(PDF) Interoperable manufacturing knowledge systems

International Journal of Production Research Fo rP ee rR ev ie w On Interoperable manufacturing knowledge systems ly Journal: International Journal of Production Research Manuscript ID TPRS-2017-IJPR-0267.R1 Manuscript Type: Special Issue Paper Date Submitted by the Author: 04-Jul-2017 Complete List of Authors: Palmer, Claire; Loughborough University, Wolfson School of Mech, Elect and Man Engineering Usman, Zahid; University of Coventry, UK, Faculty of Engineering, http://mc.manuscriptcentral.com/tprs Email:

Page 1 of 24 International Journal of Production Research 1 2 3 4 Environment and Computing Canciglieri Junior, Osiris; Pontifical Catholic University of Paraná 5 Malucelli, Andreia; Pontifical Catholic University of Paraná 6 Young, Robert; Loughborough University, Wolfson School of Mech, Elect 7 and Man Engineering 8 9 INTEROPERABILITY, INTELLIGENT MANUFACTURING SYSTEMS, 10 Keywords: KNOWLEDGE ENGINEERING, MANUFACTURING INFORMATION SYSTEMS, 11 ONTOLOGIES 12 decision support, reference ontologies, manufacturing systems, knowledge 13 Keywords (user): sharing 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 2 of 24 1 2 3 4 Interoperable manufacturing knowledge systems 5 6 7 8 Claire Palmera, Zahid Usmanb, Osiris Canciglieri Juniorc, Andreia Malucellic, 9 Robert I.M.Younga. 10 a 11 Loughborough University, Loughborough, Leicestershire, UK 12 b 13 Coventry University, Coventry, UK 14 c Pontifical Catholic University of Paraná, Curitiba, Brazil Fo 15 16 17 18 rP 19 Corresponding author: 20 21 22 Prof R I M Young, (

), Wolfson school of mechanical, electrical ee 23 and manufacturing engineering, Loughborough University, Loughborough, LE11 3TU, 24 25 UK. 26 rR 27 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 3 of 24 International Journal of Production Research 1 2 3 4 5 Interoperable manufacturing knowledge systems 6 7 8 ABSTRACT. For many years now, the importance of semantic technologies, that provide a formal, logic based route to 9 sharing meaning, has been recognized as offering the potential to support interoperability across multiple related 10 applications and hence drive manufacturing competitiveness in the digital manufacturing age. However, progress in 11 support of manufacturing enterprise interoperability has tended to be limited to fairly narrow domains of 12 applicability. This paper presents a progression of research and understanding, culminating in the work undertaken in the recent EU FLEXINET project, to develop a comprehensive manufacturing reference ontology that can (a) 13 support the clarification of understanding across domains, (b) support the ability to flexibly share information across 14 interacting software systems and (c) provide the ability to readily configure company knowledge bases to support Fo 15 interoperable manufacturing systems. 16 17 18 KEYWORDS: interoperability, decision support, reference ontologies, manufacturing systems, knowledge sharing. rP 19 20 21 1. Introduction 22 Competitive manufacturing industry must be able to react to change and to understand ee 23 24 the balance of possible options when making decisions on complex multi-faceted 25 problems. To be able to make high quality, timely decisions across a range of complex 26 factors that impact manufacturing decision making requires high quality information rR 27 and knowledge to be available at the time key decisions are made. However each 28 business domain requires its own specific discipline’s view of the necessary information 29 that it needs. While current software solutions are very good a providing local domain 30 support they do not provide the well-integrated, trans-disciplinary and holistic ev 31 approaches that are critical to long-term competitive solutions (Huber 2014). 32 33 The effective exploitation of semantic technologies has the potential to solve a 34 critical part of this problem through the provision of a formal logic based route to ie 35 sharing meaning across multiple domains (Borgo et al. 2007, Chungoora et al, 2012). 36 These, through the application of formal logic, have the potential to provide 37 substantially more comprehensive solutions than the industrial data standard approaches w 38 that have been employed to date (Chungoora et al, 2013). 39 40 The paper begins by providing the background to the problem and argues On 41 progressively for the need for a manufacturing reference ontology, the methods needed 42 to construct and use such an ontology and the future work that is still needed to fully 43 exploit the approach to provide significant industrial benefit. 44 45 ly 46 2 Background to the problem 47 48 49 2.1 A business perspective 50 51 While there are many different manufacturing sectors they all have the same sorts of 52 problems with information, most especially with access to information and the sharing 53 of information. Information to support decisions in manufacturing comes from many 54 different sources and is needed to support many different types of decisions. Each group 55 within a manufacturing business, such as in design, production, maintenance, 56 procurement and sustainability, will have it’s own types of information. Different 57 factories, customers, suppliers and systems will work with different information and 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 4 of 24 1 2 3 4 different global constraints will have importance dependent on global supplier and 5 market locations. 6 Each business area of activity needs multiple sources of knowledge to best 7 support decisions, with the efforts towards improving concurrency in product design 8 being a very good example that highlights the need for not only design functional 9 knowledge but also design for manufacture and assembly, design for operation, design 10 11 for sustainability and even business and market knowledge. 12 A critical business requirement is to ensure that any potential problems in the 13 ability of new systems to interoperate within complex, flexible, scalable and re- 14 configurable software environments can be clearly identified in order to ensure Fo 15 16 interoperable solutions. This is essential in minimising the substantial cost and time loss 17 implications of introducing new systems that are incompatible with the holistic 18 requirements of the business. A clear understanding of the implications of change is of rP 19 major importance when operating in a dynamic environment with on-going cost, time 20 and quality pressures in globally competitive markets. Providing solutions to meet this 21 need are very much in line with the present day drive towards Industrie 4.0, Smart 22 Manufacturing and the exploitation of cyber-physical systems (Kagermann et al, 2013). ee 23 24 25 2.2 An ICT perspective 26 rR 27 Manufacturing industry employs an extensive range of software tools to support their 28 business activities. A simple illustration of a some of the main tools are also shown in 29 30 figure 1: with Product Lifecycle Management (PLM), Enterprise Resource Planning (ERP), Supply Chain Management (SCM), Manufacturing Execution Systems (MES) ev 31 32 Supervisory Control and Data Acquisition (SCADA) along with specific machine 33 control systems providing perhaps the most significant coverage from high level design 34 and tactical planning activities through to shop floor machine control and monitoring. ie 35 Figure 1 also indicates the complexity of the problem of sharing information and 36 knowledge in a meaningful way. For example the arrows related to the computational 37 w systems indicate an ideal situation where information can be communicated in all 38 directions. This is not the case in reality. 