ORIGINAL ARTICLE

Key technologies for ASP-based product customizationservice system for SMEs: a case study

Yan Su & Wenhe Liao & Yu Guo & Qiulin Ding

Received: 4 January 2007 /Accepted: 29 May 2008 / Published online: 17 September 2008# The Author(s) 2008. This article is published with open access at Springerlink.com

Abstract Current product customization systems seldomconsider the whole process of product customization, norprovide integrated application services for small to medi-um-sized enterprises (SMEs). In this paper, an ASP-basedproduct customization service system with integratedapplication services is proposed for SMEs. A lifecycle-oriented customization service mode is presented tosmoothly drive the whole process of product customization.For supporting product customization effectively andefficiently, key technologies on distributed product resourcesharing, distributed product data reorganization, and prod-uct configuration are deeply studied. A dynamic resourcepublishing method and a general resource evaluation modelwith reward measures to promote resource sharing aregiven. A method of XML-based data mapping for productstructure tree reorganization is proposed. A parameter graycorrelation degree algorithm is established for quickproduct configuration. To greatly meet customers’ individ-ual requirements, an interactive virtual product configura-

tion platform is designed. Finally, product customizationservice system orienting to construction machinery isdeveloped and validates the above studies.

Keywords Product customization . Application serviceprovider (ASP) . Small tomedium-sized enterprises (SMEs) .

Resource evaluation . Product configuration

1 Introduction

Mass customization, which refers to the ability of providingcustomized products or services through flexible processesat reasonably low cost with high quality [1], has become acompetitive strategy by an increasing number of companiesin highly segmented global service market [2]. In recentyears, product customization has attracted great interest dueto the emergence of mass customization [3]. Networkedproduct customization systems as the bridge connectingenterprises with customers are important for enterprises tobring out innovative and profitable products to the marketquickly and cause mass customization (MC) enterprisesincreasing attention. Some MC enterprises have self-developed their own networked product customizationsystem. For example, General Motors Corp. and FordMotors Corp. developed their own Internet-based motorcustomization system, respectively; Dell Inc. provides acustomization service in their own website; Haier Corp.developed a networked product order system. A successfulnetworked product customization system needs an enter-prise put in a great amount of capital, technology,manpower, and time into system development and mainte-nance. It is easy to drive MC enterprises away fromconcentrating on their efforts on core competence duringself-developing and maintaining networked product cus-

Int J Adv Manuf Technol (2009) 42:381–397DOI 10.1007/s00170-008-1589-z

Y. Su (*)College of Civil Aviation,Nanjing University of Aeronautics & Astronautics,No 29 Yudao Street,Nanjing, Jiangsu 210016, People’s Republic of Chinae-mail: suyannj@nuaa.edu.cn

W. Liao :Y. GuoCollege of Mechanical & Electrical Engineering,Nanjing University of Aeronautics & Astronautics,No 29 Yudao Street,Nanjing, Jiangsu 210016, People’s Republic of China

Q. DingCollege of Information Science & Technology,Nanjing University of Aeronautics & Astronautics,No 29 Yudao Street,Nanjing, Jiangsu 210016, People’s Republic of China

tomization system. Otherwise, configurable product re-source on self-developed product customization systemmainly came from the MC enterprise itself. The limitedproduct resource is sometimes deficient to well satisfycustomers’ needs. However, customers are usually weary oflogging in to different systems many times. Customersmore intend to get their products or services on anintegrated service platform, only logging in once.

In China, about 90% of registered companies aresmall to medium-sized enterprises (SMEs). Most of theseSMEs may not make the socially optimal investment inproduct customization. It is even difficult for SMEs todevelop Internet-based customization systems because oftechnology difficulties [4]. Fact reports [5] that SMEs inChina can not keep up with new information technology,and most of them do have insufficient money to buy andno technical capability to utilize advanced software suchas CAD, CAM, PDM, SCM, and CRM. It is difficult forSMEs to develop and maintain its own networked productcustomization system for lack of the capital, IT staffs, andspecialists to manage product resource and to supportsales personnel. It is necessary to find an easier andcheaper operation mode to implement networked productcustomization. An integrated service platform, whichevolves e-business environment by integrating knowledge,technologies, and resources among customers, suppliers,and other business partners to improve SMEs’ competitivecapability and meet customers’ individual demands effi-ciently, has been a trend in recent years [6].

The newly emerged mode of application service provider(ASP), which is viewed as a subset of e-commerce [7], inmany cases, is a third party service organization to offerproducts of software and hardware through networks. ASPis also defined as a single point of contact for all thetelecommunications, hardware, software, and consultingnecessary to deploy, run, and maintain hosted applicationsremotely [8]. Its core value propositions are to lower totalcost of ownership, make monthly fees predictable, reducetime-to-market, provide access to market-leading applica-tions, and allow businesses to focus on their corecompetencies [9]. ASPs particularly target SMEs byproviding applications that these firms need to adopt, butcannot afford or are commonly incapable of developing andmaintaining [10]. The ASP mode is also helpful by settingup an ASP service platform that provides services such assystem development, system integration, and technologysupport [11–13]. For above characteristics, ASP has beenone of the high points of e-commerce in the twenty-firstcentury and is increasingly welcomed by many SMEswhich typically suffer from resource poverty and the lack ofIT capabilities [14]. Software application using an ASPinstead of the customer’s local installation is expandingrapidly. For example, Zhou et al. [15] built up a remote

ASP-oriented simulation tool on the World Wide Web. Withthe growth in the development of distributed manufacturingsystems, Huang et al. [16] have developed an integratednetwork management framework based on mobile agenttechnology for SMEs to reduce complexity and expense inmanaging networked manufacturing devices.

However, the integration of the product customizationsystem and ASP mode is seldom considered on currentresearches of product customization systems. After analyz-ing the status of existing networked product customizationsystem and the features of ASP mode, the applicationpredominance of ASP mode in product customizationsystem can conclude as follows:

(a) Resource predominance. Information technology andmanagement methods are two important aspects whichhave made mass customization a standard businesspractice [17]. Networked product customization sys-tem is, first, an information system. Networkedproduct customization system running in ASP modecan promote distributive resource sharing. Productresource comes from different enterprises and evendifferent industries and will be more abundant thanthat from single enterprise self-developed custom-ization system.

(b) Cost predominance. ASPs can not only bear the cost onsystem development, system maintenance and softwareupdate, but also provide software service, techniquesupport, and production equipment at low rent.

(c) Efficiency predominance. Since ASPs may serve alarge customer base with different customizationrequirements, a product designed to facilitate custom-ization may be easier to deliver through ASP.

(d) Market predominance. ASPs provide services to wideapplication fields, including geographically distribu-tive enterprises and even differently industrial enter-prises. Any users who can log in to Internet can accessASP services through the portal of ASP serviceplatform. The more enterprises join in to ASP serviceplatform, the more customers are attracted to log in toit, and, thus, it has a wide market.

