Discuss the relationship between data, information, and knowledge. Support your discussion with at least 3 academically reviewed articles.

Knowledge integration and sharing for complex product development

Z.Y. Wu*, X.G. Ming, L.N. He, M. Li and X.Z. Li

Shanghai Key Lab of Advanced Manufacturing Environment, School of Mechanical Engineering, Shanghai Research Center for Industrial Informatics, Institute of Computer Integrated Manufacturing, Shanghai Jiao Tong University, Shanghai, P.R. China

(Received 12 May 2013; accepted 30 April 2014)

Product development is a highly creative and knowledge-intensive process that involves extensive information and knowledge exchange and sharing among geographically distributed teams and developers. How to best integrate such heterogeneous product knowledge has become an extremely important knowledge management (KM) subject associated with product development. Product development knowledge integration and sharing is becoming a key issue in the enter- prise KM. This paper addresses the challenges of product development knowledge integration and sharing during product development. The aim of this research work effort is to develop a method to enhance the integration and sharing of product knowledge during the development phase. This study presents a systematic approach to developing knowledge integration and sharing for product development. The proposed approach includes the steps for designing a framework of product development knowledge sharing, developing representation model for product development knowledge, designing product development knowledge sharing process, designing product development knowledge integration ontol- ogy, developing knowledge integration and sharing method for product development based on ontology, and implementa- tion of product development knowledge integration and sharing method. The proposed method could bring out an efficient and proactive way for knowledge integration and sharing among product developers in the product development process.

Keywords: knowledge management; knowledge representation; knowledge integration; knowledge sharing; product development

1. Introduction

Information technology has promoted our society gradually into knowledge economy from product economy. Organisa- tion for Economic Co-operation and Development (OECD) defines that knowledge economy is the economy which is built on knowledge and information production, distribution and application. In knowledge economy era, value creation requires more and more knowledge and information to support enterprise activities. Knowledge management (KM) as an activity has already existed, however the conceptual use of KM to describe the activities of enterprise knowledge process began in Drucker. KM will make the organisation and individual have stronger competition strength, and make better decisions.

Product development knowledge in the manufacturing industry is a very special type of knowledge that supports how products can be best produced. These development knowledge and other intangible properties have become the most important and valuable assets for product development. Product development knowledge, including standards, development data, development tools, product development documentation and case as well as many trivial aspects, still rests in the information system or in the minds of experienced designers (Denkena et al. 2007). They need to work in this job or field for several years before becoming experts of product development. And without these experts, tacit knowledge is not available to be shared or be transferred as needed (Liao 2005). For most manufacturing companies in China, the shortage of knowledge workers or experts is one of the biggest gaps standing in the path of their growth. This situation has become a difficulty for most China manufacturing companies. New methods need to be developed and proposed to tackle such a problem. With the development of product knowledge modelling and information soft- ware technologies, classification and representation are now possible. The classification and representation of knowledge is still difficult but serves as the foundation in managing knowledge.

The motivation of this study mainly lies in: the knowledge recommendation based on the personal profile that could act as a supplement to the traditional knowledge query. In the product development process, it spends a lot of time for engineers to search knowledge from a repository with large amounts of data and information. Product development is

*Corresponding author. Email: [email protected]

© 2014 Taylor & Francis

International Journal of Production Research, 2014 Vol. 52, No. 21, 6296–6313, http://dx.doi.org/10.1080/00207543.2014.923121

becoming increasingly knowledge intensive in manufacturing enterprises. Product development engineers sometimes are not sure on what knowledge they really need and sometimes do not clearly describe which keywords should be used to search knowledge form the knowledge repository. Due to the above disadvantages in the traditional knowledge query mode, the knowledge recommendation based on the personal profile could recommend some potentially useful knowl- edge to those product development engineers who may really need them.

To overcome these challenges, this paper proposes a framework for product development, knowledge representation and sharing. Key techniques for knowledge representation model and knowledge sharing are elaborated. A case study is also illustrated to show the implementation and potentials of the framework.

The remainder of this paper is structured as follows. Some state-of-art reviews are briefly introduced in Section 2. A framework for product development knowledge sharing is depicted in Section 3. For the key techniques and solutions, i.e. knowledge representation model and sharing method, a detail of the construction is given in Section 4. A case study is pro- vided to demonstrate the implementation and potentials of the framework in Section 5. In Section 6, an experiment is per- formed for the method proposed in this paper. And finally, conclusions and future research directions are made in Section 7.

2. Current research status

In recent years, there have been significant and considerable developments in product KM, especially in product devel- opment. Therefore the relationship between KM and product development is becoming an important issue in academic and industry areas (Liao and Wu 2010). Bernard and Xu analyse the knowledge evolution process in product develop- ment activities and propose an integrated knowledge reference system (Bernard and Xu 2009). Kim and Kim propose causal knowledge, present a mathematical comparison of procedural knowledge and causal knowledge, and discuss the potential roles and feasibility of causal knowledge across product development KM (Kim and Kim 2011). To speed up the product design efficiency, product designers would like to utilise the past experience and know-how in assisting the design of new products or in the enhancement of existing ones. Design structure matrix is considered to be a suitable means to capture and manage the system-level design knowledge (Tang et al. 2010). Bradfield and Gao investigated KM problems in the new product development process of a multi-national manufacturing company, and proposed an ontology-based methodology to disseminate knowledge to users (Bradfield and Gao 2007). Zhang et al. develop an inte- grated KM and reuse framework for Product-Service Systems business in construction machinery industry (Zhang et al. 2012). Dekkers, Chang and Kreutzfeldt present a systematic review of the literature focused on the interface between these two areas and identify six core themes based on 49 retrieved papers (Dekkers, Chang, and Kreutzfeldt 2013). Chen presents a systematic approach to developing a knowledge integration and sharing mechanism for collaborative moulding product design and process development in his study (Chen 2010).