39 40 These manufacturing software support tools each provide valuable information On 41 support to decision makers but they are not well integrated and traditional methods of 42 integration, via interfacing, are expensive and error prone (Ray & Jones, 2006). The use 43 of industrial data standards can be helpful, but currently these tend to be focused on 44 narrow domains and are not well suited to the breadth of information sharing that needs 45 ly to be managed across manufacturing (Chungoora et al, 2013). Perhaps ISA-95 46 (IEC/ISO 62264) provides the most useful reference standard towards the interoperation 47 of manufacturing operations tools, but that too is limited by the lack of semantic 48 49 consistence in interpretation. For example (Hastilow, 2013) illustrates the problem of 50 misalignment of interpretation between users of ISA-95, identifying 83 errors in a real 51 world example and demonstrating how these could have been avoided if formal 52 semantic approaches had been used. The need for improved standardisation 53 approaches in industrial data standards has been recognised through the work of ISO 54 TC184 SC4 on “Industrial Data Integrated Ontologies and Models (IDIOM)” (Leal & 55 Feeney, 2010) and also in the standardisation task force report of the EU FInES cluster 56 (Pattenden, Young & Zelm, 2012). 57 Finding a robust and consistently interpretable way of representing information 58 such that it can be reliably shared is a substantial problem. Formal logic based methods 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 5 of 24 International Journal of Production Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 Figure 1: Elements of a Manufacturing Ecosystem 25 26 for the definition of ontologies should offer radical improvements in the ability to share rR 27 information and knowledge, or at least to clearly understand what is and is not sharable. 28 The most effective way to exploit these technologies is still evolving. This paper 29 30 provides an input to this in terms of manufacturing reference ontologies and how in the future such ontologies should be developed and exploited to best effect. ev 31 32 33 34 ie 35 2.3 An ontological perspective 36 37 w Ontologies are used in the computing world for the purposes of communication, 38 computational inferences, information & knowledge organization, exchange and reuse 39 (Gruninger and Lee 2002). An ontology, as a common basis for shared meaning, can be 40 used for the purpose of knowledge sharing and interoperability across multiple domains. On 41 Ontologies in concurrent engineering can support interoperability between systems and 42 support product design and software development (Roche, 2000). Ontologies can play a 43 pivotal role in knowledge management by providing a better way of representing 44 knowledge and supporting the development of reusable and shareable knowledge bases 45 (Sureephong 2008). ly 46 47 Initial approaches to semantic interoperability have been achieved through 48 ontology matching approaches (Shvaiko and Euzenat, 2013; Zdravković et al, 2011; Ye 49 et al, 2008), interchange formats (Ushold et al, 1998; Chen et al, 2009) and reference 50 ontologies (Annamalai et al, 2011; Foufou et al, 2005; Borsato, M., 2014; Chungoora 51 and Young, 2011). These works start to show the potential benefits that can be gained 52 from the use of formal logic but none of them meet the complex knowledge sharing 53 requirements of manufacturing businesses, where all the key aspects of the business 54 must interact and be able to share information, covering business model development 55 56 and maintenance, new product and process development, production planning and 57 operation, and on through to product operation, service and end-of-life management. 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 6 of 24 1 2 3 4 The aim of manufacturing reference ontologies is to provide an underpinning 5 formal semantic structure that can meet the above need,supporting the development of 6 flexible systems that can share manufacturing knowledge across the multiple 7 manufacturing related domains. This approach, described in the next sub-section, 8 effectively sit between domain ontologies, which are very specific and foundation 9 ontologies such as DOLCE, SUMO, BFO (Mascardi et al., 2010) which are very 10 generic. For example BFO provides formal semantic definitions for terms such as 11 12 Continuants such as Material and Information entities and Occurrents such as Processes. 13 BFO, as other foundation ontologies, leaves the user to specialise these terms into more 14 specific meaningful terms for their own use. In the case of manufacturing the aim in our Fo 15 work is to exploit this approach to provide the core semantics for manufacturing that 16 enables cross-disciplinary sharing of knowledge. i.e. to provide the core semantics for 17 manufacturing that can then be referenced and further specialised to support the multi- 18 context domains needed in a manufacturing organisation. The idea of a reference rP 19 ontology fits with the integration model concepts of the “IIDEAS” architecture, with 20 foundation concepts that are used to build general concepts that are then used to build 21 22 specific concepts (West and Fowler, 2001), but with the important addition of the use of formal logic. ee 23 24 A broad range of research has contributed to manufacturing ontology 25 understanding with a useful review provided in (Usman et al., 2013). Since then further 26 examples can be seen in (Borsato, 2014) related to sustainable manufacturing, (Bruno et rR 27 28 al., 2016) related to Product Lifecycle Management, and (Palmer et al., 2017) related to 29 global product-service production. In addition a vision statement from the International 30 Federation of Automatic Control (IFAC) Technical Committee 5.3, Enterprise Integration and Networking suggests that a common core ontology is needed to support ev 31 32 interoperability between different models within smart sensing enterprise systems 33 (Weichhart et al., 2016). 