So, we apply ASP mode in networked product custom-ization systems and propose an ASP-based product custom-ization service system for SMEs. The main motivation fordeveloping ASP-based product customization service sys-tem is to provide a lifecycle-oriented product customizationservice to assist SMEs effectively conducting mass custom-ization. The system structure is established and lifecycle-oriented product customization mode is presented. Keytechnologies for effective and quick product customizationare deeply studied, i.e., resource sharing technology,product data transform technology, product configurationtechnology, and interactive virtual product configuration

382 Int J Adv Manuf Technol (2009) 42:381–397

platform. To promote distributive product resource sharing,a general resource evaluation model is established. Forintegrated product, data from distributed SMEs areusually expressed in different file formats; an XML-based data mapping method for reorganizing productstructure tree is presented. Considering the uncertaintyand diversification of customers’ demands, gray correla-tive technology is applied to product instantiationconfiguration and a parameter gray correlative-degreealgorithm model is built to provide customers contentproducts quickly. To meet customers’ individualdemands, an interactive virtual product configurationplatform is designed. Finally, a system is developedusing the above technologies and has been successfullyapplied in construction machinery.

2 System structure

ASP-based product customization service system mainlyserves as a software resource rent center, a design/manufacture resource sharing center, and a technologyconsultation center. It provides open, reconfigurable, andweb-enabled applications to support product customizationfor SMEs. The framework of ASP-based product custom-

ization service system includes four layers: resourcesupport layer, management layer, application service layer,and user layer, as shown in Fig. 1. The four layers cover thetotal-life-cycle of product customization, from requirementsanalysis, product configuration, design, manufacture tomanaging and marketing.

As the foundation of ASP-based product customizationservice system, resource support layer provides data andresource for system running. It includes system database,sharing resource base, and product configuration knowl-edge base. The sharing resource base, which consists of anumber of product development software, manufacturedresources and enterprise information, is available to allenterprises in industry. System database are used to drivethe whole service system running smoothly. Productconfiguration knowledge base is an important part ofresource support layer and supplies sufficient knowledgeof product configuration. It includes case base, rule base,graphics base, and products database.

The management layer aims to ensure that the platformis running smoothly and all function modules well-servingthe SMEs. It consists of four main function modules whichare used to manage this service system. They areconfiguring process management module, run controlmodule, data transform module, resource management

Fig. 1 Structure of ASP-basedproduct customization servicesystem

Int J Adv Manuf Technol (2009) 42:381–397 383

module, system maintenance module, and user rightmanagement module. Configuring process managementmodule ensures that the configuring process smoothlyand successfully provides products to satisfied customers.The resource management module serves mainly forproduct customization and emphasizes particularly onmanagement of product hierarchy tree, product data, andproduct version. Run control and system maintenancemodules are designed to make system and all functionmodules act smoothly. Isomeric/isomorphic product datafrom distributed customers or enterprises are transformedinto uniform data format through data transform module.Any users who want to log in to the ASP service platformmust be managed and controlled by user right manage-ment module.

Application service layer is the core layer of thisplatform and provides three kinds of services for threepurposes. The first purpose is to meet SMEs’ resourcedemands of design and manufacture, including advancedsoftware sharing (e.g., Computer Aided Design (CAD),Computer Aided Manufacturing (CAM), Supply ChainManagement (SCM), Customer Relationship Management(CRM)) lease, design/manufacture knowledge sharing,product marketing, and technology consultation. Thesecond purpose is to integrate distributed resources ownedby SMEs, scientific institutions, colleges and universities,etc. The third purpose is to provide users (SMEs orcustomers) product customization functions, includingrequirement analyzing, case searching, design type ana-lyzing, virtual product configuration, and configurationvalidity analyzing, etc.

The user layer is the portal which users access servicesprovided by ASP service platform. It presents users thefollowing interactive function interfaces: informationupload/download, software tools upload/download, re-

source evaluation, requirements analysis, product config-uration, product assembly and disassembly, design validityanalysis, configuration validity analysis, and so on. All theinterfaces are embedded on ASP. Using these interfaces,SMEs, and other customers can access all the tools andapplication services provided by product customizationservice system.

High-tech process innovation (e.g., computer hardwareand software) offers enterprises the potential of reducingproduction costs and enhancing product quality. Thewhole product customization service system is supportedby scientific institutions, colleges and universities, andadministrative organizations, which can well act as high-tech software and technology service center. In addition, itaims to provide not only quick product customizationfunctions but also design/manufacture resources at lowrent. SMEs could develop their Internet-based sales andcustomization effectively in a short time and at low costby using this platform.

3 Lifecycle-oriented product customization service mode

Efficient customization service mode can shorten the timeof product customization and reduce the cost consequently.So, we proposed a lifecycle-oriented product customizationservice mode that can help cut product customization time.The customization service mode is shown in Fig. 2.

Customer requirement parameters have strong correla-tion to enterprises’ parameters requirements, and productcustomization is driven by customer requirements. Whencustomers log on ASP-based product customization servicesystem and input requirements’ parameters, the datatransform module parses them into XML documents. Thencase-based searching module makes the first search (as

Fig. 2 Lifecycle-orientedproduct customization mode

384 Int J Adv Manuf Technol (2009) 42:381–397

search1 in Fig. 2) for counterpart case from resourcesupport layer. If the counterpart is not found, the analyzerstarts up and analyzes exported XML model documentsfrom CAX scene. For assembly product, each child node inXML model documents is captured for next search. If thereis no counterpart or similar counterpart and the child nodecan be divided further, the second search (as search2 inFig. 2) continues until gaining the counterpart or similar,or until the node not being divided any more. When asimilar product is achieved, the approach of partialimproving design is adopted. If none of the counterpartor similar is found in the database, innovation design isappointed. Distributed design groups who receive designtasks can rent software tools of product developmentprovided by the product customization service systemweekly or monthly. Domain experts take part inevaluating the design results together with customersand designers in interactive virtual product configurationenvironment. Design results of assembly parts arereassembled and submitted to the analyzer for validityanalysis. All finished parts and assembly parts aresubmitted into product configuration knowledge base forfuture reuse. While the customer is content with theresults of the design, product orders are sent to remotedistributed manufactories for production. The ASP-basedproduct customization integrated service platform wellprovides information and technologies for distributeddesigners and manufacturers at low rent.

4 Key technologies

It can be seen from Figs. 1 and 2 that requirement analysis,distributed resource sharing, product data transform, andproduct configuration are four important aspects forsupporting product customization effectively. Requirementanalysis has been deeply studied in [18, 19]. In thefollowing, resource sharing mechanism, XML data map-ping for creating product structure tree, quick retrievingalgorithm and interactive virtual product configurationplatform for product configuration are focused on.