Mountney, Gao, and Wiseall consider the use of manufacturing knowledge in the earlier, preliminary stage of design where the geometry is not so developed and where development of manufacturing technology may be required, and a pro- totype KM tool was developed by them to meet those requirements (Mountney, Gao, and Wiseall 2007). Cai et al. attempt to shed some light on the mechanisms underpinning knowledge sharing in supply chains in their paper. In particular, they focus on knowledge sharing in a dyadic buyer–supplier relationship (Cai et al. 2012). Zhen and Jiang design a knowledge grid-based knowledge query platform to support innovative product design activities (Zhen and Jiang 2008). Martín et al. propose a new paradigm where intelligent KM is integrated into the conceptual repository of management information, they present a technique for the design and implementation of a distributed intelligent system that is designed through the normalisation of KM (Martín et al. 2012). Dantas and Farias present conceptual navigation and NavCon, made use of ontology as metadata to contextualise user search for knowledge (Villela Dantas and Muniz Farias 2010). Huang, Lin and Chan develop a system that comprises a semantic tagging mechanism and triple-pattern and visual searching mechanisms by taking advantage of Semantic Web technology and the topological characteristics of knowledge maps (Huang, Lin, and Chan 2012). Jayaram and Pathak propose a finer grained view of knowledge integration mechanisms in collaborative sup- ply chains. Specifically, they propose two different types of mechanisms that are in play within a collaborative supply chain; short-term knowledge sharing and iterative knowledge enrichment (Jayaram and Pathak 2012). Hou and Pai develop a knowledge visualisation methodology that computer systems can automatically convert the textual knowledge into visual- ised display (Hou and Pai 2009). Zhen, Song and He make an explorative study on the personal KM, and analyse various forms of personal knowledge resources in the product development process and propose a model of recommender systems for personal KM for knowledge sharing among members in the collaborative environment (Zhen, Song, and He 2012). Lin and Hsueh propose a knowledge map management system to facilitate KM in virtual communities of practice and develop knowledge map creation and maintenance functions by utilising information retrieval and data mining techniques (Lin and Hsueh 2006). Chen, Chen and Wu present a systematic approach to develop a framework for managing empirical knowledge to support a professional virtual community in knowledge-intensive service industries (Chen, Chen, and Wu

International Journal of Production Research 6297

2012). Chau, Moghimi and Popovic propose a knowledge ecosystem to frame the rehabilitation engineering KT process from need to product (Chau, Moghimi, and Popovic 2013).

Besides the above KM or knowledge sharing techniques, Wang and his team propose an agent-based KM system which builds on organisational knowledge resources and works with the advanced product quality planning concept as well as KM agents, the proposed framework reflects how the tool with KM agents seeks significant inroads into process knowledge repository to recommend a solution for decision-making in semiconductor manufacturing enterprises (Wang et al. 2010). Zhen, Huang and Jiang propose a workflow-based recommender system model on supplying proper knowl- edge to proper members in collaborative team contexts rather than daily life scenarios (Zhen, Huang, and Jiang 2009). Young et al. presents a view of the current status of manufacturing information sharing and discuss the potential for heavy weight ontological engineering approaches so as to facilitate sharing knowledge in cross-disciplinary product development teams (Young et al. 2007). Shaw and Edwards study key components of a knowledge sharing strategy, and the relationships between it and manufacturing strategy. They also make comments on the development of action plans for better KM (Shaw and Edwards 2006). In recent years, knowledge recommendation is becoming a new hotspot in both academics and industries, and shown its efficiency in many areas. Li et al. propose a social recommender system which generates discussion thread and expert recommendations based on semantic similarity, profession and reliability, social intimacy and popularity, and social network-based Markov Chain models for knowledge sharing in online forum communities (Li, Liao, and Lai 2012). Zhen et al. propose a model of distributed knowledge recommender system to facilitate knowledge sharing among collaborative team members. Their study also provides a new framework for simu- lating and studying individual or organisational behaviours of knowledge sharing in a collaborative team (Zhen, Jiang, and Song 2010). Urwin and Young focus upon a rapid development and deployment method in their work that enables the capture and representation of machining knowledge so that it may be shared and reused by design engineers to accelerate the design-make process (Urwin and Young 2013).

From the above literature review, we can see that they are mainly based on a centralised knowledge repository. By these KM methods, product development engineers usually need input keywords to query their knowledge from knowl- edge repository. It is cumbersome for engineers to get the knowledge they may need after browsing all the query result. Product development is an intensive knowledge involved, often complex, fuzzy and iterative process, and the needs and specifications of the knowledge is more refined as product development process develop towards its goal. Product devel- opment engineers sometimes may not clearly describe what kind of knowledge keywords should be keyed in to query knowledge from the knowledge repository. To make up for the gap, this study proposes a framework that contains knowledge representation model and active knowledge sharing mode, which can efficiently supply useful knowledge for engineers to support their development tasks in the process of product development.

3. A framework of product development knowledge sharing

The framework proposed in this paper has two key techniques: knowledge representation model and knowledge sharing mode. In product development process, there are various types of knowledge resources that are stored in the form of document, software or tacit knowledge from knowledge experts. A proper knowledge representation model can support effectively to knowledge sharing in product development. The framework of product development knowledge sharing is illustrated in Figure 1.