34 ie 35 36 3 The reference ontology approach to interoperable manufacturing knowledge 37 systems w 38 39 40 The reference ontology approach takes the view that we need an ontological development method that can support information sharing across multiple domain On 41 42 perspectives to provide a level of semantic consistency across these perspectives. It 43 follows the view that if any pair of domain ontologies should share information then 44 there should 45 ly 46 47 48 49 50 51 52 53 54 55 56 Figure 2: An overview of the specialisation approach 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 7 of 24 International Journal of Production Research 1 2 3 be a higher-level ontology that supports these ontologies. Also, that the highest level, 4 there will be a foundation ontology that can be exploited to support reference ontologies 5 for other business sectors e.g. healthcare. Figure 2 provides an illustration of this basic 6 specialisation approach. 7 8 From this basic approach the resulting reference ontology can then be readily 9 specialised to suit the needs of any domain and/or mapped to any specific companies 10 11 needs in order to both interoperably link multiple software applications and to construct 12 a manufacturing knowledge base for the company concerned. This concept is illustrated 13 in figure 3. 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 Figure 3: Exploiting a reference ontology for KB construction and software On 41 interoperability 42 43 In our research we have employed both a top-down approach and a bottom-up approach 44 to identify both the core concepts that support information sharing across domains and 45 the specific concepts that are used in a domain. In the bottom up approach specific ly 46 domains are investigated and used to identify, through analyse of their 47 interrelationships, the core concepts that apply to both. In the top down approach a view 48 of the necessary core concepts is proposed and then specialised to suit the needs of each more specific domain. 49 50 51 52 3.1 New understanding from a bottom-up approach 53 54 In our research we have explored the development and exploitation of reference 55 ontologies in a number of use cases. These are: 56 57 1. Product Design / Design for Machining (Usman et al., 2013) 58 2. Design for Assembly /Assembly Planning (Imran and Young, 2013) 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 8 of 24 1 2 3 4 3. The introduction of new shop floor systems into an already integrated 5 environment. (Hastilow and Young, 2012). 6 4. Product Design / Design for Manufacture of plastic injection moulded parts 7 (Szeika, 2016) 8 5. Global supply network configuration to meet new business opportunities. (Young et al., 2014) 9 10 In each case the approach was based on the guidelines of Blomqvist and Ohgren 11 (2008) and Noy and McGuiness (2000). Industrial inputs were used to develop a view 12 of knowledge domains of interest and built an understanding of the key concepts and 13 relationships involved. From that, and with an awareness of already existing ontologies, 14 a reference ontology was progressively evolved and tested to support the particular Fo 15 16 domains of interest. The following theses (Usman 2012; Imran 2015; Hastilow 2013; 17 Szejka 2016) provide the full development and evaluation of the cases listed above with 18 the exception of item five which is discussed in section 5. A simple illustration of the rP 19 first three of these is provided in figure 4, not to show any detail, but to highlight that 20 each one is different, even though there was a general aim to build a common reference 21 ontology. This highlights the difficultly of defining a common reference ontology as 22 each targeted area of activity has their own perception of what is needed even though ee 23 they agree on the same overall intention. A significant number iterations are likely to be 24 needed before a truly industrially effective manufacturing reference ontology can be 25 26 defined. rR 27 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 Figure 4: A graphical illustration of three reference ontologies. 49 50 51 These differences in results are important to understand if a common manufacturing 52 reference ontology is ever to be achieved. The bottom up approach provides a clear set 53 of requirements to be met by each use case. In the examples provided the only 54 constraint that is common across the examples is the use of a common foundation 55 ontology. Each of these research activities used the Upper Level Ontology (ULO) of 56 Highfleet (Highfleet, 2014). The high level semantic constraints that this provides are 57 insufficient to ensure any significant consistency across the reference ontologies. There 58 is therefore a need to identify a more constrained or structured way of building 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 9 of 24 International Journal of Production Research 1 2 3 reference ontologies if they are to be commonly exploitable. If a reference ontology 4 needs to be revised rather than extended then it cannot be commonly exploitable. This 5 emphasises the need for an iterative approach to reference ontology development as 6 well as the need for a combined top-down and bottom-up approach as is proposed in the 7 next section. 8 In addition to the work above significant progress, through the use of highly structured 9 ontology development processes, has been made in the medical ontology development 10 area (Arp et al, 2015). The application of this approach to manufacturing through the 11 development of an Industry Ontology Foundry is now under extensive discussion 12 (Morris KC and Kulvatunyou S. 2017) and is consistent with the understanding 13 developed through the the work reported in this paper. 14 Fo 15 16 3.2 Combining a bottom-up and a novel top down specialisation approach 17 18 From the evaluation of the bottom-up approaches it is clear that a more structured rP 19 approach to reference ontology development is needed if consistent and generally 20 applicable manufacturing reference ontology is to be developed. The idea proposed 21 follows from the recognition and supposition that all manufacturing activities can be 22 thought of from a systems perspective and that therefore a generic systems level ee 23 ontology should provide a consistent basis from which to build a manufacturing 24 reference ontology. The approach proposed by the authors and researched in the EU 25 FLEXINET project has the following main elements: 26 rR 27 1. Continue to build on a specialisation approach starting from a foundation 28 29 ontology and specialising down to domain levels; 30 2. Introduce a systems context level on which all subsequent ontology specialisation levels should be built; ev 31 32 3. Identify specialisation levels based on a combination of bottom-up and top down 33 approaches i.e. identify the common requirements that can be specialised down 34 to suit each set of domain requirements, but ensure that they remain consistent ie 35 with any higher level ontologies, but especially the systems level ontology; 36 4. Use the resulting manufacturing reference ontology, as illustrated in figure 3, as 37 w a means to readily construct mappings to (i) meet company specific ontological 38 39 requirements, (ii) build and maintain company knowledge