4.1 Distributed resource sharing mechanism

Regardless of which category is focused on and whichmethod or logical structures are adopted, the base andcore element determining the failure or success of productcustomization is product resources [20] including designresources, manufacturing resources, product cases, productdata, etc. Lack of sufficient resource is also one of theimportant factors that the business mode of applicationservice providers in China cannot diffuse effectively asexpected [21]. Since a product platform on the proposed

product customization service platform is structured fromdistributed enterprises, how to promote geographicallydistributed enterprises to submit their product resourcesonto the integrated ASP service platform is a bottleneckproblem of the product. Current researches in the field ofdistributed resource sharing are mainly focused on resourceintegration system design, information extraction technolo-gy, and knowledge resource management. For example,Wang et al. [22] proposed a resource integration frameworkof regional networked manufacturing ASP service platform.Zha [23] developed a web-enabled database system formicro electro-mechanical system which can provide thenetworked design and manufacturing services over theInternet. Reinoso-Castillo [24] presents a novel ontology-driven approach to support customizable informationextraction and integration in scientific discovery. Tointegrate the legacy of manufacturing resources intonetworked manufacturing systems, Zhou and Jiang [25]develop a novel platform for the configuration andencapsulation of legacy manufacturing resources in net-worked manufacturing systems. Knowledge resource man-agement has experienced two waves. Through the firstwave of knowledge management and the second wave ofknowledge management, large numbers of researches havefocused on knowledge resource management [26]. But fewresearches target establishing resource sharing mecha-nism to promote resource-holders to actively submittingtheir product resources. To achieve that, a “win–win”strategy for ASPs and SMEs is expected to be a goodmeasure, i.e., submitted resource must be useful forASPs and SMEs who submit useful resources shouldobtain certain rewards.

Based on the above analysis, following this paper, adynamic resource publishing method for distributedresource sharing is given (see Fig. 3). Original resourceson the resource support layer are collected by adminis-trators from typical members, relative domain institutions,ongoing Internet resources, and relative technique booksand periodicals. Most product resources and the update ofservice resources need to be done dynamically on clientsby users. Platform administrators have the highest right tooperate the whole resource, involving constructing initialresource database and integrating distributed resources.All non-members must register and pay certain money forspecial resources or for resource space to deposit theirprivate resources into database. Any uploaded resourcewill be taken into the database when it is evaluated to beuseful. When the added resource is evaluated to be avalue-added resource, a certain reward is given to theresource provider. It can be seen from the resourcepublishing process that resource evaluation model andcertain award measures are important aspects to promoteresource sharing.

Int J Adv Manuf Technol (2009) 42:381–397 385

4.1.1 General resource evaluation model

Existing resource evaluation models are established for aspecial resource with certain evaluation indices, such ascollaborative design resource evaluation [27], manufac-turing resource selection evaluation [28], Web informationresource evaluation [29], and injection mould cost evalu-ation [30], so on. Product resources on ASP serviceplatform comes from different enterprises and evendifferent industries. The category and quantity of productresources are abundant. Establishing an evaluation modelfor each category of product resource is not practical. So, ageneral resource evaluation model is established toevaluate the value of resources submitted from clients.Evaluation indices and evaluation algorithm are establishedas follows:

Hypothetically set U is an evaluation index,

U ¼ u1; u2; � � � ; ukð Þ;

where ui (i=1, 2,…, k) is assigned by evaluation experts. Ifset W is a relevant weight aggregate of U and wi is theweight of element ui, then W ¼ w1;w2; � � � ;wkð Þ; and

Pki¼1

wi ¼ 1. Each index can be divided into different grades

according resource categories. Each grade is marked byevaluation experts. Let R* be the mark matrix of U and Pbe the matrix of the number of evaluation experts, then

R* ¼r11 r12 � � � r1mr21 r22 � � � r2m

� � � � � � ... � � �

rk1 rk2 � � � rkm

26664

37775; P ¼

N11 N12 � � � N1m

N21 N22 � � � N2m

� � � � � � ... � � �

Nk1 Nk2 � � � Nkm

26664

37775

Let A be the evaluation matrix, then

A ¼ W � R* � P ¼ W � R* � PT� � ¼ a1; a2; a3; a4; a5ð Þ:

Log files memorize the frequency t of submittingresource and the number of resource provided by eachresource-holder at a time. The context of resources andother information are memorized in temporary informationarea. Domain experts and task-assigner mark each item.Generally, the sum of resources provided by each resource-holder changes along with time t. Let g(t,t) be number of

Fig. 3 Dynamic resourcepublishing

386 Int J Adv Manuf Technol (2009) 42:381–397

resources per second at a time and Gj ( j=1, 2,…, k) be thetotal quantity of the j th kind of resource, then

Gj ¼ZZ

g t; tð Þ d t d t;

Suppose ξi (i=1, 2,…, k) is the value of unit mark and N isthe number of experts who attend evaluation. ξi isappointed by task-assigner. Let Val be the evaluation resultof resource value, then

N ¼ PPNij; i ¼ 1; � � � ; k; j ¼; � � � ;m;

Val ¼ x1G1; x2G2; � � � ; xkGkð Þ � A=N-

Set l=(l1, l2, l3) represents the grade of useful degree, wherel1, l2, and l3, respectively, represent very useful, useful, and

useless. Therein, 0≤li≤1 andP3i¼1

li ¼ 1. Hypothetically,

relative useful degree Vper ¼ Val

� Pki¼1

xiGið Þ� �

, and

the lower limit of value-added resource is δ, which isappointed by experts, then

If Vper≥l1, the resource is very useful;If l2≤Vper≤l1, the resource is useful;If Vper≤l3, the resource is useless;If the resource is useful and Vper≤δ, the resource isvalue-added.

Uploaded resource information from non-administratorsis first deposited in a temporary information area. Task-assigners and experts start up resource evaluation module toevaluate uploaded resources and give the results of usefuldegree and value-added degree. Useful resources areaccepted and deposited in sharing resource base, productconfiguration knowledge base, or system database respec-tively according to resource characters.

4.1.2 Awarding measures

To encourage and attract resource-holders to submit productresources and more customers to feedback helpful infor-mation, three reward measures are presented according tothe value of Val.

(a) PaymentPay cash or mail goods to resource supplier directly

according the rate of exchange of Val which assigned bytask distributor.(b) Awarded marks

Every time when the published resource is evaluated byestimation module, the calculated value is added to Val anddeposited as a type of awarded marks. The resourcesupplier can use the rewarded marks to exchange services

which are provided by ASP service platform at any timeuntil the awarded marks equal to zero.(c) Widen authorization

When Val is large enough and exceeds a certain valueappointed by the administrator, the resource supplier hasthe following choices:

& To be amember if his/her present status is non-member& To be an advanced member if he/she has been a

general member& To prolong period of validity to operate the ASP-

based product customization service system

The rate of exchange for the above-three award measuresis assigned by ASPs after forecasting the future value ofrelevant resource. ASP service platform backups andevaluates the resource submitted from client in real time.The Val is termly updated and possible reward informationis timely sent to resource-holders. Resource providers arefree to select permissive reward measures.

4.2 XML-based data mapping for product structure tree

Integrated product data from distributed SMEs are usuallyexpressed in different file formats. It is necessary totransform these isomeric product data into uniform format.Product hierarchy structure tree is an intuitive organizationform of product data and convenient for users querying andbrowsing. Current researchers mainly focus on productfamily design [31] and product information representationfor effective case retrieval [32]. Few studies emphasize onproduct data reorganization.

