.-Y. Lin, C.-F. Chien, R. Kerh, UNISON framework of data-driven innovation for
Engineering: Review and Trends, IEEE Trans. Syst., Man, Cybernet., C (Appl. extracting user experience of product design of wearable devices, Comput. Ind.
Rev.) 37 (5) (2007) 766–778, https://doi.org/10.1109/TSMCC.2007.900669. Eng. 99 (2016) 487–502, https://doi.org/10.1016/j.cie.2016.05.023.
[21] Y. Liu, X. Xu, Industry 4.0 and Cloud Manufacturing: A Comparative Analysis, [51] R. Chaklader, M.B. Parkinson, Data-Driven Sizing Specification Utilizing
J. Manuf. Sci. Eng. 139 (3) (2017), https://doi.org/10.1115/1.4034667. Consumer Text Reviews, J. Mech. Des. 139 (11) (2017) pp, https://doi.org/
[22] E. Marilungo, A. Papetti, M. Germani, M. Peruzzini, From PSS to CPS Design: A 10.1115/1.4037476.
Real Industrial Use Case Toward Industry 4.0, Procedia CIRP 64 (2017) 357–362, [52] A. Seitz, M. Grunow, R. Akkerman, Data driven supply allocation to individual
https://doi.org/10.1016/j.procir.2017.03.007. customers considering forecast bias, Int. J. Prod. Econ. 227 (2020), 107683,
[23] Y. Liao, F. Deschamps, E.D.F.R. Loures, L.F.P. Ramos, Past, present and future of https://doi.org/10.1016/j.ijpe.2020.107683.
Industry 4.0 - a systematic literature review and research agenda proposal, Int. J.

15
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

[53] K. Afrin, B. Nepal, L. Monplaisir, A data-driven framework to new product [81] D. Suryadi, H.M. Kim, A Data-Driven Methodology to Construct Customer Choice
demand prediction: Integrating product differentiation and transfer learning Sets Using Online Data and Customer Reviews, J. Mech. Des. 141 (11) (2019),
approach, Expert Syst. Appl. 108 (2018) 246–257, https://doi.org/10.1016/j. https://doi.org/10.1115/1.4044198.
eswa.2018.04.032. [82] B. Song, J. Luo, K. Wood, Data-Driven Platform Design: Patent Data and Function
[54] M.-C. Chiu, K.-Z. Lin, Utilizing text mining and Kansei Engineering to support Network Analysis, J. Mech. Des. 141 (2) (2018), 021101, https://doi.org/
data-driven design automation at conceptual design stage, Adv. Eng. Inf. 38 10.1115/1.4042083.
(2018) 826–839, https://doi.org/10.1016/j.aei.2018.11.002. [83] A. Tufano, R. Accorsi, F. Garbellini, R. Manzini, Plant design and control in food
[55] Y. Zhan, K.H. Tan, B. Huo, Bridging customer knowledge to innovative product service industry. A multi-disciplinary decision-support system, Comput. Ind. 103
development: a data mining approach, Int. J. Prod. Res. 57 (20) (2019) (2018) 72–85, https://doi.org/10.1016/j.compind.2018.09.007.
6335–6350, https://doi.org/10.1080/00207543.2019.1566662. [84] Z. Zhang, L. Liu, W. Wei, F. Tao, T. Li, A. Liu, A Systematic Function
[56] D. Suryadi, H.M. Kim, A Data-Driven Approach to Product Usage Context Recommendation Process for Data-Driven Product and Service Design, J. Mech.
Identification From Online Customer Reviews, J. Mech. Des. 141 (12) (2019), Des. 139 (11) (2017), https://doi.org/10.1115/1.4037610.
https://doi.org/10.1115/1.4044523. [85] D. Ghosh, A. Olewnik, K. Lewis, J. Kim, A. Lakshmanan, Cyber-Empathic Design:
[57] H. Alrezaamiri, A. Ebrahimnejad, H. Motameni, Parallel multi-objective artificial A Data-Driven Framework for Product Design, J. Mech. Des. 139 (9) (2017),
bee colony algorithm for software requirement optimization, Requirements Eng. https://doi.org/10.1115/1.4036780.
25 (3) (2020) 363–380, https://doi.org/10.1007/s00766-020-00328-y. [86] S. Lim, C.S. Tucker, A Bayesian Sampling Method for Product Feature Extraction
[58] M. Ahrens, K. Schneider, Improving requirements specification use by From Large-Scale Textual Data, J. Mech. Des. 138 (6) (2016), https://doi.org/
transferring attention with eye tracking data, Inf. Softw. Technol. 131 (2021), 10.1115/1.4033238.
106483, https://doi.org/10.1016/j.infsof.2020.106483. [87] M. Wang, W. Chen, A Data-Driven Network Analysis Approach to Predicting
[59] Y. Wang, M.M. Tseng, A Naïve Bayes approach to map customer requirements to Customer Choice Sets for Choice Modeling in Engineering Design, J. Mech. Des.
product variants, J. Intell. Manuf. 26 (3) (2015) 501–509. 137 (7) (2015), https://doi.org/10.1115/1.4030160.