It mainly has three parts: knowledge ontology, product development knowledge sharing process and product devel- opment process. Firstly, knowledge ontology is divided into four types: know-what, know-how, know-why and know- who. The details are elaborated in Section 4.1. The second part in this framework is the knowledge process that contains knowledge demands, problem description, knowledge query, knowledge matching, knowledge recommendation and knowledge application. The details are elaborated in Section 4.2. The development process is the third part in this framework that composes of idea generation, concept design, preliminary design, detail design, technical implementation et al. Each stage of product development knowledge characteristic can map to a knowledge process. And engineers may involve in various kinds of products.

4. Technology solutions for product development knowledge sharing

4.1 Product development knowledge representation model

In order to propose an appropriate model, the product development knowledge is divided into four categories defined by OECD in product development stage (Foray and Lundvall 1997): know-what knowledge, know-why knowledge, know- how knowledge and know-who knowledge.

6298 Z.Y. Wu et al.

� Know-what: the definitions and contents from activities of product development process. � Know-why: the executive motivation or intention from developmental activities in each phase of product devel-

opment process, such as product functional requirements, customer requirements, design specification, experi- ence knowledge, and mouldability assessment data.

� Know-how: the executive steps for activity from the process of product development as well as the required product operation knowledge such as CAD model views, FMEA diagram, virtual prototypes, computer algo- rithms and design procedures.

� Know-who: the domain expert and resource application in executing every activity from the process of product development such as product design engineer, expert, CAD\CAE and material database.

The proposed model for knowledge representation in product development is shown in Figure 2. There are three lay- ers in this representation model, namely resource layer, XML topic map (XTM) layer and ontology layer. The resource layer provides information and knowledge resource contain various types of information and knowledge source to meta- data. The XTM layer establishes links between the knowledge resource and the knowledge application. The ontology layer provides a knowledge ontology which consists of know-what ontology, know-why ontology, know-how ontology and know-who ontology.

The resource layer provides product development data, information, tangible and intangible knowledge. There are various types of information and knowledge through product development process, and they need to be classified and organised. These information or knowledge contain customer requirements, decision tables, product development cases, design rules, detailed drawings, CAD model views, FMEA diagram, computer algorithms multimedia, flow charts and photographs. And these information and knowledge resource is the foundation of knowledge representation.

The XTM layer provides a mechanism for representing information about the structure of information and organising it into topics. The topic has associations and occurrences that it can represent and define relationships between these topics. Information and knowledge about the topics can be connected by traversing the associations and occurrences linked to the topic. The hyper-graph method ensures the knowledge representation model is unified in the same

Know-what ontology Know-how ontology Know-who ontologyKnow-why ontology

knowledge resources

knowledge ontology

Knowledge demands

Problem description

Knowledge query Knowledge matching

Knowledge recommendation

Knowledge application

Service Process

Development Process

Idea generation

Concept design Preliminary design

Detail design

Technical Implementation

Product 1 Product 2 Product 3 Product n··· Product category

Engineer 1 Engineer 2 Engineer 3 Engineer n

Users

Figure 1. The framework of product development knowledge sharing.

International Journal of Production Research 6299

resolution of knowledge representation, which guarantees the model is reusable and robust. It is important to facilitate applications to check effectively and consistency of this model. The XTM layer establishes links between the informa- tion/knowledge and knowledge ontology.

The ontology layer provides definitions for product development knowledge, attributes and relationships between each other to achieve knowledge sharing. The ontology is defined to four types: know-what ontology, know-why ontol- ogy, know-how ontology and know-who ontology. This ontology can help to associate design processes, design objects and knowledge objects together, realise the representation of four kinds of knowledge and lay the foundation for the fol- lowing knowledge sharing and reuse.

4.2 Product development knowledge sharing

In this section, the process of knowledge development knowledge for knowledge integration and sharing is first devel- oped. Based on this sharing process, the ontology-based framework for product knowledge integration and sharing are then designed. The integration and sharing process for product development knowledge is designed based on the ontol- ogy-based framework for product development knowledge integration.

4.2.1 Product development knowledge sharing process design

Based on the knowledge integration (Chen, Chen, and Chu 2009; Chen 2010) and knowledge sharing (Wang and Noe 2010; Zhen, Jiang, and Song 2011) techniques, the process of knowledge development knowledge for knowledge inte- gration and sharing is developed, as shown in Figure 3. The process model of product development knowledge for knowledge integration and sharing mainly are combined with four layers: the product development process, the knowl- edge sharing process, the knowledge operation process and operation supporting process. The product development pro- cess layer has five main phases: idea generation, concept design, preliminary design, detail design and technical implementation. The knowledge sharing process layer has three main phases: knowledge requirement establishment, knowledge matching and sharing and knowledge recommendation. There are six phases in the knowledge operation pro- cess: need definition, need dissolution, knowledge resource, knowledge integration, and knowledge sharing and product development knowledge recommendation. The operation supporting process layer comprises of the following phases: knowledge definition, knowledge classification, knowledge representation, integration ontology definition, ontology schema definition, ontology mapping and integration, integration ontology establishment, problem description, knowl- edge query, knowledge learning, decision-making, problem-solving and knowledge matching.

Know-what ontology

Know-how ontology

Know-who ontology

Know-why ontology

XML Topic Maps XML Topic Maps XML Topic Maps XML Topic Maps

Ontology layer

XTM layer

Resource layer

Knowledge object

Figure 2. The proposed model for product development knowledge representation.

6300 Z.Y. Wu et al.

4.2.2 Ontology-based framework for product development knowledge integration and sharing

Based on the knowledge sharing process of product development, the ontology-based framework for product develop- ment knowledge integration and sharing is developed to support knowledge integration and sharing in each phase and activity of product development. This designed framework comprises knowledge representation layer, knowledge classi- fication layer, knowledge integration layer and knowledge sharing layer. Each layer is defined as follows:

� Knowledge representation layer An effective product development knowledge integration and sharing framework establishes a sharable knowl- edge representation method and model, records the linking address physical knowledge. The physical knowl- edge provides product development data, information, tangible and intangible knowledge.