On ASP-based product customization service system,relational database is used as back-end database. To realizedata mapping between different file formats, XML documentis used to exchange and search among different modules.The main benefit of adopting XML format is that there arepowerful parsers to process XML files and XML format hasa clear structure to organize data [33]. Furthermore, it isbecause of its feature of standard language and itsinteroperability, XML would allow, by means of eXtensibleStylesheet Language Transformations (XSLT), the genera-tion of eXtensible HTML (XHTML) pages which arevisualized on a browser with a representation of any otherdocument that includes their elements [34].

There are two types of data mapping for creating anddepositing product hierarchy structure tree. One is the datamapping between relational database and product hierarchystructure. The other is data mapping between relationaldatabase and product data connecting with tree nodes ofproduct hierarchy structure. For the two types of datamapping, two XML-based mapping methods are proposed:

Int J Adv Manuf Technol (2009) 42:381–397 387

XML template-driven mapping and object-oriented XMLtable mapping. The two proposed mapping methods arerealized by following pre-designed XML template andobject-oriented XML table.

hierarchy structure tree. According to object-orientedmapping, data in the table or nodes of hierarchy tree aretransformed into XML documents and deposited intorelational database. Fig. 5 illustrates the above-mentionedmapping relationships in reorganizing product hierarchystructure tree of Road-surface Machinery. When the1220MAXI-PAV Paver meets a customer’s needs, productdata of 1220MAXI-PAV Paver is mapped to a producthierarchy structure tree driven by an XML template.Product data of performance parameters or others aremapped into the data table by object-oriented XML tableand connects with relevant tree nodes, such as Extension

Frame of 1220MAXI-PAV Paver. If modifications aremade, product hierarchy structure tree and data connectingwith tree nodes can also be deposited in a relationaldatabase driven by XML template and XML table.

4.3 Product configuration

4.3.1 Case-based product configuration algorithm

To solve the product configuration problem, several differentapproaches have been adopted. For example, the generic billof material (GBOM) concept to solve the problem of productconfiguration management [35], Product Family Structure(PFA) [1] and the object-oriented concept [36] to organizeand express product configuration model, and ConstraintSatisfactory Problem (CSP) Algorithm [37] to solve rapidconfiguration problem. These researches in the field mostlyfocus on the issues about creating product configurationinformation system environment and solving optimization-based problem for product family design. Reusing precioussuccessful cases can lower the error rate of configurationdesign and case-based reasoning (CBR) can help solve theproblem through the retrieval of similar previous cases [38].In the domain of product design and development, case-based reasoning algorithm is used to quickly help generateFig. 4 XML-based data mapping for product structure tree

For data-mapping principle, see Fig. 4. According tothe XML template, data are retrieved from a relationaldatabase and expressed in XML. On the front-end browser,the XML document displays in the user interface as a product

388 Int J Adv Manuf Technol (2009) 42:381–397

a right new BOM [39] or new product ideas [40] whensolving a new problem. The preconditions of the aboveproduct configuration methods are that customers’ needsmust be transformed into coincidence expression withconfigurable attribute.

As a matter of fact, customers’ needs take on uncertaintyand diversification. The transformation of customers’ needswill not only add load to the system but sometimes be hardto be unified. The goal of case-based product configurationis to find the most similar cases which meet withcustomers’ needs. Case-based product configuration pro-cess is also a process of sorting cases by the similar degree.Due to the uncertainty and diversification of customers’needs, product configuration process, as a system, isactually a gray system according to the theory of graycorrelative technology [41]. If taking attributes of custom-

ers’ needs as referent sequence and attributes of cases ascomparable sequence, case-based product configuration canbe realized using gray correlative-degree algorithm model[42]. Usually, attributes of customers’ needs and attributesof cases are discrete. Here, point gray correlative degree[42] is adopted and parameter gray correlative-degreealgorithm is established as follows:

Set the p0 attribute aggregate of customers’ needs, p0(k)the kth attribute among attribute aggregate of customers’needs, then

p0 ¼ p0 1ð Þ; p0 2ð Þ; � � � ; p0 lð Þf g

Set pi as the attribute aggregate of the ith case, pi(k) the kthattribute among attribute aggregate of the ith case. If oneattribute is absent in the attribute aggregate of a case, the

Fig. 5 Creating product struc-ture tree of road-surfacemachine

Int J Adv Manuf Technol (2009) 42:381–397 389

value of the attribute item is set to zero. All the attributeaggregates of cases are expressed as follows:

p1 ¼ p1 1ð Þ; p1 2ð Þ; � � � ; p1 lð Þf g;p2 ¼ p2 1ð Þ; p2 2ð Þ; � � � ; p2 lð Þf g;

� � � � � �pm ¼ pm 1ð Þ; pm 2ð Þ; � � � ; pm lð Þf g:

Taking p0 as the reference sequence and pi (i=1, 2,…,m) asthe comparable sequence, parameter gray correlative degreep0 between pi and is

r p0; pið Þ ¼ Plk¼1

wkr p0 kð Þ; pi kð Þð Þ

¼ Plk¼1

wk

Δmini

mink

���p0 kð Þ�pi kð Þ���þ0:5Δmax

imaxk

���p0 kð Þ�pi kð Þ���

p0 kð Þ�pi kð Þ þ0:5Δmaximaxk

���p0 kð Þ�pi kð Þ���������

ðFormula4:1Þ

where ωk is the weight of performance parameter k, 0≤ωk≤1,and

Pnk¼1

wk ¼ 1

In order to make all attributes to be compared and ensureretrieving correct results, dimensions of attributes need tobe eliminated. From attributes of customers’ needs andattributes of cases exist three types of expression, i.e. text,numerical value, and range value. Existing data transformmethods in gray correlative technology can not be directlyused. The following defines gray distances betweenattributes by different type of attributes:

(a) Text gray distance

For text attributes among referent sequence and compa-rable sequence, the improved algorithm model of textsimilarity [43] is as follows

sim OT0;OTið Þ ¼ count OT0 \ OTið Þcount OT0ð Þ þ count OTið Þ � count OT0 \ OTið Þ

ðFormula4:2ÞAnd define text gray distance

OT0;OTij j ¼ count OT0 \ OTið Þ � w0 � sTi

w0j jcount OT0ð Þ þ wij jcount OTið Þ � count OT0 \ OTið Þ � w0 � wTi

ðFormula4:3ÞWhere OT0, OTi are text attributes of p0 and text

attributes of pi respectively, the function of count is used tocompute the quantity of text attributes, ω0 and ωi, are theweight of test attribute of p0 and the weight of test attributeof pi respectively.