[60] Y. Wang, W. Zhao, W.X. Wan, Needs-Based Product Configurator Design for Mass [88] K. Kang, C. Kang, Y.S. Hong, Data-driven optimized vehicle-level engineering
Customization Using Hierarchical Attention Network, IEEE Trans. Autom. Sci. specifications, Ind. Manage. Data Syst. 114 (3) (2014) 338–364, https://doi.org/
Eng. 18 (1) (2021) 195–204, https://doi.org/10.1109/TASE.2019.2957136. 10.1108/IMDS-08-2013-0363.
[61] H. Sun, W. Guo, H. Shao, B.o. Rong, Dynamical mining of ever-changing user [89] C.-F. Chien, R. Kerh, K.-Y. Lin, P.- I. Yu, Data-driven innovation to capture user-
requirements: A product design and improvement perspective, Adv. Eng. Inf. 46 experience product design: An empirical study for notebook visual aesthetics
(2020), 101174, https://doi.org/10.1016/j.aei.2020.101174. design, Comput. Ind. Eng. 99 (2016) 162–173, https://doi.org/10.1016/j.
[62] A.K. Tripathy, P.K. Tripathy, Fuzzy QoS requirement-aware dynamic service cie.2016.07.006.
discovery and adaptation, Appl. Soft Comput. 68 (2018) 136–146. [90] C. Yu, L. Zhu, Product design pattern based on big data-driven scenario, Adv.
[63] Y.T. Chong, C.-H. Chen, Management and forecast of dynamic customer needs: An Mech. Eng. 8 (7) (2016), https://doi.org/10.1177/1687814016656805.
artificial immune and neural system approach, Adv. Eng. Inf. 24 (1) (2010) [91] L. Zhang, X. Chu, H. Chen, B.o. Yan, A data-driven approach for the optimisation
96–106. of product specifications, Int. J. Prod. Res. 57 (3) (2019) 703–721, https://doi.
[64] K.-Y. Lin, User experience-based product design for smart production to empower org/10.1080/00207543.2018.1480843.
industry 4.0 in the glass recycling circular economy, Comput. Ind. Eng. 125 [92] F. Tao, J. Cheng, Q. Qi, M. Zhang, H.e. Zhang, F. Sui, Digital twin-driven product
(2018) 729–738, https://doi.org/10.1016/j.cie.2018.06.023. design, manufacturing and service with big data, Int J Adv Manuf Technol. 94
[65] J. Xie, Y. Bi, Z. Sha, M. Wang, Y. Fu, N. Contractor, L. Gong, W. Chen, Data-Driven (2018) 3563–3576, https://doi.org/10.1007/s00170-017-0233-1.
Dynamic Network Modeling for Analyzing the Evolution of Product Competitions, [93] A. Liu, S.C.Y. Lu, A crowdsourcing design framework for concept generation,
J. Mech. Des. 142 (3) (2020) pp, https://doi.org/10.1115/1.4045687. CIRP Ann. 65 (1) (2016) 177–180, https://doi.org/10.1016/j.cirp.2016.04.021.
[66] S. Lim, C.S. Tucker, Mitigating Online Product Rating Biases Through the [94] Y. Wu, F. Zhou, J. Kong, Innovative design approach for product design based on
Discovery of Optimistic, Pessimistic, and Realistic Reviewers, J. Mech. Des. 139 TRIZ, AD, fuzzy and Grey relational analysis, Comput. Ind. Eng. 140 (2020),
(11) (2017), https://doi.org/10.1115/1.4037612. 106276, https://doi.org/10.1016/j.cie.2020.106276.
[67] H. Voet, M. Altenhof, M. Ellerich, R.H. Schmitt, B. Linke, A Framework for the [95] D. Hwang, W. Park, Design heuristics set for X: A design aid for assistive product
Capture and Analysis of Product Usage Data for Continuous Product concept generation, Des. Stud. 58 (2018) 89–126, https://doi.org/10.1016/j.
Improvement, J. Manuf. Sci. Eng. 141 (2) (2018) pp, https://doi.org/10.1115/ destud.2018.04.003.
1.4041948. [96] W. Yan, C.-H. Chen, M.-D. Shieh, Product concept generation and selection using
[68] P. Zheng, C.-H. Chen, S. Shang, Towards an automatic engineering change sorting technique and fuzzy c-means algorithm, Comput. Ind. Eng. 50 (3) (2006)
management in smart product-service systems–A DSM-based learning approach, 273–285, https://doi.org/10.1016/j.cie.2006.05.003.
Adv. Eng. Inf. 39 (2019) 203–213. [97] M.R. Bohm, J.P. Vucovich, R.B. Stone, Using a Design Repository to Drive
[69] M. Wang, Z. Sha, Y. Huang, N. Contractor, Y. Fu, W. Chen, Predicting product co- Concept Generation, J. Comput. Inf. Sci. Eng. 8 (1) (2008) pp, https://doi.org/
consideration and market competitions for technology-driven product design: a 10.1115/1.2830844.
network-based approach, Des. Sci. 4 (2018), https://doi.org/10.1017/dsj.2018.4. [98] D. Chang, C.-H. Chen, Digital design and manufacturing of wood head golf club in
[70] M. Fargnoli, N. Haber, A practical ANP-QFD methodology for dealing with a cyber physical environment, Ind. Manage. Data Syst. 117 (4) (2017) 648–671,
requirements’ inner dependency in PSS development, Comput. Ind. Eng. 127 https://doi.org/10.1108/IMDS-07-2016-0280.