� Knowledge classification layer According to the knowledge classification defined in Section 4.1, the product development knowledge ontology is classified into four types: know-what ontology, know-why ontology, know-how ontology and know-who ontology. This represented knowledge ontology is mapped to the know-X ontology for knowledge integration.

� Knowledge integration layer Using the knowledge integration method and mechanism, the represented knowledge ontologies are mapped and integrated into the knowledge integration ontology ‘Know-X ontology’ through the adaptors. Using the know-X ontology, this knowledge integration ontology can connect physical knowledge address. Thus, product developers can share their own product development knowledge with other knowledge owners in product devel- opment process to increase product value.

� Knowledge sharing layer Product developers can retrieve knowledge from integrated product development knowledge according to their knowledge demands. The primary process in knowledge sharing is as follows: problem description, knowledge query, knowledge learning, decision-making and problem-solving. Knowledge searching identifies the most suit- able product development knowledge via knowledge similarity computations for all product development knowledge (Figure 4).

Knowledge Requirement Establishment

Knowledge Mmatching &

Sharing

Knowledge definition

Knowledge classification

Knowledge representation

Knowledge integration

Knowledge sharing

Knowledge resource

Need definition

Need dissolution

Integration ontology definition

Ontology Schema

definition

Ontology mapping & integration

Problem description

Knowledge query

Knowledge learning

Decision making

Problem solving

Integration ontology

establishment

Knowledge recommendation

Product development knowledge

recommendation

Knowledge matching

Legend

Developmental process

Mapping process

S e rvice sMa rk e t in gPro du ct io nPro du ct D e v e lo pm e n tMark e t S u rve yProduct Development

Process

Idea generation Concept design Preliminary design

Detail design Technical

Implementation

Knowledge Sharing Process

Knowledge Operation

Process

Operation Supporting

Process

Figure 3. Knowledge sharing process of product development.

International Journal of Production Research 6301

4.3 Product development knowledge integration and sharing method development based on ontology

To realise the knowledge integration and sharing for ontology-based product development knowledge, the integration and sharing method for product development knowledge is developed in this section. This developmental process involves the establishment of knowledge integration ontology, development of selecting the target knowledge for knowl- edge query and design of a filtering method for supplied knowledge for developers.

4.3.1 Establishment of a knowledge integration ontology

The product development knowledge integration ontology is present in Figure 5, which encompasses development pro- cess, knowledge object and development object. The sub-concepts of these ontologies are defined as follows:

� Sub-concepts from the development process are combined with some important elements, which are process element, design activity, design task, participant, design guide and guide model.

� Sub-concepts from the development object encompass structure, material, person, product, component, home part, out part, standard part, external cooperation part, function and bill of the material.

� Sub-concepts from knowledge object consist of several ontology elements, which are mainly project, functions, structure, department and persons involved in the project. The relationship between properties is defined as shown in figure.

4.3.2 Development of selecting the target knowledge for knowledge query

After establishment of a knowledge integration ontology described in Section 4.3.1, an ontology-based search method for product development knowledge is designed that allows product developers to search for product development knowledge.

Know-what ontologyKnow-what ontologyKnow-what ontology

Knowledge integration

Request Response

Knowledge representation

Know-what ontology

Know-X ontology

Adaptor

Ontology mapping

Problem description

Knowledge query

Knowledge learning

Decision making

Problem solving

Administrators

Manage

Manage

Knowledge Sharing

Knowledge classification

Adaptor Adaptor Adaptor

Knowledge searching

Figure 4. Ontology-based framework for product development, knowledge integration and sharing.

6302 Z.Y. Wu et al.

The calculation of the similarity between knowledge users and knowledge resources is based on the knowledge ontology. The keywords of knowledge requirement match with knowledge ontology’s attribute value. Traversing all knowledge resources, the similarities between knowledge requirement keywords and knowledge ontologies are calcu- lated one by one. Then the average value of the above results is derived to measure the similarity between them:

SimðR; SÞ ¼ a � TRS TR

� � þ b � TRS

TS

� �

where R and S denote the knowledge user and knowledge resource; TR denotes knowledge characteristics of knowledge requirement set; TS denotes knowledge resource set; TXS denotes the similar characteristics set between knowledge requirement and knowledge resource; a; bð0\a; b\1Þ denote the weight coefficient, set by users according to actual condition.

Consequently, similarity matching for knowledge query is the sum of key attributes similarity as shown:

KSim ¼ Xn i¼1

Simðr; sÞ

where i denotes the attribute set of knowledge ontology.

4.3.3 Design of a filtering method for supplied knowledge for developers

For a product developer, this section addresses how to filter the supplied knowledge queried from knowledge resource. Figure 6 shows the query and filtering process based on ontology for product development, which includes the follow- ing steps: input knowledge requirement, filter the keywords, get all knowledge ontology attributes, calculate the key- words and attributes similarity, sum keywords and ontology attributes similarity, sort results based on their similarity and output knowledge.

When a product developer connects into the product development knowledge sharing network, some potentially use- ful knowledge resources are supplied to him. Then a filtering process will select some correlative knowledge resource, and the knowledge resource will be recommended to the product developer.

The criterion of some filtering is a function defined as FCðr; qÞ, which denotes the score of the knowledge resource q as to the product developer r:

FCðr; qÞ ¼ Pn

i¼1wi � Simðri; qiÞPn i¼1Simðri; qiÞ

Knowledge_ objects

Persons Departments Structures Functions Projects

Know-X Development_

processes Development_

obiects

is_owned_by

belong_to belong_to is_related_with belong_to

belong_to

has_structure

has_function

has_participate_in

Concept classification Object attribute Concept entity

Figure 5. Knowledge integration ontology.