(b) Range gray distance

Set the attribute item of reference sequence p0(k)=[d1,d2], attribute item of comparable sequence pi(k)=[e1, e2],

wa0;k and wai;k are the weight of p0(k) and the weight ofpi(k), respectively, 0 � wa0;k � 1;0 � wai;k � 1. Improvethe definition 2.2.2 [42] and define the following rangegray distance with weight

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiwa0;k � d1 � wai;k � e1� �2 þ wa0;k � d2 � wai;k � e2

� �2P0 kð Þ; pi kð Þj j ¼ ffiffiffi2

p=2q

ðFormula4:4Þ(c) Numerical value gray distance

Set the attribute item of reference sequence p0(k)=d,attribute item of comparable sequence pi(k)=e and definenumerical value gray distance

p0 kð Þ; pi kð Þj j ¼ wa0;k � d � wai;k � e�� ��=wa0;k � d þ wai;k � e

�� ��ðFormula4:5Þ

While computing gray correlative degree betweenreference sequence and comparable sequence, gray dis-tance between attributes is calculated using the aboveFormula 4.3, Formula 4.4, or Formula 4.5 according to thetype of attribute.

From requirement-driven product customization processgraph (Fig. 2), it is obvious that searching counterparts orsimilar cases from resource support layer is one keytechnology for product customization to proceed smoothly.There are two main strategies that are used for retrieval,namely, generalization hierarchy strategy and parallelstrategy [44]. Here, we integrate generalization hierarchystrategy and parallel strategy according to product hierar-chy tree, and apply two searching algorithms for quickproduct configuration: the text similar degree algorithm(Formula 4.2) [43] and the proposed parameter graycorrelative-degree algorithm (Formula 4.1).

When a product is required to be configured, text similardegree algorithm (Formula 4.2) is applied in nodes of theproduct hierarchy structure tree. Text-name relevant to treenode is regarded as text attribute of the retrieving object. Theprocess of retrieving starts from the root-node and downwardswhen none of the counterpart or similarity is retrieved at thecurrent level and ends when the node is a leaf-node. Whenmulti-similarities of a retrieving object are found, customerrequirements are decomposed into performance parameters. Afurther search based on parameter gray correlative-degreealgorithm (Formula 4.1) is done according to performanceparameters. When a text-type parameter is met within aparameter table, text gray distance (Formula 4.3) is utilizedand the value of OT0;OTij j is added into Formula 4.1. Whenone parameter item is a range value, range gray distance(Formula 4.4) is used to calculate the similar degree of twoperformance parameters. The similar degree of two param-eters of numerical value type is computed by numericalvalue gray distance (Formula 4.5)

(Formula 4.1)

(Formula 4.2)

(Formula 4.4)

(Formula 4.3)

(Formula 4.5)

390 Int J Adv Manuf Technol (2009) 42:381–397

For example, a Cement–Concrete Paving Machinery isrequired by customers (Table 1). The text-name “Cement–Concrete Paving Machinery” is an object to be retrieved.First, text similar degree algorithm is applied and multi-similar products are retrieved (Table 2). In order to get themost similar product from Table 2, parameter graycorrelative-degree algorithm is used for further search byperformance parameters. When the item of Engine type ismet, text similar degree algorithm is utilized and the textitem is replaced by sim(OT0, OT9). Results indicate thatRP802 is the most similar. Improvement design orinnovation design can be performed upon RP802. Testsshow that case-based product configuration by use ofproposed retrieval method can significantly improve re-trieval efficiency.

4.3.2 Interactive virtual product configuration platform

At present, end users often want to configure the productwith individual preferences and are allowed to chooseproduct specifications with real-time update in the product

appearance or overall shape. Only, supplying a bill ofproduct configuration or 2D image for showing customer’sconfigurations may not be effective when the productstyling (shape, color, and appearance in 3D space) is a keyfactor that influences the purchase decision. There is a needfor web-based interactive visualization environment. Lotsof Web VR-supported interactive or collaborative systemshave been studied and developed in recent years. Kan et al.[45] employ virtual reality modeling language (VRML) forsynchronized 3D-viewing of a product design over theInternet and developed a web VR collaborative environ-ment. Cera et al [46] employ Java and Java 3D to develop acollaborative 3D environment for designers separatedgeographically to author design semantics. Huang [47]employs VRML to construct a whiteboard to display designinformation. Lau et al. [48] employ VRML to develop aninteractive environment for sharing product design data andthe visualization of design changes in real time. Zhang et al.[49] propose a Web-based framework for product informa-tion sharing and visualization. Gierach et al. [50] present anapproach for managing networked hardware devices and

Table 2 Performance parameters of multi-function bitumen-concrete paving machine cases retrieved from case base

Item/parameter Product-type

RP955 RP800J RP951A RP751 RP802 …

Basic working width 3.0 3.0 3.0 3.0 3.0 …Max working width 9.5 8.0 9.5 7.5 8.0 …Max paving thickness 350 330 350 380 380 …Paving speed 0∼18 2.0∼5.5 0∼18 0∼10 0∼14 …Traveling speed 0∼2.4 0∼2.4 0∼2.4 0∼3.5 0∼3.0 …Theoreticproductivity

700 500 700 600 600 …

Hopper capacity 14 14 14 13 13 …Gradient ability 20 20 20 20 20 …Engine type BF6M1013E D6114ZG1B BF6M1013 BF4M1013C BF4M1013C …Rated output 137/2300 140/2300 133 112 112 …Generator power 28 28 28 20 20 …Weight 21.3∼28.8 22∼25 21.3∼28.6 16.5∼19.5 16.5∼23.5 …Overall dimension 6,731×3,000×

3,8506,636×3,000×3,880

6,757×3,000×3,650

6,130×3,000×3,815

6,230×3,000×3,855

…

Table 1 Performance parameters of multi-function bitumen-concrete paving machine put forward by customer

Multi-function bitumen–concrete paving machine

Basic workingwidth (ω1)

Max workingwidth (ω2)

Max pavingthickness (ω3)

Paving speed(ω4)

Travelingspeed (ω5)

Theoreticproductivity (w6)

Hoppercapacity (ω7)

3.0 7.5∼9.5 380 0∼15 0∼3.0 600∼650 18∼22Gradient ability (ω8) Engine type (ω9) Rated output (ω10) Generator power (ω11) Weight (ω12) Overall dimension (w13)20 – 110∼120 20∼25 16∼25 Not exceed 6,300×3,000×3,900

(ω1=ω2=ω3=ω4=ω5=ω7=ω7=ω8=ω10=ω11=ω12=ω13=0.076, ω9=0.088,P13i¼1

wi ¼ 1)

Int J Adv Manuf Technol (2009) 42:381–397 391

software services in the field of virtual manufacturing andhave applied it for dynamically monitoring and controllingremote facilities in virtual manufacturing.

The activities and functions involved in product design andmanufacturing on the above systems or methods are differentfrom those for configuration design. Functions related toproduct design may include generation and editing of newshapes, measurement of dimensions, specification of manu-facturing parameters such as tolerances, surface finish, etc.Functions for configuration design of customized productsinclude quick selection of proper base products and compo-nents from part database, configuration of a virtual productsimilar to the customized product, 3D visualization of thevirtual product, and interactive improvement of the virtualproduct. For interactive configuration design, Sun and Chiang[51] presented a Web-based collaborative VR environmentfor customer–company interaction support. But it is mainlydesigned to acquire customer’s requirements through thecommunication between the company’s salesperson andcustomer. It only provides a simple function of the 3Ddisplay of the virtual product. Both customers and companycan not mark, modify, and partially reconfigure the 3Dmodel of the virtual product.