(2019) 536–548, https://doi.org/10.1016/j.cie.2018.10.042. [99] D. Chang, C.-H. Chen, K.M. Lee, A crowdsourcing development approach based
[71] Z. Chen, X. Ming, X. Zhang, D. Yin, Z. Sun, A rough-fuzzy DEMATEL-ANP method on a neuro-fuzzy network for creating innovative product concepts,
for evaluating sustainable value requirement of product service system, J. Cleaner Neurocomputing 142 (2014) 60–72, https://doi.org/10.1016/j.
Prod. 228 (2019) 485–508, https://doi.org/10.1016/j.jclepro.2019.04.145. neucom.2014.03.044.
[72] Z. Liu, X. Ming, S. Qiu, Q.u. Yuanju, X. Zhang, A framework with hybrid approach [100] S. Khan, M.J. Awan, A generative design technique for exploring shape variations,
to analyse system requirements of smart PSS toward customer needs and co- Adv. Eng. Inf. 38 (2018) 712–724, https://doi.org/10.1016/j.aei.2018.10.005.
creative value propositions, Comput. Ind. Eng. 139 (2020), 105776, https://doi. [101] P. Zheng, Y. Liu, F. Tao, Z. Wang, C.-H. Chen, Smart Product-Service Systems
org/10.1016/j.cie.2019.03.040. Solution Design via Hybrid Crowd Sensing Approach, IEEE Access 7 (2019)
[73] Z. Liu, X. Ming, W. Song, A framework integrating interval-valued hesitant fuzzy 128463–128473, https://doi.org/10.1109/ACCESS.2019.2939828.
DEMATEL method to capture and evaluate co-creative value propositions for [102] J. Ma, H. Kim, Product family architecture design with predictive, data-driven
smart PSS, J. Cleaner Prod. 215 (2019,) 611–625, https://doi.org/10.1016/j. product family design method, Res. Eng. Des. 27 (1) (2016) 5–21, https://doi.
jclepro.2019.01.089. org/10.1007/s00163-015-0201-4.
[74] H.J. Long, L.Y. Wang, J. Shen, M.X. Wu, Z.B. Jiang, Product service system [103] S. Yoo, S. Ha, M. Shin, K. Noh, H. Nam, D. Lee, A Data-Driven Approach for
configuration based on support vector machine considering customer perception, Identifying Medicinal Combinations of Natural Products, IEEE Access 6 (2018)
Int. J. Prod. Res. 51 (18) (2013) 5450–5468. 58106–58118, https://doi.org/10.1109/ACCESS.2018.2874089.
[75] Y. Wang, X. Li, D. Mo, Knowledge-Empowered Multi-Task Learning to Address [104] P. Zheng, X. Xu, C.-H. Chen, A data-driven cyber-physical approach for
the Semantic Gap Between Customer Needs and Design Specifications, IEEE personalised smart, connected product co-development in a cloud-based
Trans. Ind. Inf. (2021) 1, https://doi.org/10.1109/TII.2021.3067141. environment, J. Intell. Manuf. 31 (2020) 3–18, https://doi.org/10.1007/s10845-
[76] Y. Wang, X. Li, Mining Product Reviews for Needs-Based Product Configurator 018-1430-y.
Design: A Transfer Learning-Based Approach, IEEE Trans. Ind. Inf. (2020) 1, [105] C. Ranscombe, P. Kinsella, J. Blijlevens, Data-Driven Styling: Augmenting
https://doi.org/10.1109/TII.2020.3043315. Intuition in the Product Design Process Using Holistic Styling Analysis, J. Mech.
[77] Y. Park, S. Lee, How to design and utilize online customer center to support new Des. 139 (11) (2017), https://doi.org/10.1115/1.4037249.
product concept generation, Expert Syst. Appl. 38 (8) (2011) 10638–10647, [106] C.S. Tucker, H.M. Kim, Data-Driven Decision Tree Classification for Product
https://doi.org/10.1016/j.eswa.2011.02.125. Portfolio Design Optimization, J. Comput. Inf. Sci. Eng. 9 (4) (2009), https://doi.
[78] D. Chang, C. Lee, A product affective properties identification approach based on org/10.1115/1.3243634.
web mining in a crowdsourcing environment, J. Eng. Des. 29 (8–9) (2018) [107] S.L. Mountney, J.X. Gao, S. Wiseall, A knowledge system to support
449–483, https://doi.org/10.1080/09544828.2018.1463514. manufacturing knowledge during preliminary design, Int. J. Prod. Res. 45 (7)
[79] R. Dou, C. Zong, Application of Interactive Genetic Algorithm based on hesitancy (2007/04/01 2007,) 1521–1537, https://doi.org/10.1080/00207540600942300.
degree in product configuration for customer requirement, Int. J. Comput. Int. [108] L. Hou, R.J. Jiao, Data-informed inverse design by product usage information: a
Sys. 7 (sup2) (2014) 74–84. review, framework and outlook, J. Intell. Manuf. 31 (2020) 529–552, https://doi.