International Journal of Production Research 6303

where wi denotes the score given by one as the important degree of knowledge ontology attribute; we could simplify that process by assigning from 0 to 1 to that wi by knowledge user or supplier; Simðri; qiÞ denotes the similarity between knowledge resource q and product developer r.

5. An example of application in product development

Based on the proposed methods and techniques for knowledge integration and sharing for product development, a proto- type of knowledge integration and sharing mechanism is developed to support auto headlamp product development.

The headlamp is also a very personalised product. The headlamp is a complicated auto unit, and it is often difficult for product developers to understand the problems and issues incurred during the usage of car headlamp. This case study will demonstrate how product developed knowledge integration and sharing method can help developers to obtain knowledge they required during product development stage. The structure of auto headlamp is shown in Figure 7 as follows:

5.1 Establishment of knowledge representation model

The headlamp product development knowledge object ontology model is shown in Figure 8, which consists of four kinds of knowledge ontology such as know-what knowledge, know-why knowledge, know-how knowledge and know- who knowledge. The sub-concepts of these knowledge ontologies are identified as follows:

� Sub-concepts from know-what consist of several ontology elements, which are mainly data properties, func- tions, features, plans and markets.

� Sub-concepts from know-why ontology consist of several ontology elements, which are mainly persons, resources, rules and projects.

� Sub-concepts from know-how ontology consist of several ontology elements, which are mainly data properties, functions, features, plans and markets.

� Sub-concepts from know-who consist of several ontology elements, which are mainly persons, department, workgroups, roles and projects.

According to the ontology developed above, the XTM layer establishes links between the knowledge resource and the knowledge application. A part of prototype model can be seen in Figure 9. The development object is a panel of purfling assembly. According to the definition above, there are three layers about headlamp development knowledge rep- resentation model.

The resource layer is provided with a knowledge resource that contains a database of information system, Internet/ Intranet web information and the participant’s form for completing the project. In the XTM layer, there are XTM topics

START

Input knowledge requirement

Filter the keywords

Get all knowledge ontology attributes

Calculate keywords and attributes similarity

Sum keywords and attributes similarity

Sort results based on their similarty

Output knowledge

END

Figure 6. Query and filtering process based on ontology for product development.

6304 Z.Y. Wu et al.

and the relation between them. According to their attributes, there are different types of relationships between the XTM topics. The XTM topic and XTM relation connect to knowledge resource through XTM link as shown in figure. The ontology layer provides a knowledge ontology which consists of design process ontology, design object ontology and knowledge object ontology. In Figure 9, it shows the know-how ontology, and also expresses the property of know-how ontology. The ontology and ontology relation connect to XTM topics through ontology link as shown in figure.

5.2 Design of integration ontology for product development

The panel colour development knowledge integration ontology model is shown in Figure 10, which encompasses devel- opment process ‘preliminary design’, knowledge object ‘colour’ and development object ‘panel’. The sub-concept from knowledge object is knowledge contents and resources for know-what, know-why, know-how and know-who. The sub- concept from development process is activities, participants, task, et al. The sub-concept from development object is the product or part name which is developed.

The knowledge integration ontology schema is used to build the product development knowledge maps to integrate and share knowledge. The schema is described as follows:

headlamp

lamp assembly

dimmer motor assembly

PES assembly

matched mirror assembly

reflector assembly

standard parts

harness assembly

lamp body

motor

bracket motor

bracket dimmer

lens

lens holder

mask assembly

PES reflector

matched mirror

purfling assembly

inner lens

reflector

connecting piece

rear cover assembly

breather cap

lamp holder

lamp bulb

purfling

panel

Figure 7. The structure tree of auto headlamp.

International Journal of Production Research 6305

� Product/part name: Expressing a product or part name which is developed. � Development process: Recording the stage for a product development activity. Its sub-concept is activity name,

participant name, task name, et al.

headlamp

Know-what Know-why Know-how Know-who

Features Plans

Markets

Persons Resources Rule Projects

Persons Projects

Persons Roles

Projects

is_ow ned_

by

be lo

ng _t

o

is_ow n

ed _b

y

b elon

g_to

is_related_with

belong_to

b el

on g_

to

belong_tois_related_with

is_ow ned_by

b elon

g_to belong_to

is_related_with

belong_to

has_p roper

ty

ha s_f

un cti

on

ha s_

fe at

ur e has_plan

is_related_to

Structures Departments

FunctionsFunctions

Data_ properties

Departments

Workgroups

Figure 8. Headlamp product knowledge ontology.

Figure 9. Representation model for panel development knowledge.

6306 Z.Y. Wu et al.

� Knowledge object: Recording the linking address of knowledge resource and knowledge content. This is the main part of knowledge integration ontology.

� Development object: Recording the development object information which is object name, product name and process name.

� Essential information: Presenting information related to a product or part, including function, input, output, con- straint and resource (Figure 11).

5.3 Method application of knowledge sharing for product development

When a product developer requests knowledge during the development of a product or a part, he connects into the knowledge sharing network, the first thing is to find some help form knowledge repository. There are defined knowledge resources in knowledge repository. The similarities between product developer and knowledge ontology are calculated. Then the next step is to filter the supplied knowledge queried from knowledge resource. The top-five most similar knowledge resources are selected for product developer. The process of knowledge sharing for product development is shown in Figures 12–16.

(1) Parts of the headlamp product development knowledge are shown in Figure 12. There are eight topics in Figure 12 that include motorcycle type, material, CAD, 3D model, colour, structure design, surface and hard- ness. Then knowledge ontology is divided into four types: know-what, know-how, know-why and know-who.