bases and (ii) provide 40 a means of flexibly interfacing to a range of software services. On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 10 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 Figure 5: Specialisation from foundation to enterprise specific ontologies ee 23 24 The investigation of levels within the reference ontology was explained by Palmer et al 25 (2016) and is illustrated in figure 5, with the areas in white boxes being of direct 26 relevance to the investigations in this paper. The levels can be explained as follows: rR 27 Level 0: This is the foundation ontology level. In the research reported here Highfleet’s 28 (2014) Upper Level Ontology (ULO) has been used for this purpose; 29 Level 1: This provides the system’s context and potentially the most important ontology 30 level. This is explained in detail below; ev 31 32 Level2: While this research is focused on manufacturing systems it is clear that many 33 concepts could have wider applicability to any designed system. Also systems 34 that are not developed by people, called natural systems, could also employ the ie 35 level 1 concepts; 36 Level 3: This level contains no concepts of its own but purely provides the context for 37 the sort of designed system for which subsequent ontology levels are to be w 38 developed. In our case this is manufacturing business systems, but it could be 39 one of very many possibilities; 40 Level 4: This provides the detail level of specialisation for manufacturing business On 41 42 systems, with our focus being on the production aspect of the product lifecycle. 43 While this level is developed further in section 4, the focus of this paper is on 44 support for high-level strategic and tactical decisions. This level will need 45 further development and extension to support in factory operational decisions; ly 46 Level 5: This provides the ontology to support a specific enterprise and, as far as 47 possible, will be based on mappings from level 4, limiting new specialisations as 48 far as possible. 49 50 Note: 51 1. In our presentation of the ontology we use UML class diagrams to provide an 52 53 easier to comprehend visual illustration of the concepts and relations defined. In 54 the full ontology these are defined in Highfleet’s Common Logic based 55 Knowledge Frame Language that importantly includes the logic based axioms 56 that apply to the concepts and relations. 57 2. As we combine top-down and bottom-up approaches we explain in this section 58 the top-down definition of the level 1 ontology. Levels 2,3 and 4 have resulted 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 11 of 24 International Journal of Production Research 1 2 3 from eth bottom-up investigations in the FLEXINET project and are therefor 4 explained in section 4 below. 5 6 The level 1 ontology has developed over time with the current concepts and relations 7 illustrated in figure 6. One key aspect to the level 1 is the representation of Activity and 8 System Function through the use of formalised IDEF0 concepts (PUBs 1993). The 9 10 Activity concept being equivalent to an IDEF0 activity and the Roles of Input, Output, 11 Resource and Control playing the part of IDEF0 flows. An Input represents what is 12 brought into and is transformed or consumed by the activity to produce Outputs. A 13 Control is a specialised Input that provides a condition required to produce the correct 14 activity output (PUBs 1993). Resource represents the means by which an activity is Fo 15 performed. 16 17 Another key aspect is the definition of the two main parent classes at level 1; 18 those of Basic and Role. A Basic concept is independent of the system or context so its rP 19 definition does not depend on any other concept and an instance of a Basic always 20 21 retains its identity. Role defines how Basics are used in any particular Activity. Basics 22 occurring at level 1 are classified as Activity and Entity, with an Entity being anything of interest. Subtypes of Entity at level 1 are Information, Material and Energy. Note ee 23 24 that Material is used to denote any material thing and not raw materials. A Basic can be 25 comprised of Basics. This recursion can also be applied to Activity to represent sub- 26 activities or sub-system Functions. An Activity is also a subtype of Basic and provides a rR 27 context for the Roles that other Basics play in its performance. 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 Figure 6: level 1 generic systems concepts and relationships 53 54 Role depends on Activity for its context and an instance of a Role cannot exist 55 without such a context. Basics are used to play a Role within an Activity. Roles are also 56 57 represented as being played within a Scenario, so the playsRole relation is a quaternary 58 relation. A Basic playing a Role for certain period of time is represented using the 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 12 of 24 1 2 3 4 TimeSpan concept and modelled using the relation “playsRole”. The Scenario concept 5 has been provided in order to represent alternative ways of meeting the same overall 6 system requirements i.e. in production terms each scenario provides one way of 7 providing the system output. Alternative scenarios offer the potential for companies to 8 perform ‘what-if” analysis on each alternative. 9 Level 2 and Level 4 concepts cover an extensive range, which will continue to 10 expand as manufacturing reference ontologies develop. How these levels have been 11 12 developed within the FLEXINET project is described in the following section. 13 14 Of course, given the use of a logic based approach, there are many axioms (i.e. Fo 15 statements that are taken to be true) and rules that have been developed in order to 16 constrain the use of the concepts to their intended meaning. The level 1 axioms and 17 rules are limited in number. These are then added to as the levels of specialisation are 18 developed and in the FLEXINET reference ontology some 55 axioms and 26 rules were rP 19 developed. The level 1 axioms and rules are listed as: 20 21 22 1. Axiom – a role requires an activity to provide a context; ee 23 24 2. Axiom – an activity cannot contain a role and play the role; 25 3. Rule – role requires and activity 26 (if a role requires an activity as a context then the activity contains the role) rR 27 4. Rule – activity containing a role 28 (if an activity contains a role then the role requires the activity as a context) 29 30 As an illustration the coding of axiom 1 is: ev 31 notation for ‘a role requires an activity to provide a context’: 32 :Use 1SYSCtx 33 (=> (Role ?r) 34 (exists (?a) ie 35 (and (Activity ?a) 36 (requiresA ?r ?a)))) :IC hard "The Role ?r requiresA Activity to provide a context." 