[80] J. Shen, B. Wu, L. Yu, Personalized configuration rules extraction in product org/10.1007/s10845-019-01463-2.
service systems by using Local Cluster Neural Network, Ind. Manage. Data Syst.
115 (8) (2015) 1529–1546.

16
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

[109] B. Liu, Y. Zhang, G. Zhang, P. Zheng, Edge-cloud orchestration driven industrial [137] M.V. Martin, K. Ishii, Design for variety: developing standardized and
smart product-service systems solution design based on CPS and IIoT, Adv. Eng. modularized product platform architectures, Res. Eng. Des. 13 (4) (2002)
Inf. 42 (2019), 100984. 213–235, https://doi.org/10.1007/s00163-002-0020-2.
[110] Y. Li, U. Roy, J.S. Saltz, Towards an integrated process model for new product [138] F. Alizon, S.B. Shooter, T.W. Simpson, Improving an existing product family based
development with data-driven features (NPD3), Res. Eng. Des. 30 (2) (2019) on commonality/diversity, modularity, and cost, Des. Stud. 28 (4) (2007)
271–289, https://doi.org/10.1007/s00163-019-00308-6. 387–409, https://doi.org/10.1016/j.destud.2007.01.002.
[111] N.P. Suh, Axiomatic Design: Advances and Applications, Oxford University Press, [139] Q.-J. Peng, Y.-H. Liu, J. Zhang, P. Gu, Personalization for Massive Product
2001. Innovation Using Open Architecture, Chin. J. Mech. Eng. 31 (1) (2018), https://
[112] M.H. Rahman, C. Xie, Z. Sha, Predicting Sequential Design Decisions Using the doi.org/10.1186/s10033-018-0239-0.
Function-Behavior-Structure Design Process Model and Recurrent Neural [140] F.J. Emmatty, S.P. Sarmah, Modular product development through platform-
Networks, J. Mech. Des. 143 (8) (2021), https://doi.org/10.1115/1.4049971. based design and DFMA, J. Eng. Des. 23 (9) (2012) 696–714, https://doi.org/
[113] J.-Y. Kuo, C.-H. Chen, S. Koyama, D. Chang, Investigating the relationship 10.1080/09544828.2011.653330.
between users’ eye movements and perceived product attributes in design [141] J. Delgado-Maciel, G. Cortés-Robles, C. Sánchez-Ramírez, J. García-Alcaraz, J.
concept evaluation, Appl. Ergon. 94 (2021), 103393, https://doi.org/10.1016/j. M. Méndez-Contreras, The evaluation of conceptual design through dynamic
apergo.2021.103393. simulation: A proposal based on TRIZ and system Dynamics, Comput. Ind. Eng.
[114] J.-Y. Kuo, C.-H. Chen, J.R. Roberts, D. Chang, Evaluation of the user emotional 149 (2020), 106785, https://doi.org/10.1016/j.cie.2020.106785.
experience on bicycle saddle designs via a multi-sensory approach, Int. J. Ind. [142] B. Agard, A. Kusiak, Data-mining-based methodology for the design of product
Ergon. 80 (2020), 103039, https://doi.org/10.1016/j.ergon.2020.103039. families, Int. J. Prod. Res. 42 (15) (2004) 2955–2969, https://doi.org/10.1080/
[115] D. Chang, C.-H. Chen, Product concept evaluation and selection using data mining 00207540410001691929.
and domain ontology in a crowdsourcing environment, Adv. Eng. Inf. 29 (4) [143] S.K. Moon, S.R.T. Kumara, T.W. Simpson, Data Mining and Fuzzy Clustering to
(2015) 759–774, https://doi.org/10.1016/j.aei.2015.06.003. Support Product Family Design, in: ASME 2006 International Design Engineering
[116] S. Aydoğan, E.E. Günay, D. Akay, G.E. Okudan Kremer, Concept design evaluation Technical Conferences and Computers and Information in Engineering
by using Z-axiomatic design, Comput. Ind. 122 (2020) 103278, https://doi.org/ Conference, Volume 1: 32nd Design Automation Conference, 2006, pp. 317–325,
10.1016/j.compind.2020.103278. https://doi.org/10.1115/detc2006-99287.
[117] X. Deng, W. Jiang, An Evidential Axiomatic Design Approach for Decision Making [144] Q. Le, Z. Sha, J.H. Panchal, A Generative Network Model for Product Evolution,
Using the Evaluation of Belief Structure Satisfaction to Uncertain Target Values, J. Comput. Inf. Sci. Eng. 14 (1) (2014), https://doi.org/10.1115/1.4025856.
Int. J. Intell. Syst. 33 (1) (2018) 15–32, https://doi.org/10.1002/int.21929. [145] S. Li, W.u. Yongming, X.u. Yanxia, H.u. Jie, H.u. Jianjun, A Bayesian Network
[118] L. Siddharth, N. Madhusudanan, A. Chakrabarti, Toward Automatically Assessing Based Adaptability Design of Product Structures for Function Evolution, Appl. Sci.
the Novelty of Engineering Design Solutions, J. Comput. Inf. Sci. Eng. 20 (1) 8 (4) (2018) 493, https://doi.org/10.3390/app8040493.