(2) The headlamp product development knowledge integration ontology is shown in Figure 13. It describes the related knowledge of the headlamp product development created by Protégé.

(3) In Figure 14, it presents the knowledge topic 5 and its similarity value from Figure 12 which consists of cate- gory knowledge, colour feature knowledge, colour suitability knowledge and file updater/creator.

(4) Finally, Figure 15 summarises the query and knowledge sharing result of product development and the related knowledge of the headlamp product development.

panelcolor_ know-x

color

Knowledge object

Know-what

Know-why

Know-how

Know-who

Part_of

Part_of

Contains

Contains

Contains

Contains

preliminary design

activities participants task

panel

Part_of

Part_of

Is_a Is_a Is_a

Song

color design headlamp

designWu

Xu

Figure 10. The knowledge integration ontology for panel product development.

International Journal of Production Research 6307

6. Experiment

An experiment is needed to be designed to evaluate the method proposed in this paper. Zhen and Jiang’s experiment (Zhen, Jiang, and Song 2011), which is to evaluate the performance of the proposed active knowledge supply model for distributed knowledge sharing in their paper greatly inspires the authors. They have made an excellent work. The exper- iment designed by the authors consists of defining and explaining of the data-set, metrics and methodologies, results and analysis from the experiment.

6.1 Data-set

The authors design some experiments based on data collected from a manufacturing enterprise in Shanghai, China. The data which is collected records 100 product development engineers’ knowledge browsing content tracks in two months. There are almost 20,000 knowledge resources in the knowledge repository in the centre server of the product develop- ment department. On a knowledge portal system, product development engineers share and query many knowledge con- tents in work day. The data-set consist of 53,624 tracking data records. And these records note which knowledge file has been read by which user on which time. All the knowledge resources have a uniform presentation format described in knowledge presentation section above. The uniform presentation format information for knowledge is one of basis for this experiments’ implementation.

product/part name: panel

developmemt process

knowledge object

development object

essential information

Activities: color design

Participants: song,xu,wu

Task: headlamp design

Object name:panel

Product name: headlamp

Process name :preliminary design

Knowledge name: know-how

Knowledge content: CAD

Function: panel color design

Output: panel product model

Input: panel product model

Constraint: headlamp contracture

Resource: design repository

Relationship

is_part_of

belong_to

has_function

has_participant_in

Figure 11. The knowledge integration ontology schema.

Topic1: motorcycle type motorcycle type information

target customer information

Topic2: mate rial panel material information physical property knowledge pressure rules knowledge material experts Topic3: CAD technical parameters modeling knowledge modeling skill knowledge file updater/creator Topic4: 3D mode l technical parameters modeling knowledge modeling skill knowledge file updater/creator Topic5: color category knowledge color feature knowledge color fitness knowledge file updater/creator Topic6: structural de sign panel structural knowledge external dimension

knowledge pressure rules knowledge designers/updater

Topic7: surface pattern type knowledge smoothness knowledge surface curve knowledge designers/updater Topic8: hardne ss hardness knowledge hardness function knowledge hardness rules

knowledge file updater/creator

Cate gorie s Topic

Know-what knowle dge know-why knowle dge know-how knowle dge know-who knowle dge

Figure 12. Headlamp product development knowledge.

6308 Z.Y. Wu et al.

6.2 Metrics and methodology

The collected 60 days’ data-set is divided into two parts: matching set and testing set. The matching set contains the ear- lier 40 days’ data records, which was used to select similar colleagues for each product development engineer; and

Figure 13. Knowledge integration ontology of headlamp product development.

category knowledge color feature knowledge color suitability knowledge file updater/creator

pressure, time curve material property knowledge pressure rules knowledge material experts

forming process knowledge physical size formula time rules knowledge file updater/creator

fuzzy membership degree length calculation formula transparency knowledge file updater/creator

physical parameter knowledge thickness calculation formula dustproof knowledge file updater/creator

mechanical properties table hardness calculation formula high temperature rules designers/updater

FMEA knowledge Cav No table surface curve knowledge designers/updater

material PTV curve physical dimension knowledge low temperature rules designers/updater

Know-what knowle dge know-why knowle dge know-how knowle dge know-who knowle dge

category knowledge color feature knowledge color suitability knowledge file updater/creator 0.684 0 0.698 0 0 0.353 0 0 0 0 0 0.825 0 0.463 0.345 0 0.267 0 0.58 0.454 0 0 0.589 0.896 0.12 0 0.577 0.832

Know-what knowle dge know-why knowle dge know-how knowle dge know-who knowle dge

Topic 5:color

Similarity value

Figure 14. Knowledge ontology mapping result for headlamp colour development.

International Journal of Production Research 6309

testing set contains the latter 20 days’ data records, in which if a product development engineer has knowledge in read- ing track records, it would be stored as the engineer’s connection in the knowledge sharing network. The results of knowledge supplied to user will be compared with the user knowledge browsing track records on that day. If the prod- uct development engineer i has knowledge track records in j days during the latter 20 days (testing period, j � 20). In these j days, we can obtain j sets of supplied or recommended knowledge, which are denoted by Rm;im 2 ½1; j�. The actual track records for product development engineer i are denoted by Tm;im 2 ½1; j�. For the product development engi- neer i, the precision based on the method proposed by Zhen et al. is defined as follows (Zhen, Jiang, and Song 2011):

Pi ¼ Sj

m¼1 Rm;i � �

\ Sjm¼1 Tm;i � ��� ��Sj

m¼1 Rm;i �� ��

In this formula, the numerator denotes the number of common data units in both the sets: Sj

m¼1 Rm;i and Sj

m¼1 Tm;i. The denominator denotes the number of data units contained in the set

Sj m¼1 Rm;i. The final evaluating indicator ‘Preci-

sion’ is derived by the average of Pi, P ¼ PN

i¼1 Pi N

, where N is the number of product development engineer in this exper- iments and N ¼ 100.