37 w 38 39 40 4 Applying the new approach in FLEXINET On 41 42 4.1 The FLEXINET Project 43 44 FLEXINET is focused on the early decision making processes that cut across the 45 ly combination of Product-Service engineering decisions and commercial business 46 decisions to support the dynamic evolution of global production networks. To be able 47 to create new networks and modify old ones, flexibility is required to introduce new 48 systems into existing networks and also to introduce new facilities or suppliers. This 49 requires information sharing between systems across different facilities within an 50 enterprise and between systems across multi-enterprise networks. To enhance the 51 viability of business decisions FLEXINET draws on prior knowledge of production 52 networks and their related global location knowledge. It provides an ideal set of 53 business scenarios from which to research the necessary manufacturing reference 54 ontology to support the complex knowledge sharing requirements of manufacturing 55 businesses. These include business model development, new product development, 56 production network configuration and risk analysis. This is illustrated in figure 7 where 57 the “Collaboration Environments” capability is used for new product development. 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 13 of 24 International Journal of Production Research 1 2 3 The objectives of FLEXINET from an ontological perspective were: 4 • To identify underpinning semantic concepts and their inter-relationships to 5 support the range of software tools and business information being exploited in the 6 project, taking into account the need for flexibility and interoperability. 7 8 • To create reference ontologies and compliance queries to support these tools 9 from initial new ideas through to the development of new business models and to the 10 development and evaluation of global production network design configurations. 11 12 13 The FLEXINET tools in these four areas provide decision makers with 14 information and knowledge critical to making effective decisions drawn from the KB Fo 15 and analysed or updated via the applications. FLEXINET as a whole is provided as a " 16 platform", i.e. a set of applications that can be used in combinations that can be 17 configured to suit the specific needs of particular end users. 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 28 29 30 ev 31 32 33 34 ie 35 Figure 7. The FLEXINET Global Business Support Environment 36 37 4. 2 The FLEXINET Reference Ontology w 38 This section illustrates the comprehensive range of concepts, relationships, and 39 constraints that have been developed and validated to demonstrate the capability of a 40 manufacturing reference ontology. The full ontology description containing some 600 On 41 42 concepts, 55 axioms and 26 rules can be found in the FLEXINET project public 43 deliverables (FLEXINET 2016). 44 45 4.2.1 An Overview of the ontologies and inter-relationships ly 46 47 The software selected to develop the reference ontology is the Highfleet Integrated 48 Ontology Development Environment (IODE) as it utilises an expressive common logic 49 based approach. The Common Logic based language used to implement the 50 FLEXINET reference ontology is the Knowledge Framework Language (KFL) 51 (Highfleet 2014). 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 14 of 24 1 2 3 4 The FLEXINET reference ontology contains a broad range of knowledge across a 5 number of interrelated domains as shown in figure 8. These nine domains have been 6 defined (Project, Network, Product, Business, Risk, Scenario, Indicators, Metrics and 7 Location) to provide the knowledge support for the applications involved the scope of 8 FLEXINET i.e. from a bottom-up perspective. 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 Figure 8. FLEXINET related knowledge domains rR 27 These domain ontologies use the top down levels described in section 3.2 and 28 29 exploit concepts defined as specialisations of the level 1 ontology concepts. Each of the 30 related knowledge domains are introduced in the following sub-sections, with the exception of metrics and indicators, as they do not add to the general argument of the ev 31 32 approach. 33 To support interoperability across the applications supported by the ontology 34 multiple relationships exist between the domains of the reference ontology. Three types ie 35 of relationship exist, as illustrated by the arrows shown in Figure 8: 36 37 w • Direct Relationships 38 39 • Containment Relationships 40 • Indirect Relationships On 41 42 Direct relationships are defined when a relation exists between properties which 43 are situated in two or more domains (a cross-domain relation). Direct relationships can 44 take the form of binary, ternary or quaternary relations. Examples of binary direct 45 ly relationships are: “locationHasExternalFactor” which links the Location and Indicator 46 domains; and “hasIndicatorValue” which specifies “Indicator” and “Metric” as its 47 48 arguments and forms a connection between the corresponding domains. 49 A containment relation occurs when a property in a domain is a container for a 50 property in a separate domain. It is specialism of a binary direct relationship. For 51 example, the “Project” property from the Project domain contains a “Business Model 52 Canvas” from the Business Model domain. The top level containment relation in 53 FLEXINET is “basicContainsBasic”. 54 Indirect relationships require two or more relations to link separate domain areas 55 within the reference ontology. They are formed from a chain of arguments which 56 connects the relations, i.e. Argument A which is a member of domain X is associated 57 58 with argument B through relation 1. Argument B is associated with argument C 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 15 of 24 International Journal of Production Research 1 2 3 (contained in domain Y) through relation 2. An example of an indirect relation is the 4 connection between the Location domain and the Business domain. The two connecting 5 relations which connect the domains are “facilityLocatedAt” and “playsRole”. An 6 instance of a Facility (a general Level 2 property) is associated with a Location through 7 the “facilityLocatedAt” relation. Within the Business Canvas model system the Facility 8 instance plays the role of a KeyPartner (relation “playsRole”, domain “Business”). 9 10 Examples of level 2 system concepts are illustrated in figure 9 and role concepts 11 that are played by Basics are illustrated in figure 10. Of particular importance for 12 FLEXINET high-level decisions are the system concepts of Organisation and Facility 13 and the Roles of Supplier, Producer and Customer. 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 Figure 9. Examples of level 2 System concepts 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 Figure 10. Examples of level 2 Role concepts 40 On 41 4.2.2 Organisation 42 The Organisation and Facility concepts are illustrated in figure 9 as specialisations of 43 the System concept. Organisation is synonymous with an enterprise which Facility 44 45 represents a system within an Organisation which has a specific location e.g. a factory ly 46 or a service centre. A Factory would then be further developed into its operational 47 capability, although this development has been outside the scope of FLEXINET. 