(2019), https://doi.org/10.1115/1.4044318. [146] T.W. Simpson, J.R. Maier, F. Mistree, Product platform design: method and
[119] S. Altavilla, F. Montagna, A Product Architecture-Based Framework for a Data- application, Res. Eng. Des. 13 (1) (2001) 2–22, https://doi.org/10.1007/
Driven Estimation of Lifecycle Cost, J. Manuf. Sci. Eng. 141 (5) (2019), https:// s001630100002.
doi.org/10.1115/1.4043195. [147] A. Messac, M.P. Martinez, T.W. Simpson, Introduction of a Product Family
[120] C. Yang, Y. Zou, P. Lai, N. Jiang, Data mining-based methods for fault isolation Penalty Function Using Physical Programming, J. Mech. Des. 124 (2) (2002)
with validated FMEA model ranking, Appl. Intell. 43 (4) (2015/12/01 2015,) 164–172, https://doi.org/10.1115/1.1467602.
913–923, https://doi.org/10.1007/s10489-015-0674-x. [148] Z. Liu, Y.S. Wong, K.S. Lee, A manufacturing-oriented approach for multi-
[121] X. Chen, X. Tao, X. Zeng, L. Koehl, J. Boulenguez-Phippen, Control and platforming product family design with modified genetic algorithm, J. Intell.
optimization of human perception on virtual garment products by learning from Manuf. 22 (6) (2011) 891–907, https://doi.org/10.1007/s10845-009-0365-8.
experimental data, Knowl.-Based Syst. 87 (2015) 92–101, https://doi.org/ [149] R. Kumar, V. Allada, Scalable platforms using ant colony optimization, J. Intell.
10.1016/j.knosys.2015.05.031. Manuf. 18 (1) (2007) 127–142, https://doi.org/10.1007/s10845-007-0009-9.
[122] W.M. Cheung, R. Marsh, L.B. Newnes, A.R. Mileham, J.D. Lanham, Cost data [150] W. Wei, Z. Tian, C. Peng, A. Liu, Z. Zhang, Product family flexibility design
modelling and searching to support low-volume, high-complexity, long-life method based on hybrid adaptive ant colony algorithm, Soft. Comput. 23 (20)
defence system development, P. I. Mech. Eng. B-J. Eng. 229 (5) (2015) 835–846, (2019) 10509–10520, https://doi.org/10.1007/s00500-018-3622-y.
https://doi.org/10.1177/0954405414534226. [151] T.W. Simpson, Product platform design and customization: Status and promise,
[123] L. Wang, X. Chen, D. Henkel, R. Jin, Family learning: A process modeling method Artif. Intell. Eng. Des. Anal. Manuf. 18 (1) (2004) 3–20, https://doi.org/10.1017/
for cyber-additive manufacturing network, IISE Trans. 54 (1) (2022) 1–16, S0890060404040028.
https://doi.org/10.1080/24725854.2020.1851824. [152] M.A. Ilhami, Subagyo, N.A. Masruroh, A mathematical model at the detailed
[124] B. Song, J. Luo, Mining Patent Precedents for Data-Driven Design: The Case of design phase in the 3DCE new product development, Comput. Ind. Eng. 146
Spherical Rolling Robots, J. Mech. Des. 139 (11) (2017), https://doi.org/ (2020), 106617, https://doi.org/10.1016/j.cie.2020.106617.
10.1115/1.4037613. [153] S. Yoo, S. Lee, S. Kim, K.H. Hwang, J.H. Park, N. Kang, Integrating deep learning
[125] W.-T. You, H. Jiang, Z.-Y. Yang, C.-Y. Yang, L.-Y. Sun, Automatic synthesis of into CAD/CAE system: generative design and evaluation of 3D conceptual wheel,
advertising images according to a specified style, Front. Inform. Tech. El. 21 (10) Struct. Multidiscip. Optim. 64 (4) (2021) 2725–2747, https://doi.org/10.1007/
(2020) 1455–1466, https://doi.org/10.1631/FITEE.1900367. s00158-021-02953-9.
[126] C.-H. Lee, C.-H. Chen, F. Li, A.-J. Shie, Customized and knowledge-centric service [154] C. McComb, Toward the Rapid Design of Engineered Systems Through Deep
design model integrating case-based reasoning and TRIZ, Expert Syst. Appl. 143 Neural Networks, in: International Conference On-Design Computing and Cognition,
(2020), 113062, https://doi.org/10.1016/j.eswa.2019.113062. Springer, Cham, 2019, pp. 3–20.
[127] Z. Su, S. Yu, J. Chu, Q. Zhai, J. Gong, H. Fan, A novel architecture: Using [155] X. Guo, W. Li, F. Iorio, Convolutional Neural Networks for Steady Flow
convolutional neural networks for Kansei attributes automatic evaluation and Approximation, in: Proceedings of the 22nd ACM SIGKDD international
labeling, Adv. Eng. Inf. 44 (2020), 101055, https://doi.org/10.1016/j. conference on knowledge discovery and data mining, Association for Computing
aei.2020.101055. Machinery, 2016. doi:10.1145/2939672.2939738.