In this paper, in order to allow users to participate inproduct configuration vividly and make the configurationproduct meet the customer’s individual requirements aspossible as it can, an interactive virtual product configura-tion platform is presented. By analyzing the content of

product configuration, successful configuration productsare based upon valid organization of product structure,sufficient product data, and 2D/3D models of the virtualproduct. Thus, the platform consists of five major modules:product configuration hierarchy organization module, vir-tual configuration scene, color module, PDM module,intercommunication module, and collaborative configura-tion design module. Configuration product hierarchy tree isstructured by product configuration hierarchy organizationmodule. Each node of the hierarchy tree corresponds to aconfigurable product function module. If a customer isdiscontented with some function modules of the configuredproduct, he/she can start up the product configurationhierarchy organization module and re-organize the hierar-chy of the configured product using a new replaceableproduct function module. Product data are managed byPDM module and are well-integrated with the producthierarchy tree by the keywords of tree nodes. Participants inproduct configuration, including the customer, designers,and experts, air their opinions by intercommunicationmodule and change colors of 3D model by color moduleto meet individual favor. When customers expect partiallymodifying or improving some product function modules,customers mark it out and inform product designers.Product designers and even customers carry collaborativedesign through collaborative configuration design module.

For the structure of interactive virtual product configu-ration platform, see Fig. 6. It adopts three layers of B/S

Fig. 6 Structure of interactivevirtual product configurationplatform

392 Int J Adv Manuf Technol (2009) 42:381–397

structure and developed on modular thought. The userlayer is an interactive interface connecting users andinteractive configuration design system. Users can browsethe information of parts, product, and product structuretree through Web browser and 3D browser. Users can alsocarry out actions, such as loading virtual product models,disassembling products, and assembling parts, so on. Theoperational logic layer takes charge of the followingoperations: (1) handling information and data uploadedby users; (2) querying, browsing, and modifying config-urable parts of a product; (3) controlling assembly processof 3D models; configuration optimizing and makingdecision. The data layer stores and manages system data,users’ information, product data, 2D/3D product models,and configuration rules.

The interactive configuration process between customersand designers shows in Fig. 7. After the authentication ofusers, users can query and browse all the virtual productinformation provided by the ASP service platform throughWeb browser. If some modifications are needed, they caninform remote designers through Publish Questions &Suggestions module. Users and designers cooperativelycarry out interactive configuration design using offeredinteractive configuration tools, such as Create ProductStructure Tree, Modify Product Structure Tree, ModifyPerformance Parameter, Mark 2D Drawing, etc.

Design communication can be implemented using FTP,VRML, E-mail, Net-meeting, or video conferencing. The3D display is controlled and responds to an outside eventthrough the self-defined feeling nodes and script nodes on

the VRML scene, comprising adding scene nodes, chang-ing the appearance, color, location, and motion track ofnodes on the VRML scene. Real-time interactive operationbetween users and VRML scene is realized by not onlyutilizing Script nodes of VRML itself, but Java Applet andEAI. To achieve product resource reuse, virtual productsmodeled on CAD tools or other 3D modeling software canbe exported into VRML files for display on the VRMLscene. For direct exported VRML files from CAD tools andother 3D modeling software are commonly large anddifficult for transport and coloration, a method of optimi-zation and reconstruction of VRML files [52] is given andresolves the above problems.

5 A case study

Based on aforementioned studies, a prototype of ASP-basedproduct customization service system is developed byutilizing object-oriented technology, module design methodand Internet technology. Resource support layer is builtupon the SQL Server 2000 system. All modules areimplemented in .NET framework by making use oftechnologies of ASP.NET with Visual Basic and JavaScript. The interactive virtual product configuration plat-form is developed by improving a self-developed web-based interactive virtual assembly system [53]. The virtualconfiguration scene is developed by use of EAI technolo-gies of Cortona [54] on Java development platform. Thevirtual configuration environment is developed based onVRML and embedded in an IE6.0 browser that allowsclients to access the environment over the Internet. Thedeveloped product customization system has been diffu-sively applying in construction machinery. Road-surfacepavers in construction machinery are adopted to verify anddemonstrate above studies. Fig. 8 shows several typicalinterfaces developed based on above studies.

Fig. 8(a) is a typical interface of resource evaluationsystem for collecting product configuration resource of aroad-surface paver. It is developed on the proposeddynamic resource collecting method with certain rewardmeasures. The evaluation object is the manufacture re-source for a Frame of 1220MAXI PAV Paver. Experts login to the system to assign evaluation indices and mark therelevant weight of evaluation indices. Here, the indicesaggregate U specialization degree, innovation degree, creditdegree, quality level, forecast-value respectively. Eachindex is divided into six grades, i.e. very high, high,general, low, very low, and none. The grade of usefuldegree is set as, g1=0.8, g2=0.5, g1=0.1 and the lower limitto be value-added resource is δ=0.75. Through statisticallycomputing and analyzing, the resource value is Val=68 andthe relative useful degree is Vper=0.779. Evaluated resultsFig. 7 Interactive configuration design process

Int J Adv Manuf Technol (2009) 42:381–397 393

show that the submitted manufacturing resource is a value-added resource.

Fig. 8(b)~(d) are typical interfaces of product custom-ization. Fig. 8(b) provides customers rapid configurationfunctions of primary case search by keywords. Retrievedassembly is organized as a product structure tree andmapped into XML files for storage. Customers canreconstruct the product structure tree by adding nodes,deleting nodes, replacing a module with a configurablemodule or parts on a node. Fig. 8(c) is the typical interfaceof interactive virtual product configuration platform whichoffers all users functions of product models preview,collaborative configuration analysis, and collaborativeconfiguration design. The platform includes seven mainmodules shown at the left of Fig. 8(c). The ConfigurationStructure Tree Management module takes charge oforganizing and maintaining configuration product hierarchystructure tree. The Product Data Management module, theProduct Version Management module, and the PartsManagement module are responsible for managing andmaintaining detailed information of products or parts linkedon tree nodes. The 3D Model Control module bears twofunctions: one is to control 3D model assembly/disassem-bly and movement, the other is to mark the 3D model whenany modification is demanded. The Match Colors Manage-ment module provides customers’ auto-match the colorsfunction of the 3D model according to customers’ favorites.Customers, experts, designers, and ASPs have different

authorizations respectively in operation, management, andmaintenance. Their authorizations are managed by UsersAuthorization Management module. The interface ofinteractive virtual product configuration environment isdivided into four display plats. The center consists ofabove-part and under-part. The above-part is a virtual sceneof the 3D model display and the under-part lists theperformance parameters of configuration product. The rightis divided into the display area of product structure tree andcommunication area. Detailed information of the config-ured product, including product hierarchy structure tree,performance parameters table, dimension parameter table,and 3D model can be queried and provided to furtherdesign or manufacture (Fig. 8(d)). Product hierarchystructure trees in Fig. 8(b), (c), and (d) are created byXML template mapping. Parameter tables in Fig. 8(c) and(d) are retrieved by the combination of text similar degreealgorithm and parameter gray correlative-degree algorithm,and created by object-oriented XML table mapping. ASPsare responsible for system maintenance and management.