48 49 4.2.3 Scenario 50 As mentioned in section 3.2 alternative scenarios offer the potential for companies to 51 perform ‘what-if” analysis on each alternative. So scenarios can be developed for 52 products or business models or risk or, of particular importance in manufacturing, of 53 production networks. The Network Level 2 ontology, illustrated in figure 11, can then 54 55 be specialised at level 4 into a production network or global production network. 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 16 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 Figure 11. The level 2 Network ontology 22 4.2.4 Project ee 23 Project is defined as a planned set of interrelated tasks to be executed within time and 24 cost limitations. The elements of a project of importance in FLEXINET were related to 25 Business models, represented as a Business Model Canvas, Decision Event, Scenario, 26 Concept (here with the meaning of a product concept) and Product. These are illustrated rR 27 28 in figure 12. 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 Figure 12. The level 2 Project properties 43 44 45 ly 4.2.5 Product 46 Product is clearly a domain that has a huge range of potential concepts within its scope. 47 In FLEXINET the requirement has been to distinguish between physical products and 48 service products. This has been achieved as illustrated in figure 13, where a Physical 49 50 Product” is ‘a material artefact’. A “Service Product” refers to ‘an offering’. A “Service 51 Using a Product” is ‘an offering that employs a physical product’ e.g. power by the 52 hour. A “Product Service” is ‘a physical product that also provides a service offering 53 that delivers value in use’ (Annamalai et al., 2011). A “Manufactured Product” is a 54 product that exploits/consumes raw materials. A Manufactured Product Service is a 55 specialisation of Manufactured Product that also provides a service offering. 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 17 of 24 International Journal of Production Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 28 Figure 13. The level 2 and level 4 product properties 29 30 4.2.6 Business ev 31 32 The Business Model properties linking business objectives to the elements of a Business 33 Model Canvas is represented in figure 14. In addition the Business model includes 34 support for the use Balanced Score Card techniques and as such has links defined from ie 35 balanced scorecard views to indicators as illustrated in figure 15. 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 Figure 14. The level 2 Business Model properties 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 18 of 24 1 2 3 class Lev el 2 - Business Balanced Score Card 4 Information System 5 Strategic Value 0..* hasStrategicValue 1 Level 2 - Designed Systems::Facility 6 + ID :char notes 7 Calculation Model Level 0 0..* 8 9 1..5 Information 10 BSC View 11 + + hasWeighting :Percent ID :char 12 + + lowerUncertainty :Percent upperUncertainty :Percent 13 notes Business Score Card - 14 Calculation Model Level1 Fo 15 16 1..* 17 Information hasValue KPI 18 rP + ID :char 19 + hasWeighting :Percent notes 20 Key Performance Indicator - Calculation Model Level 2 21 22 ee 23 1..* Quantitative Measure 24 PI Information hasFuzzyError Metrics::FuzzyError + hasWeighting :Percent Level 2 - Designed Systems:: 0..* 0..1 25 notes applies 0..* 1 Indicator 26 Company Performance Indicator - Calculation rR Model Level 3 0..* 0..* 27 hasFuzzyValue Quantitative Measure Metrics::FuzzyMeasure 28 0..1 1 29 30 Information ev 31 hasValue/ hasT argetValue 0..1 Metrics::Metric 32 33 34 Figure 15. Balanced Score Card linked to indicators and metrics ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 19 of 24 International Journal of Production Research 1 2 3 2.2.7 Risk 4 The risk applications in FLEXINET are used to evaluate risk scenarios in a Global 5 Production Network. Figure 16 illustrates the risk ontology developed to support these 6 applications. 7 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 Figure 16. Risk properties 28 29 30 4.2.8 Location Location is an important concept for globally based decision making. The appropriate ev 31 32 definitions for Location are dependent on use and can be difficult to define precisely as 33 location boundaries are often determined by human demarcation (Smith & Varzi, 2000). 34 In FLEXINET Location supports decisions related to global production networks, ie 35 business rules and risk assessment. The ontology for location used in FLEXINET is 36 shown in figure 17. 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 20 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 28 29 30 ev 31 32 Figure 17. Level 2 Location properties. 33 34 4.3 Exploiting the FLEXINET Reference Ontology ie 35 36 The FLEXINET project has constructed and interacted with four knowledge bases in 37 order to test the capability of the reference ontology. These KBs were one for each of w 38 our manufacturing company partners in pumps manufacture, white goods and in food & 39 drink plus a fictitious bike company that could be used for generic dissemination. 40 Figure 18 illustrates 6 of the main application software tools set in the context of the On 41 business, product and production development timeline. The applications being idea and 42 concept management, business model canvas development, business model evaluation, 43 product-service configuration, product network configuration and risk analysis. In 44 addition to these applications there are also some pre-configuration applications which 45 allow the company to pull in legacy data, existing supplier knowledge, business rules ly 46 and external global Social, Technological, Economic, Environmental and Political 47 (STEEP) information. 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 21 of 24 International Journal of Production Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 Figure 18. Example FLEXINET application tools in the context of a business, product ee 23 24 and production development timeline 25 26 These applications all interact via the knowledge base for information inputs and rR 27 outputs, all based on the underlying reference ontology. Examples of key interactions 28 are (i) between the generation of a new product concept in the Idea Manager application 29 and the use of this as a value proposition in the Business Model Canvas application and 30 (ii) in the Global Product Network application which checks the Location of any ev 31 potential Facility against the company’s business rules and the external STEEP data for 32 that Location. Figure 19 illustrates a user interface with a global production network 33 configuration application. 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 Figure 19. User’s view of global facility location along with defined flows in the 50 network. 