[128] Z. Wang, X. Li, P. Zheng, C.-H. Chen, L. P. Khoo, Smart product-service system [156] A. Haug, Acquiring materials knowledge in design education, Int. J. Technol. Des.
configuration: a novel hypergraph model-based approach, in: 2020 IEEE 16th Educ. 29 (2) (2019) 405–420, https://doi.org/10.1007/s10798-018-9445-4.
International Conference on Automation Science and Engineering (CASE), 20-21 [157] C.B. Chapman, M. Pinfold, Design engineering—a need to rethink the solution
Aug. 2020, doi: 10.1109/CASE48305.2020.9216890. using knowledge based engineering, Knowl.-Based Syst. 12 (5) (1999) 257–267,
[129] G. Pahl, W. Beitz, Engineering design: a systematic approach, Springer Science & https://doi.org/10.1016/S0950-7051(99)00013-1.
Business Media, 2013. [158] E. Rauch, P. Dallasega, D.T. Matt, The Way from Lean Product Development
[130] S. Krish, A practical generative design method, Comput. Aided Des. 43 (1) (2011) (LPD) to Smart Product Development (SPD), Procedia CIRP 50 (2016) 26–31,
88–100, https://doi.org/10.1016/j.cad.2010.09.009. https://doi.org/10.1016/j.procir.2016.05.081.
[131] Z. Nie, T. Lin, H. Jiang, L.B. Kara, TopologyGAN: Topology Optimization Using [159] K.K.H. Ng, C.-H. Chen, C.K.M. Lee, R.J. Jiao, Z.-X. Yang, A systematic literature
Generative Adversarial Networks Based on Physical Fields Over the Initial review on intelligent automation: Aligning concepts from theory, practice, and
Domain, J. Mech. Des. 143 (3) (2021), https://doi.org/10.1115/1.4049533. future perspectives, Adv. Eng. Inf. 47 (2021), 101246, https://doi.org/10.1016/j.
[132] K. Xing, H.F. Wang, W. Qian, A sustainability-oriented multi-dimensional value aei.2021.101246.
assessment model for product-service development, Int. J. Prod. Res. 51 (19) [160] X. Li, C.-H. Chen, P. Zheng, Z. Wang, Z.H. Jiang, Z. Jiang, A Knowledge Graph-
(2013) 5908–5933. Aided Concept-Knowledge Approach for Evolutionary Smart Product-Service
[133] J. Jiao, T.W. Simpson, Z. Siddique, Product family design and platform-based System Development, J. Mech. Des. 142 (10) (2020) pp, https://doi.org/10.1115/
product development: a state-of-the-art review, J. Intell. Manuf. 18 (1) (2007) 1.4046807.
5–29, https://doi.org/10.1007/s10845-007-0003-2. [161] Z.Y. Wu, X.G. Ming, L.N. He, M. Li, X.Z. Li, Knowledge integration and sharing for
[134] X. Du, J. Jiao, M.M. Tseng, Architecture of Product Family: Fundamentals and complex product development, Int. J. Prod. Res. 52 (21) (2014) 6296–6313. htt
Methodology, Concurrent Eng. 9 (4) (2001) 309–325, https://doi.org/10.1177/ p://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=99001021&site
1063293X0100900407. =ehost-live.
[135] Z. Liu, Y.S. Wong, K.S. Lee, Modularity analysis and commonality design: a [162] T.K. Das, P.M. Kumar, Big data analytics: A framework for unstructured data
framework for the top-down platform and product family design, Int. J. Prod. Res. analysis, Int. J. Eng. Sci. Tech. 5 (1) (2013) 153.
48 (12) (2010) 3657–3680, https://doi.org/10.1080/00207540902902598. [163] S.K. Chandrasegaran, K. Ramani, R.D. Sriram, I. Horváth, A. Bernard, R.F. Harik,
[136] K. Baylis, G. Zhang, D.A. McAdams, Product family platform selection using a W. Gao, The evolution, challenges, and future of knowledge representation in
Pareto front of maximum commonality and strategic modularity, Res. Eng. Des. product design systems, Comput. Aided Des. 45 (2) (2013) 204–228, https://doi.
29 (4) (2018) 547–563, https://doi.org/10.1007/s00163-018-0288-5. org/10.1016/j.cad.2012.08.006.

17
Z. Wang et al. Advanced Engineering Informatics 54 (2022) 101793

[164] X. Chen, C.-H. Chen, K.F. Leong, X. Jiang, An ontology learning system for [188] Guan Sai-Ping, Xiao-Long Jin, Yan-Tao Jia, Yuan-Zhuo Wang, and Xue-Qi Cheng,
customer needs representation in product development, Int. J. Adv. Manuf. Tech. “Knowledge graph oriented knowledge inference methods: A survey,” (in
67 (1) (2013) 441–4533. Chinese), Ruan Jian Xue Bao/Journal of Software, vol. 29, p. 10, 2018 2018.
[165] I. Nonaka, H. Takeuchi, The knowledge-creating company, Harvard Bus. Rev. 85 [Online]. Available: http://www.jos.org.cn/1000-9825/5551.htm.