Obviously, a product development engineer’s knowledge browse data is more than knowledge recommendation data. In another words, the set T is larger than the set R, which makes recall value very small. So the recall is not contained in this experiment.

6.3 Results and analysis

The authors design the size of team varying from 5 to 50 in this experiment, and the queue length varying from 10 to 40. From this experiment, the sensitivity for team size could be analysed. From Figure 16, it could be concluded that the precision enlarges with the growing team size and tends to a steady value when the size is larger than 20. The authors set 20 as a proper size of the team. Because that too large of the team size may have too much irrelevant knowledge for product development engineers. Too large of the team size will damage knowledge sharing effectiveness.

In Figure 16, there are four curves displaying four queue lengths from 10 to 40, in which the influence of the queue length on the final results can be reflected. From Figure 16, we can know that the quality of knowledge sharing

Figure 15. Related knowledge of the product development knowledge.

6310 Z.Y. Wu et al.

improves visibly as the queue length increases. But we can also know from Figure 16 that when queue length exceeds 30, the improvement effect slows down and even has a worse trend.

Actually, the authors propose a knowledge sharing model that is based on product development engineers’ personal knowledge repositories in product development team. However, the traditional knowledge sharing environment is based on the centralised knowledge repository, in which knowledge resources are distributed to product development engineers according to their demands. Then the authors design some experiments to compare the personal knowledge repositories- based knowledge sharing model with the knowledge distribution model based on the centralised knowledge repository.

The proposed knowledge sharing model in this paper and the traditional model are based on same data-set. In this experiment, the matching set of the data is use to generate the demands of product development engineers, then the selected knowledge is sent to some product development engineers. The precision would be calculated to compare the sent knowledge to the knowledge actually read by the product development engineers in the testing data-set generated in testing period. The final evaluating indicator ‘Precision’ is calculated to measure the quality for method proposed in this paper and the traditional knowledge sharing method. The size of team is set from 5 to 50.

The results of the comparison experiments are shown in Figure 17. From Figure 17, we can see that the knowledge sharing model proposed in this paper has an advantage over the traditional centralised knowledge sharing mode.

The results validate the efficiency of the proposed knowledge sharing systems. However, the experiment in this paper is a simulation test. It doesn’t build on the base of product development engineers’ real routines. From this exper- iment results, the knowledge sharing quality could be analysed under different parameters.

Figure 16. The results under different sizes of teams.

Figure 17. The comparison between the proposed method and the traditional method.

International Journal of Production Research 6311

7. Conclusion and future perspective

This study develops knowledge integration and sharing approach for complex product development that effectively inte- grates product development knowledge distributed among various information systems and product developers with the ability to share product development knowledge. The detailed results and contributions of this study are as follows:

(1) This study establishes product development knowledge integration ontology. Additionally, the ontology schema is also designed that includes product/part name, development process, knowledge object, develop- ment object, essential information and relationship. The core techniques involved in the integration and shar- ing method designed for product development knowledge include development of selecting the target knowledge for knowledge query and design of a filtering method for supplied knowledge for developers. These techniques are developed to achieve ontology-based knowledge integration and sharing during product development.

(2) Based on the proposed integration and sharing methods for product development knowledge, the knowledge integration and sharing method for product development was implemented to demonstrate the feasibility of the proposed methods.

However, there are still some limitations for the current model and methods:

(1) The method proposed in this paper lies in that all engineers in enterprise are willing to share their knowledge with each other. In reality, lots of engineers want to receive more knowledge from others than to share their knowledge with others. Then this knowledge sharing model will not work well. So, the enterprise should to establish some incentive mechanisms embedded into this model to improve its practicality.

(2) This study only considers the similarity relationship among engineers. Actually, the relationships between engineers are very complex. In order to make the knowledge sharing network more intelligent, we should consider some semantic relationships between engineers.

(3) In addition, the scale of the experiments in this paper is not large enough. The number of product develop- ment engineers and knowledge resources that are involved in this study are a bit small.

Future work of the proposed approach include: to develop knowledge configuration and sharing mechanism accord- ing to user requirement, to construct knowledge recommendation mechanism based on product development process. Finally, detailed empirical studies for various types of product development projects are sought in order to further vali- date the representation model proposed in this paper.

Acknowledgements The author would like to thank Shanghai Research Center for industrial Informatics for the funding support to this research.

References

Bernard, A., and Y. Xu. 2009. “An Integrated Knowledge Reference System for Product Development.” CIRP Annals – Manufactur- ing Technology 58 (1): 119–122.

Bradfield, D. J., and J. X. Gao. 2007. “A Methodology to Facilitate Knowledge Sharing in the New Product Development Process.” International Journal of Production Research 45 (7): 1489–1504.

Cai, S., M. Goh, R. de Souza, and G. Li. 2012. “Knowledge Sharing in Collaborative Supply Chains: Twin Effects of Trust and Power.” International Journal of Production Research 51 (7): 2060–2076.

Chau, T., S. Moghimi, and M. R. Popovic. 2013. “Knowledge Translation in Rehabilitation Engineering Research and Development: A Knowledge Ecosystem Framework.” Archives of Physical Medicine and Rehabilitation 94 (1, Supplement): S9–S19.

Chen, Y.-J. 2010. “Knowledge Integration and Sharing for Collaborative Molding Product Design and Process Development.” Com- puters in Industry 61 (7): 659–675.

Chen, Y.-J., Y.-M. Chen, and H.-C. Chu. 2009. “Development of a Mechanism for Ontology-based Product Lifecycle Knowledge Integration.” Expert Systems with Applications 36 (2): 2759–2779.