When a client inputs keywords of required product, partsof counterparts or similarity are organized as a hierarchytree by using text similar degree algorithm and listed at theleft of Fig. 8(b). When multi-counterparts or similarity areretrieved, the client switches to parameter configuration byusing combined algorithm of text similar degree andparameter gray correlative degree. Initial configurationproduct is organized as hierarchy tree (at the center of

Resource evaluation

Product customization

Interactive virtual product configuration environment

Configuration results display

a c

db

Fig. 8 Typical interfaces ofconstruction-machinery-orientedproduct customization serviceplatform (in Chinese)

394 Int J Adv Manuf Technol (2009) 42:381–397

Fig. 8(b)) and deposits on product knowledge resource basein XML format as a new case for future use (at the left ofFig. 8(b)). While further modification for meetingindividual requirements or configuration validity analysisis required, clients (customers, designers, experts andmanufacturers) can enter into interactive virtual productconfiguration environment (Fig. 8(c)) and carry collabo-rative analysis, design, and modification (color, shape, andappearance in 3D). Detailed information of configurationproduct, including product hierarchy organization, per-formance parameters table, dimension parameter tableand 3D model, can be queried and provided to furtherdesign or manufacture (Fig. 8(d)). Any further modifica-tion of the configuration product can be carried on throughswitching to interactive virtual product configurationenvironment.

For example, a customer wants to customize a multi-function cement–concrete paver produced. The customerlogins onto the product customization system through theportal of ASP service platform and first inputs keywords ofpaver, machinery, and cement concrete for primary search.Fifteen counterparts and some selections of relevant partslist at the left of Fig. 8(b). The customer is asked tocontinue inputting detailed parameters; then the customerinputs parameters of basic paving width 4~4.5 m, maxpaving width 8 m, paving thickness 360 mm, paving speed0~6 m/min, and so on. When the customer taps the buttonof searching, product text similar degree algorithm andparameter gray correlative-degree algorithm model start up.Product listing is immediately updated and only thehierarchy structure of 1220MAXI-PAV remains. Accordingto different paving width, each kind of frames of1220MAXI-PAV has several selections. For instance,Right-front End Frame has three selections, i.e. Right-frontEnd Frame1, Right-front End Frame2, and Right-front EndFrame3. The customer can select favorite or more suitableparts to form assembly through function modules of AddNodes, Delete Selected Nodes and reconstruct structureTree (see at the center of Fig. 8(b)). After primary productconfiguration, the customer switches to interactive virtualproduct configuration environment to preview and carry outindividualized design. If the customer expects Cabinet Aand C to be coffee color, Cabinet B with chamfer angle of30°, relevant location of 3D model in virtual scene can belabeled and modification information can be published oncommunication area. The customer himself can modify thecolor of Cabinet A and Main Frame C through the functionmodule of Match Colors. Designers and experts executecollaborative design and validity analysis, and feed backresults to the customer. If the customer put forward furtherindividuation needs, the same process can continue till theconfiguration product meets with the customer’s require-ments completely.

The construction-machinery-oriented product customiza-tion service platform has been applied to some SMEs inconstruction machinery in Jiangsu Province, People’sRepublic of China. Feedback information shows that afterusing this service platform, these SMEs are able to conducttheir Internet-based sales and product customization effec-tively in a shorter time and at a lower cost than before.

6 Conclusion and future work

In this paper, we have proposed an ASP-based productcustomization service system operating in lifecycle-orientedcustomization service mode. With the ASP mode, theproduct customization service system offers significantcommercial and organizational benefits such as easysoftware maintenance and future upgrade, low cost of entryinto the business for SMEs. Key technologies for support-ing lifecycle-oriented product customization service effec-tively are deeply studied. A general resource evaluationmodel with reward measures for facilitating distributedresource sharing is established. XML template-drivenmapping and object-oriented XML table mapping forproduct hierarchy structure tree reorganization are pre-sented. Considering the uncertainty and diversification ofcustomers’ demands, a new case-based retrieving algorithmof parameter gray correlative-degree algorithm is estab-lished to speed product configuration. In order to meetcustomers’ individual requirements and allow customers toparticipate in product configuration vividly, an interactivevirtual product configuration platform is designed. Aconstruction-machinery-oriented product customization ser-vice system is developed. Road-surface pavers in construc-tion machinery are taken as a case and the above studies areverified.

In the future, we intend to constitute a set of uniformevaluation criterion through statistical analysis of history datato make resource evaluation more effective. In addition,product configuration validity analysis, system service mech-anism, and data safety need further investigation.

Open Access This article is distributed under the terms of theCreative Commons Attribution Noncommercial License which per-mits any noncommercial use, distribution, and reproduction in anymedium, provided the original author(s) and source are credited.

References

1. Jiao JX, Ma QH, Tseng MM (2003) Towards high value-addedproducts and services: mass customization and beyond. TCH23:809–821

Int J Adv Manuf Technol (2009) 42:381–397 395

2. Duray R, Ward TP, Milligan WG, Berry LW (2000) Approachesto mass customization: configurations and empirical validation.JOM 18(6):605–625

3. Kotha S (1995) Mass customization: implementing the emergingparadigm for competitive advantage. Strateg Manage J 16:21–42

4. Wang ZQ, Li XN, Jiang CY (2003) Networked manufacturing forhigh technology industry region of Shanxi Province’s center.CIMS 9(8):710–715

5. Zhao HJ, Ju WJ, Wang SY, Yin CF (2003) Software resourcesharing and its application in networked manufacturing system.CIMS 9(7):608–612

6. Choi SY, Whinston BA (1999) The future of e-commerce:integrate and customize. Comput 32(1):133–1 34 138

7. Gehling RG (2002) Development of a general application serviceprovider (ASP) model: a case study. PhD thesis, AuburnUniversity, Alabama, USA

8. Smith MA, Kumar LR (2004) A theory of application serviceprovider (ASP) use from a client perspective. Inf Manage 41:977–1002

9. Kern T, Lacity MC, Willcocks LP (2002) Netsourcing: rentingbusiness applications and services over a network. Prentice-Hall,New York

10. Lockett NJ, Brown DH (2005) A SME perspective of verticalapplication service providers. Int J Enterprise Inf Syst 1(2):37–55

11. Yang Y, Zhang XD, Liu F, Xie Q (2005) An Internet-basedproduct customization system for CIM. Robot Comput IntegrManuf 21:109–118