51 52 The ontology supports reasoning in order to offer appropriate information to the 53 applications by means of applying queries to a populated knowledge base. Queries are 54 presented which guide user decisions. In the query shown in figure 20 an assessment is 55 being made as to whether a facility is suitable to be a Key Partner within a Business 56 Canvas Model. The end user “DrinksCo” has specified business policy requirements 57 within the knowledge base using the “specifiesBusinessPolicy” and 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 22 of 24 1 2 3 4 “hasThresholdValueForIndicator” relations. For example, it can be seen that 5 “DrinksCo” has stipulated that GDP must be greater than or equal to $2000. This query 6 is requesting the business policies specified by “DrinksCo” and comparing these values 7 with the external factors of the locations of facilities within the knowledge base. 8 9 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 28 29 30 Figure 20 Query to check for suitable Key Partners ev 31 32 33 The user is able to use the results returned by this query to select a suitable 34 facility to be a key partner within a Business Canvas Model which will create an ie 35 indirect relationship between the Location domain and the Business Canvas Model 36 domain. 37 w 38 5. Concluding discussion 39 40 Whilst there are many projects which have capture sets of manufacturing concepts what On 41 makes this work distinctive is that it captures the logical relationships between these 42 concepts as well as constraints on their use, so that the knowledge of their interactions 43 can be effectively captured and applied in complex business development, product 44 development and production network configuration activities. Not only does it define a 45 formal manufacturing reference ontology but it does this using a flexible and extensible ly 46 approach that is fundamental to future development as businesses evolve. 47 The FLEXINET reference ontology comprises approximately 450 properties and 650 48 relations. As FLEXINET is focused on the strategic and tactical levels of business 49 decision making most of the concepts and relations within the reference ontology occur 50 at a high generic level. The previous domain descriptions indicate Level 2 contains the 51 majority of constructs (330 properties and 435 relations). Level 4 contains 110 52 properties and 190 relations specialised to the manufacturing business systems domain. 53 The FLEXINET ontology is focused on strategic and tactical decision support 54 55 for new business opportunities. It has been used successfully by our industrial partners 56 and with our software partners as a basis for knowledge capture, use and sharing. 57 Importantly this has been successful across the three industrial sectors involved in the 58 project. 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 23 of 24 International Journal of Production Research 1 2 3 4 Perhaps the most important aspect of the FLEXINET ontology is the level 1 5 ontology as this provides an important bridge from foundation ontologies to clarify the 6 meaning of the concepts at level 2 and level 4. This helps to avoid the problems of 7 developing multiple incompatible reference ontologies. 8 9 Manufacturing reference ontologies contribute to supporting multi-faceted 10 decision-making by providing a knowledge framework, including key cross-domain 11 relationships. This means that the implications of decisions from one viewpoint can be 12 applied to other related viewpoints. Hence the user has the most appropriate information 13 to make intelligent multi-perspective decisions, reacting rapidly and reducing cost 14 through the use of high quality information. The inter-relationships within the Fo 15 reference ontology enable: 16 17 • Flexible approaches to information sharing across systems; 18 • Decisions taken in one domain to be understood within another domain; rP 19 20 • The ability to readily configure “common” business knowledge bases; 21 22 • Enable a migration path towards interoperable standards across a more broadly based manufacturing scope; ee 23 24 25 The Inter-relations between the different focused knowledge domain ontologies, as 26 demonstrated in FLEXINET and illustrated in figure 8, support interoperability across rR 27 28 multiple applications. However, this raises issues in the maintenance and development 29 of a reference ontology, as the interrelationships between these domain ontologies need 30 to be managed with care. Nonetheless it is clear that it would be of great benefit for future ontology development if a pre-defined set of relatively narrow ontologies could ev 31 32 be available for use in building the reference ontology. This would effectively provide a 33 library of ontologies or potential knowledge libraries that could be used and reused. 34 This would both minimise the cost and time in building the reference ontology and also ie 35 improve the ability of multiple reference ontologies to interoperate. 36 37 w 38 The FLEXINET ontology requires extension to broaden its scope of application 39 to in-factory decision support and to support operational level supply networks to be 40 more fully be aligned with the aims of Indusrie 4.0. The upper levels of the ontology On 41 could also be exploited for other domains outside of manufacturing. 42 43 44 6. Acknowledgements 45 ly 46 We wish to acknowledge the FLEXINET consortium and especially the financial support from 47 the European Union Seventh Framework Programme FP7-2013-NMP-ICT-FOF (RTD) under 48 grant agreement no 688627. 49 50 We also wish to acknowledge the support of the EPSRC in the early research understanding 51 presented in this paper who funded the Interoperable Manufacturing Knowledge Systems 52 (IMKS) under project 253 of the Loughborough University Innovative Manufacturing and 53 Construction Research Centre (IMCRC). 54 55 We also wish to acknowledge the Brazilian Science without Borders programme who supported 56 inputs through the “Intelligent Knowledge Libraries: exploiting new ICT technologies for improved Manufacturing Intelligence” project. 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