(7/8) (2007), 162. [189] M.K. Thompson, et al., Design for Additive Manufacturing: Trends, opportunities,
[166] F.H. Wang, Z.H. Jiang, X. Li, G. Li, Cognitive factors of the transfer of empirical considerations, and constraints, CIRP Ann. 65 (2) (2016) 737–760, https://doi.
engineering knowledge: A behavioral and fNIRS study, Adv. Eng. Inf. 47 (2021), org/10.1016/j.cirp.2016.05.004.
101207, https://doi.org/10.1016/j.aei.2020.101207. [190] Z. Doubrovski, J.C. Verlinden, J.M.P. Geraedts, Optimal Design for Additive
[167] X. Li, Z.H. Jiang, B.o. Song, L. Liu, Long-term knowledge evolution modeling for Manufacturing: Opportunities and Challenges, in: ASME 2011 International
empirical engineering knowledge, Adv. Eng. Inf. 34 (2017) 17–35, https://doi. Design Engineering Technical Conferences and Computers and Information in
org/10.1016/j.aei.2017.08.001. Engineering Conference, Volume 9: 23rd International Conference on Design
[168] L. Liu, Z.H. Jiang, B.o. Song, H. Zhu, X. Li, A Novel Method for Acquiring Theory and Methodology; 16th Design for Manufacturing and the Life Cycle
Engineering-Oriented Operational Empirical Knowledge, Math. Probl. Eng. Conference, 2011, pp. 635–646, https://doi.org/10.1115/detc2011-48131.
(2016), 9754298, https://doi.org/10.1155/2016/9754298. [191] Y.i. Xiong, et al., Data-Driven Design Space Exploration and Exploitation for
[169] J. Liu, X. Hu, A reuse oriented representation model for capturing and formalizing Design for Additive Manufacturing, J. Mech. Des. 141 (10) (2019) pp, https://doi.
the evolving design rationale, AI EDAM 27 (4) (2013) 401–413, https://doi.org/ org/10.1115/1.4043587.
10.1017/S0890060413000395. [192] G.A.O. Adam, D. Zimmer, Design for Additive Manufacturing—Element
[170] J. Rowley, The wisdom hierarchy: representations of the DIKW hierarchy, Journal transitions and aggregated structures, CIRP J. Manuf. Sci. Technol. 7 (1) (2014)
of Information Science 33 (2) (2007) 163–180, https://doi.org/10.1177/ 20–28, https://doi.org/10.1016/j.cirpj.2013.10.001.
0165551506070706. [193] B. Vayre, F. Vignat, F. Villeneuve, Identification on Some Design Key Parameters
[171] J.E. Kelly III, S. Hamm, Smart machines: IBM’s Watson and the era of cognitive for Additive Manufacturing: Application on Electron Beam Melting, Procedia
computing, Columbia University Press, 2013. CIRP 7 (2013) 264–269, https://doi.org/10.1016/j.procir.2013.05.045.
[172] J.E. Kelly, Computing, cognition and the future of knowing, Whitepaper, IBM [194] B. Vayre, F. Vignat, F. Villeneuve, Designing for Additive Manufacturing,
Reseach 2 (2015). Procedia CIRP 3 (2012) 632–637, https://doi.org/10.1016/j.procir.2012.07.108.
[173] M.E. Porter, J.E. Heppelmann, How smart, connected products are transforming [195] C. Chu, G. Graf, D.W. Rosen, Design for Additive Manufacturing of Cellular
companies, Harvard Bus. Rev. 93 (10) (2015) 96–114. Structures, Comput.-Aided Des. Applic. 5 (5) (2008/01/01 2008,) 686–696,
[174] M. Abramovici, J.C. Göbel, H.B. Dang, Semantic data management for the https://doi.org/10.3722/cadaps.2008.686-696.
development and continuous reconfiguration of smart products and systems, CIRP [196] D.W. Rosen, Computer-Aided Design for Additive Manufacturing of Cellular
Ann. 65 (1) (2016) 185–188, https://doi.org/10.1016/j.cirp.2016.04.051. Structures, Comput.-Aided Des. Applic. 4 (5) (2007) 585–594, https://doi.org/
[175] A. Whitmore, A. Agarwal, X.u. Li Da, The Internet of Things—A survey of topics 10.1080/16864360.2007.10738493.
and trends, Inform. Syst. Front. 17 (2) (2015) 261–274, https://doi.org/10.1007/ [197] Y. Huang, M.C. Leu, J. Mazumder, A. Donmez, Additive manufacturing: Current
s10796-014-9489-2. state, future potential, gaps and needs, and recommendations, J. Manuf. Sci. E-T
[176] H. ElMaraghy, L. Monostori, G. Schuh, W. ElMaraghy, Evolution and future of ASME 137 (1) (2015), 014001, https://doi.org/10.1115/1.4028725.
manufacturing systems, CIRP Ann. 70 (2) (2021) 635–658, https://doi.org/ [198] A. Tukker, Eight types of product–service system: eight ways to sustainability?
10.1016/j.cirp.2021.05.008. Experiences from SusProNet, Business Strategy and the Environment 13 (4)
[177] Y. Borgianni, E. Rauch, L. Maccioni, B.G. Mark, User Experience Analysis in (2004) 246–260, https://doi.org/10.1002/bse.414.
Industry 4.0 - The Use of Biometric Devices in Engineering Design and [199] A. Tukker, Product services for a resource-efficient and circular economy – a
Manufacturing. 2018 IEEE International Conference on Industrial Engineering review, J. Cleaner Prod. 97 (2015) 76–91, https://doi.org/10.1016/j.
and Engineering Management (IEEM), IEEE, 2018. jclepro.2013.11.049.
[178] X. Zheng, J. Lu, D. Kiritsis, The emergence of cognitive digital twin: vision, [200] P. Zheng, Z. Wang, C.-H. Chen, Smart product-service systems: A novel
challenges and opportunities, Int. J. Prod. Res. (2021) 1–23, https://doi.org/ transdisciplinary sociotechnical paradigm, in: 26th ISTE International Conference
10.1080/00207543.2021.2014591. on Transdisciplinary Engineering, Tokyo, Japan, 30 July - 1 August 2019, vol. 10,
[179] S. Boschert, C. Heinrich, R. Rosen, Next generation digital twin, in: Proc. tmce, in Advances in Transdisciplinary Engineering, pp. 234-241, doi: 10.3233/
Las Palmas de Gran Canaria, Spain, 2018, pp. 7–11. ATDE190128.
[180] J. Lu, X. Zheng, A. Gharaei, K. Kalaboukas, D. Kiritsis, Cognitive Twins for [201] M.F. Filho, Y. Liao, E.R. Loures, O. Canciglieri, Self-Aware Smart Products:
Supporting Decision-Makings of Internet of Things Systems, in: Proceedings of 5th Systematic Literature Review, Conceptual Design and Prototype Implementation,
International Conference on the Industry 4.0 Model for Advanced Manufacturing. Procedia Manufacturing 11 (2017) 1471–1480, https://doi.org/10.1016/j.
Springer, Cham, 2020. promfg.2017.07.278.
[181] M.A. Al Faruque, D. Muthirayan, S.-Y. Yu, P.P. Khargonekar, Cognitive digital [202] P. Zheng, Y. Lin, C.-H. Chen, X. Xu, Smart, connected open architecture product:
twin for manufacturing systems, in: 2021 Design, Automation & Test in Europe an IT-driven co-creation paradigm with lifecycle personalization concerns, Int. J.
Conference & Exhibition (DATE), IEEE, 2021. Prod. Res. 57 (8) (2018) 1–14.
[182] Y. Shen, S.K. Ong, A.Y.C. Nee, Augmented reality for collaborative product design [203] A. Alcayaga, M. Wiener, E.G. Hansen, Towards a framework of smart-circular
and development, Des. Stud. 31 (2) (2010) 118–145, https://doi.org/10.1016/j. systems: An integrative literature review, J. Cleaner Prod. 221 (2019) 622–634,
destud.2009.11.001. https://doi.org/10.1016/j.jclepro.2019.02.085.
[183] Y.M. Tang, K.M. Au, Yohana Leung, Comprehending products with mixed reality: [204] Y. Ji, Z.H. Jiang, X. Li, Y. Huang, F.H. Wang, A multitask context-aware approach
Geometric relationships and creativity, Int. J. Eng. Bus. Manage. 10 (2018), for design lesson-learned knowledge recommendation in collaborative product
1847979018809599, https://doi.org/10.1177/1847979018809599. design, J. Intell. Manuf. 1 (2022) 1–23, https://doi.org/10.1007/s10845-021-
[184] S.K. Ong, Y. Shen, A mixed reality environment for collaborative product design 01889-7.
and development, CIRP Ann. 58 (1) (2009) 139–142, https://doi.org/10.1016/j. [205] X. Li, C.-H. Chen, P. Zheng, Z.H. Jiang, L. Wang, A context-aware diversity-
cirp.2009.03.020. oriented knowledge recommendation approach for smart engineering solution
[185] A.E. Uva, S. Cristiano, M. Fiorentino, G. Monno, Distributed design review using design, Knowl.-Based Syst. 215 (2021), 106739, https://doi.org/10.1016/j.
tangible augmented technical drawings, Comput. Aided Des. 42 (5) (2010) knosys.2021.106739.
364–372, https://doi.org/10.1016/j.cad.2008.10.015. [206] Z. Wang, C.-H. Chen, X. Li, P. Zheng, L.P. Khoo, A context-aware concept
[186] X. Li, M. Lyu, Z. Wang, C.-H. Chen, P. Zheng, Exploiting knowledge graphs in evaluation approach based on user experiences for smart product-service systems
industrial products and services: A survey of key aspects, challenges, and future design iteration, Adv. Eng. Inf. 50 (2021), 101394, https://doi.org/10.1016/j.
perspectives, Comput. Ind. 129 (2021), 103449, https://doi.org/10.1016/j. aei.2021.101394.
compind.2021.103449. [207] A. Huet, R. Pinquié, P. Véron, A. Mallet, F. Segonds, CACDA: A knowledge graph
[187] J. Ren, J. Long, X.u. Zhikang, Financial news recommendation based on graph for a context-aware cognitive design assistant, Comput. Ind. 125 (2021), 103377,
embeddings, Decis. Support Syst. 125 (2019), 113115, https://doi.org/10.1016/j. https://doi.org/10.1016/j.compind.2020.103377.
dss.2019.113115.

18