12. Xu LY, Li AP, Zhang WM (2004) Networked manufacturing baseon ASP and relational technologies. CMEM 15(19):1755–1759

13. Rodney LS, Gerard AA, Ashwin WJ (2002) Managing seller–buyer new product development relationships for customizedproducts: a contingency model based on transaction cost analysisand empirical test. JPI 19(6):439–454

14. Heart T, Pliskin N (2001) Is e-commerce of IT application services(ASP) alive and well? JITTA 3(4):33–41

15. Zhou H, Zhang Y, Li D (2002) A remote injection mouldingsimulation system based on the Internet and the Web. Int J AdvManuf Technol 19:722–726

16. Huang CY, Cheng K, Holt A (2007) An integrated manufacturingnetwork management framework by using mobile agent. Int J AdvManuf Technol 32:822–833

17. Pine BJ, Victor B, Boyton AC (1993) Making mass customizationwork. Harv Bus Rev 71:108–119

18. Su Y, Liao WH, Guo Y, Gao SW (2007) Method on customizationcustomers dynamic clustering oriented to design for mass custom-ization. CMEM 18(22):2711–2716

19. Su Y, Liao WH, Guo Y, Shi HB (2007) Customer grouppreference’s configuration features merging for configurableproduct design. CMEM 18(20):2439–2443 2448

20. Choi B, Lee H (2003) An empirical investigation of KM stylesand their effect on corporate performance. Inf Manage 40:403–417

21. E-works (2004) A seminar on application and diffusibility of ASPin China. http://www.e-works.net.cn/ewk2004/dtl/asp/index.htm.Accessed 29 June 2005

22. Wang SY, Zhao HJ, Sun LF (2005) Research on resourceintegration framework of regional networked manufacturing ASPservice platform. CMEM 16(9):1729–1732

23. Zha XF (2007) A web-enabled open database system for designand manufacturing of micro-electro-mechanical systems (MEMS).Int J Adv Manuf Technol 32:378–392

24. Reinoso-Castillo JA (2002) Ontology-driven information extrac-tion and integration from heterogeneous distributed autonomousdata sources: A federated query centric approach. Masters Thesis.

Artificial Intelligence Research Laboratory, Department of Com-puter Science, Iowa State University

25. Zhou GH, Jiang PY (2005) Using mobile agents to encapsulatemanufacturing resources over the Internet. Int J Adv ManufTechnol 25:189–197

26. Mouritsen J, Larsen HT (2005) The 2nd wave of knowledgemanagement: the management control of knowledge resourcesthrough intellectual capital information. Manage Account Res16:371–394

27. He Z, Qiu CH, Wang NJ, Yao MZ (2006) Evaluation andselection of the ship collaborative design resources based onAHP and genetic and simulated annealing algorithm. JMSA 5(1):23–30

28. Tao F, Hu YF, Ding YF, Sheng BY, Zhou ZD (2005) Resourceoptimization selection evaluation model based on agents inmanufacturing grid systems. CMEM 16(24):2192–2197

29. Xu W, Liu J, Shen XX (2008) Research on the status quo andsystem architecture of the Web information resource evaluation.In: IFIP International Federation for Information Processing,Springer, Boston, USA

30. Wang H, Zhou XH, Ruan XY (2003) Research on injection mouldintelligent cost estimation system and key technologies. Int J AdvManuf Technol 3:215–222

31. Jiao JX, Simpon TW (2007) Product family design and platform-based product development: a state-of-the-art review. J IntellManuf 18:5–29

32. Chen G, Chen J, Zhao Z, Ruan XY (2005) An object-orientedhierarchical case representation of automotive panels in computer-aided process planning system. Int J Adv Manuf Technol26:1323–1330

33. O’reilly (2004) XML.com webpage. www.xml.com. Cited Febru-ary 2005

34. W3C (2004) W3C webpage. www.w3.org. Cited March 200535. Hegge HMH, Wortmann JC (1991) Generic bill-of-material: a

new product model. Int J Prod Econ 23(1–3):117–12836. Kobler A, Norrie MC (1997) A product information system based

on an object-oriented internet database system. Proceedings of the6th IEEE Workshops on Enabling Technologies Infrastructure forCollaborative Enterprises (WET-ICE ‘97). IEEE Computer Soci-ety, Washington, DC, USA, pp 43–48

37. Ryu YU (1999) A hierarchical constraint satisfaction approach toproduct selection for electronic shopping support. IEEE TransSyst Man Cybern A Syst Hum 29(6):525–532

38. Kolodner JL (1993) Case-based reasoning. Morgan Kaufman,Palo Alto, USA

39. Tseng HE, Chang CC, Shu HC (2005) Applying case-basedreasoning for product configuration in mass customizationenvironments. Expert Syst Appl 29:913–925

40. Wu MC, Lo YF, Hsu SH (2006) A case-based reasoning approachto generating new product ideas. Int J Adv Manuf Technol30:166–173

41. Deng JL (1990) Gray system theory. Huazhong University ofScience and Technology Press, Wuhan, Hubei, China

42. Xiao XP, Song ZM, Li F (2005) The basis and application of graytechnology. Science, Beijing, China

43. Hu J (2006) Research and implementation of key technologiesin product design knowledge management. PhD Thesis,Nanjing University of Aeronautics and Astronautics, Nanjing,China

44. Watson I (1996) Case-Based Reasoning Tools: An Overview. In:Proceedings of the 2nd UK Workshop on Case Based Reasoning,71–88. Salford University, UK

45. Kan HY, Duffy VG, Su CJ (2001) An internet virtual realitycollaborative environment for effective product design. CompIndus 45(2):197–213

396 Int J Adv Manuf Technol (2009) 42:381–397

46. Cera CD, Regli WC, Braude I, Shapirstein Y, Foster CV (2002) Acollaborative 3D environment for authoring design semantics.IEEE CG A 22(3):43–55

47. Huang GQ (2002) Web-based support for collaborative productdesign review. Comp Indus 48(1):71–88

48. Lau HYK, Mak KL, Lu MTH (2003) A virtual design platform forinteractive product design and visualization. J Mater ProcessTechnol 139:420–407

49. Zhang S, Shen W, Ghennewa H (2004) A review of Internet-basedproduct information sharing and visualization. Comp Indus 54(1):1–15

50. Gierach K, Thompson D, Banerjee P (2002) An approach forfacilitating service management in networked virtual manu-

facturing environments. Robot Comput Integr Manuf 18:147–156

51. Sun TL, Chiang TH (2005) Supporting customer–companyinteractive in product customization using a web-based collabo-rative VR environment. J Chin Inst Ind Engineers 22(3):235–242

52. Su Y, Liao WH, Guo Y (2007) Virtual assembly-orientedoptimization and reconstruction of VRML models. Comput EngDes 28(11):2509–2512

53. Yi L, Cheng XS, Guo Y (2006) Development and implementationof web-based interactive virtual assembly platform. Mach Build-ing Autom 35(1):75–79

54. ParallelGraphics (2005) ParallelGraphics webpage. www.parallelgraphics.com. Cited December 2005

Int J Adv Manuf Technol (2009) 42:381–397 397