Chen, Y.-J., Y.-M. Chen, and M.-S. Wu. 2012. “An Empirical Knowledge Management Framework for Professional Virtual Commu- nity in Knowledge-intensive Service Industries.” Expert Systems with Applications 39 (18): 13135–13147.

Dekkers, R., C. M. Chang, and J. Kreutzfeldt. 2013. “The Interface between ‘Product Design and Engineering’ and Manufacturing: A Review of the Literature and Empirical Evidence.” International Journal of Production Economics 144 (1): 316–333.

Denkena, B., M. Shpitalni, P. Kowalski, G. Molcho, and Y. Zipori. 2007. “Knowledge Management in Process Planning.” CIRP Annals – Manufacturing Technology 56 (1): 175–180.

6312 Z.Y. Wu et al.

Foray, D., and B. A. Lundvall. 1997. The Knowledge Based Economy. Paris: OCDE. Hou, J.-L., and S.-T. Pai. 2009. “A Spatial Knowledge Sharing Platform. Using the Visualization Approach.” International Journal of

Production Research 47 (1): 25–50. Huang, S.-L., S.-C. Lin, and Y.-C. Chan. 2012. “Investigating Effectiveness and User Acceptance of Semantic Social Tagging for

Knowledge Sharing.” Information Processing & Management 48 (4): 599–617. Jayaram, J., and S. Pathak. 2012. “A Holistic View of Knowledge Integration in Collaborative Supply Chains.” International Journal

of Production Research 51 (7): 1958–1972. Kim, K.-Y., and Y. S. Kim. 2011. “Causal Design Knowledge: Alternative Representation Method for Product Development Knowl-

edge Management.” Computer-Aided Design 43 (9): 1137–1153. Li, Y.-M., T.-F. Liao, and C.-Y. Lai. 2012. “A Social Recommender Mechanism for Improving Knowledge Sharing in Online For-

ums.” Information Processing & Management 48 (5): 978–994. Liao, S. H. 2005. “Expert System Methodologies and Applications – A Decade Review from 1995 to 2004.” Expert Systems with

Applications 28 (1): 93–103. Liao, S.-H., and C.-C. Wu. 2010. “System Perspective of Knowledge Management, Organizational Learning, and Organizational Inno-

vation.” Expert Systems with Applications 37 (2): 1096–1103. Lin, F.-R., and C.-M. Hsueh. 2006. “Knowledge Map Creation and Maintenance for Virtual Communities of Practice.” Information

Processing & Management 42 (2): 551–568. Martín, A., C. León, J. Luque, and I. Monedero. 2012. “A Framework for Development of Integrated Intelligent Knowledge for Man-

agement of Telecommunication Networks.” Expert Systems with Applications 39 (10): 9264–9274. Mountney, S. L., J. X. Gao, and S. Wiseall. 2007. “A Knowledge System to Support Manufacturing Knowledge during Preliminary

Design.” International Journal of Production Research 45 (7): 1521–1537. Shaw, D., and J. S. Edwards. 2006. “Manufacturing Knowledge Management Strategy.” International Journal of Production Research

44 (10): 1907–1925. Tang, D., R. Zhu, J. Tang, R. Xu, and R. He. 2010. “Product Design Knowledge Management Based on Design Structure Matrix.”

Advanced Engineering Informatics 24 (2): 159–166. Urwin, E. N., and R. I. M. Young. 2013. “The Reuse of Machining Knowledge to Improve Designer Awareness through the Configu-

ration of Knowledge Libraries in PLM.” International Journal of Production Research 52 (2): 595–615. Villela Dantas, J. R., and P. P. Muniz Farias. 2010. “Conceptual Navigation in Knowledge Management Environments Using

NavCon.” Information Processing & Management 46 (4): 413–425. Wang, S., and R. A. Noe. 2010. “Knowledge Sharing: A Review and Directions for Future Research.” Human Resource Management

Review 20 (2): 115–131. Wang, K.-J., V. S. Jha, D.-C. Gong, T. C. Hou, and C.-C. Chiu. 2010. “Agent-based Knowledge Management System with APQP:

Implementation of Semiconductor Manufacturing Services Industry.” International Journal of Production Research 48 (10): 2913–2936.

Young, R. I. M., A. G. Gunendran, A. F. Cutting-Decelle, and M. Gruninger. 2007. “Manufacturing Knowledge Sharing in PLM: A Progression towards the Use of Heavy Weight Ontologies.” International Journal of Production Research 45 (7): 1505–1519.

Zhang, D., D. Hu, Y. Xu, and H. Zhang. 2012. “A Framework for Design Knowledge Management and Reuse for Product-service Systems in Construction Machinery Industry.” Computers in Industry 63 (4): 328–337.

Zhen, L., and Z. Jiang. 2008. “Innovation-oriented Knowledge Query in Knowledge Grid.” Journal of Information Science and Engi- neering 24 (2): 601–613.

Zhen, L., G. Q. Huang, and Z. Jiang. 2009. “Recommender System Based on Workflow.” Decision Support Systems 48 (1): 237–245. Zhen, L., Z. Jiang, and H. Song. 2010. “Distributed Recommender for Peer-to-peer Knowledge Sharing.” Information Sciences 180

(18): 3546–3561. Zhen, L., Z. Jiang, and H. Song. 2011. “Distributed Knowledge Sharing for Collaborative Product Development.” International

Journal of Production Research 49 (10): 2959–2976. Zhen, L., H. Song, and J. He. 2012. “Recommender Systems for Personal Knowledge Management in Collaborative Environments.”

Expert Systems with Applications 39 (16): 12536–12542.

International Journal of Production Research 6313

Copyright of International Journal of Production Research is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.