International Journal of Production Research Page 24 of 24 1 2 3 4 7. References 5 Annamalai, G., Hussain, R., Cakkol, M., Roy, R., Evans, S., & Tiwari, A., 2011. An ontology for Product-Service 6 Systems. Decision Engineering Report Series, R. Roy and Y Xu (Eds.). Cranfield University. 7 Arp R., Smith B., Spear A. (2015) “Building ontologies with basic formal ontology” The MIT press. ISBN 8 9780262527811. 9 Blomqvist, E. and A. Öhgren (2008). "Constructing an enterprise ontology for an automotive supplier." Engineering 10 Applications of Artificial Intelligence 21(3): 386-397. Borgo, S., and Leitão, P. 2007, Foundations for a core ontology of manufacturing. Integrated Series in Information 11 Systems, Volume 14, pp 751-775. 12 Borsato, M., 2014. Bridging the gap between product lifecycle management and sustainability in manufacturing 13 through ontology building. Computers in Industry, 65, pp. 258-269 14 Bruno G., Korf R., Lentes J., Zimmerman N., (2016). “Efficient management of product lifecycle information Fo 15 through a semantic platform”. Int. J. Product Lifecycle Management, Vol. 9, No. 1. Chen, D., Knothe, T., & Doumeing, G., 2009, POP* Meta-Model For Enterprise Model Interoperability. Proceedings 16 of the 13th IFAC Symposium on Information Control Problems in Manufacturing. Moscow, Russia, June 3-5, 17 pp. 176-180. 18 Chungoora, N., & Young, R.I.M. (2011). The configuration of design and manufacture knowledge models from a rP 19 heavyweight ontological foundation. International Journal of Production Research, 49(15), pp. 4701-4725. 20 Chungoora, N., Gunendran, G.A., Young, R.I.M., Usman, Z., Anjum, N.A., Palmer, C., Harding, J.A., Case, K., and Cutting-Decelle, A.F., 2012. Extending product lifecycle management for manufacturing knowledge sharing. 21 Proceedings of the Institution of Mechanical Engineers Part B - Journal of Engineering Manufacture, 226 22 (A12), pp. 2047-2063. ee 23 Chungoora, N., Cutting-Decelle, A-F., Young, R.I.M., Gunendran, G., Usman, Z., Harding, J.A., Case, K., 2013. 24 Towards the ontology-based consolidation of production-centric standards, International Journal of Production 25 Research, vol. 51 no. 2, pp. 327-345. FLEXINET (2016) “product service ontology report”. http://www.flexinet- 26 fof.eu/Documents/FLEXINET%20D3.5%20Product-Service-Production%20Ontologies.pdf rR 27 Gruninger, M. and J. Lee (2002). "Introduction: Ontology Applications and Design." Communications of the ACM 28 45(2). 29 Highfleet (2014) Highfleet Ontology Library Reference, Baltimore, MA: HIGHFLEET Inc 30 Hastilow N., Young R. (2012). “Understanding ‘Manufacturing Intelligence’ – a precursor to interoperable manufacturing systems” in Enterprise Interoperability VII: Proceedings of the I-ESA Conferences 8, pub Wiley ev 31 2012. 32 Hastilow N 2013 Manufacturing systems interoperability in dynamic change environments PhD thesis, 33 Loughborough Univeristy, Loughborough, UK. https://dspace.lboro.ac.uk/dspace-jspui/handle/2134/13174 34 Huber A. (2014) presentation by Siemens CEO, World Manufacturing Forum, Milan, July 2014. ie 35 http://www.ims.org/2014/07/world-manufacturing-forum-2014/ [Accessed April 2016]. Imran M and Young R (2013) “The acclication of common logic based formal ontologies to assembly knowledge 36 sharing” Journal of Intelligent Manufacturing. Vol. 26(1), pp 139-158. Jan 2013 37 w Imran M 2015. “Towards an Assembly Reference Ontology for Assembly Knowledge Sharing”. PhD thesis, 38 Loughborough Univeristy, Loughborough, UK. https://dspace.lboro.ac.uk/dspace-jspui/handle/2134/13995 39 ISO/IEC-24707. (2007). Information technology — Common Logic (CL): a framework for a family of logic based. 40 Retrieved 08 14, 2011, from Common Logic Standard: http://standards.iso.org/ittf/licence.html Kagermann H., Wahlster W., Helbig J. (2013) “Recommendations for implementing the strategic initiative On 41 INDUSTRIE 4.0” Final report of the Industrie 4.0 working group. Pub April 2013. Office of the Industry- 42 Science Research Alliance, beim Stifterverband für die Deutsche Wissenschaft, Ulrike Findeklee, M.A. 43

or contact

44 D Leal, A Feeney. An Architecture for Industrial Data Standards, Report to ISO TC184 SC4 15 June 2010 45 Mascardi, V., Locoro, A., Rosso, P. (2010). Automatic Ontology Matching via Upper Ontologies: A Systematic ly Evaluation. IEEE Transactions on Knowledge and Data Engineering, May 2010, Vol.22(5), pp.609-623. 46 Morris KC., Kulvatunyou S. (2017) http://blog.mesa.org/2017/03/working-towards-industrial-ontology.html 47 Noy, N. F. and D. L. McGuinness (2000). Ontology Development 101: A Guide to Creating Your First Ontology. D. 48 L. McGuinness. 49 Oh, S., Yeom, H.Y. and Ahn, J. (2011). Cohesion and coupling metrics for ontology modules. Information 50 Technology and Management, 12(2), pp.81-96. Palmer, C., Urwin, E.N., Pinazo-Sánchez, J-M, Cid, F.S., Rodríguez, E.P., Pajkovska-Goceva, S. and Young, R.I.M. 51 (2016), ‘Reference Ontologies to Support the Development of Global Production Network Systems’, 52 Computers In Industry, Vol. 77, pp. 48-60. 53 Palmer, C., Urwin, E.N, Marilungo E. and Young, R.I.M., 2017, “Reference ontology approach to support global 54 product-service production”, International Journal of Product Lifecycle Management.?? 55 Pattenden, S., Young, R., and Zelm, M. Standardisation Task Force Report to Future Internet Enterprise Systems (2012) 56 Ray, S. R., & Jones, A. T., 2006, Manufacturing interoperability. Journal of Intelligent Manufacturing, 17(6), 681– 57 688. 58 59 60 http://mc.manuscriptcentral.com/tprs Email:

Page 25 of 24 International Journal of Production Research 1 2 3 Shvaiko P. & Euzenat J., 2013. Ontology Matching: Stae of the art and future challenges. IEEE Transactions on knowledge and data engineering. Vol 25. Issue 1. pp 158-176 4 Sureephong, P., N. Chakpitak, Y. Ouzrout and A. Bouras. 2008. "An Ontology-based Knowledge Management 5 System for Industry Clusters." In International Conference on Advanced Design and Manufacture (ICADAM 6 2008). Sanya: China. 7 Szejka A L (2016) “Contribution to interoperable product design and manufacturing information: application to 8 plastic injection products manufacturing” Doctoral thesis of the Pontifical Catholoic University of Parana, Brazil and the University of Lorraine, France. 9 Uschold, M. King, M., Moralee, S., & Zorgios, Y., 1998). The enterprise ontology, The Knowledge Engineering 10 Review, 13 (1), pp. 31-89. 11 Usman, Z., Young, R.I.M., Chungoora, N., Case, K., Palmer, C. and Harding J.A.,(2013), ‘Towards a formal 12 manufacturing reference ontology’, International Journal of Production Research, Vol. 51, Iss. 22, pp. 6553- 13 6572. Usman Z. 2012. “A Manufacturing Core Concepts Ontology to Support Knowledge Sharing” PhD thesis, 14 Loughborough Univeristy, Loughborough, UK. https://dspace.lboro.ac.uk/dspace-jspui/handle/2134/9857 Fo 15 Weichhart, G., Molina, A., Chen, D., Whitman, L.E. and Vernadat, F., 2016. ‘Challenges and current developments 16 for sensing, smart and sustainable enterprise systems’, Computers in Industry, vol 79, pp.34-46. 17 West M., Fowler J. (2001) “The IIDEAS architecture and integration methodology for integrating enterprises” PDT 18 Days 2001 rP Ye, Y., Yang, D., Jiang, Z., & Tong, L. ,2008, Ontology-based semantic models for supply chain management. The 19 International Journal of Advanced Manufacturing Technology, 37(11), 1250-1260. 20 Young, R. I. M., Gunendran, A. G., Cutting-Decelle, A. F., & Gruninger, M. , 2007. Manufacturing knowledge 21 sharing in PLM: a progression towards the use of heavy weight ontologies. International Journal of Production 22 Research, 45(7), 1505-1519. Young, R., Pooplewell K., Jaekel F-W., Otto B., Ghullar G., (2014) “Intelligent Systems Configuration Services for ee 23 Flexible Dynamic Global Production Networks” Workshop on Interoperability for Enterprise Systems and 24 Applications. I-ESA 2014. 25 26 rR 27 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tprs Email: