E-commerce And Digital Manufacturing 9781845446635, 9780861767274

Citation preview

ISSN 0957-6061

Integrated Manufacturing Systems The International Journal of Manufacturing Technology Management Volume 13, Number 5, 2002

E-commerce and digital manufacturing Guest Editors: George Huang, K.L. Mak and P.Y. Jiang

This issue is part of a comprehensive multiple access information service comprising:

Paper format Integrated Manufacturing Systems includes eight issues in traditional paper format. The contents of this issue are detailed below.

Internet Online Publishing with Archive, Active Reference Linking, Emerald WIRE, Key Readings, Institution-wide Licence, E-mail Alerting Service and Usage Statistics. Access via the Emerald Web site: http://www.emeraldinsight.com/ft See overleaf for full details of subscriber entitlements.

Contents

278 Access to Integrated Manufacturing Systems online

340 The Internet-based virtual machining system using CORBA

279 Abstracts & keywords

Sang-Hoon Kong, Jaehong Park, Young-Geun Han, Gibom Kim and Kyo-Il Lee

281 Guest editorial 283 E-commerce in production: some experiences Frank Dignum

295 Applying digital manufacturing technology to ship production and the maritime environment Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang

306 ASP-driven e-service platform for Web-based online manufacturing Pingyu Jiang, Guanghui Zhou and Yong Liu

319 A survey and implementation framework for industrial-oriented Web-based applications Qingjin Peng

328 Design agility and manufacturing responsiveness on the Web J. Toussaint and K. Cheng

345 Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Jason S.K. Lau, George Q. Huang and K.L. Mak

Integrated Manufacturing Systems online An advanced knowledge resource for the entire organization Access via the Emerald Web site – http://www.emeraldinsight.com/ft

Subscribers to this journal benefit from access to a fully searchable knowledge resource stretching far beyond the current volume and issue. Integrated Manufacturing Systems online is enhanced with a wealth of features to meet the need for fast, effortless, and instant access to the core body of knowledge. Furthermore, this user-friendly electronic library may be networked throughout the subscribing organization to maximize the use and value of the subscription. This is augmented with advanced search facilities and ‘‘choice of access’’ through a variety of alternative journal gateways.

Emerald online access includes: Institution-wide Licence Our liberal licence allows everyone within your institution to access the journals via the Internet, making your subscription more costeffective. It has been designed to provide a comprehensive, simple system with minimum administration. Access can be granted by IP range or username and password. ATHENS authentication is enabled.

Current Issue + Journal Archive Internet access via Emerald Fulltext to information as it becomes available online and to material from past volumes. Users can choose to browse the database for relevant articles or search using the powerful search engine provided.

Active Reference Linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article.

Emerald WIRE A fully searchable subject specific database, brought to you by Emerald Management Reviews, providing article abstracts from the world’s top management, computing and engineering journals. Directly relating to the journal subject area, the database is updated monthly and gives users details of how to obtain the full text of original articles.

Usage Statistics Online Journal Usage Statistics are now available. This feature allows Emerald Administrators to download their usage statistics with regard to their organization’s journal usage. Usage Statistics allow you to review the value of electronic dissemination of your journal subscriptions throughout your organization. They can also help determine the future trends for information within your organization. For further information go to http://www.emeraldinsight.com/stats

Key Readings Abstracts of articles, relating to keywords, are selected to provide readers with current awareness of interesting articles from other publications in the field. The abstracts are

[ 278 ]

available online and are updated regularly upon publication of the latest issue.

Emerald Alert The table of contents e-mail alerting service will e-mail you the contents page of any chosen journal whenever the latest issue becomes available online. For further information please go to http://www.emeraldinsight.com/ alerts

Support Resources A comprehensive range of resources is available online that help users learn how to effectively use online information resources and that helps information professionals market resources to their users. For further information please go to http://www. emeraldinsight.com/support

Choice of Access Electronic access to this journal is available via a number of channels, enabling end users’ libraries to reach the content through their preferred delivery system. The Emerald Web site – http://www. emeraldinsight.com/ft – is the recommended means of electronic access as it provides fully searchable and high value-added access to the complete content of the journal. Refer to the next section for ‘‘how to access’’ via the Emerald Web site. Subscribers can also access and search the article content of this journal through the following journal gateway services: Ebsco Online http://www.ebsco.com/online Huber E-Journals http://e-journals.hanshuber.com/english Minerva http://www.minerva.at/ OCLC Firstsearch ‘‘Electronic Collections Online’’ http://www.uk.oclc.org/oclc/menu/ eco.htm RoweCom’s ‘‘Information Quest’’ http://www.informationquest.com SilverPlatter http://www.silverplatter.com SwetsBlackwell’s ‘‘SwetsnetNavigator’’ http://www.swetsnetnavigator.nl

How to access this journal through Emerald Organizations must first register for online access (instructions provided at http:// www.emeraldinsight.com/register), after which the content is available to everyone within the organization’s domain. To access this journal’s content, simply log on either from the journal homepage or direct through the Emerald Web site. Emerald Customer Support Services For customer service and technical help, contact: E-mail: [email protected] Telephone: (44) 1274 785278 Fax: (44) 1274 785204

Abstracts & keywords

E-commerce in production: some experiences

ASP-driven e-service platform for Web-based online manufacturing

Frank Dignum

Pingyu Jiang, Guanghui Zhou and Yong Liu

Keywords Internet, Catalogues, Shopping, Relationship marketing, Logistics

Keywords Internet, Computer-integrated manufacturing, Applications software

This paper will report on a number of e-commerce projects at production companies in The Netherlands. It has been shown that there is no unique model for the introduction of e-commerce in production companies. Companies can start to use e-commerce at different ends of the organization, with different objectives and different (positive) outcomes, depending on the market situation and the state of ICT awareness within the company. Objectives might range from cutting costs on purchase procedures to increasing customer intimacy. ICT awareness can range from having one PC (personal computer) for a department with no Internet connection to a completely functioning intranet and Internet connection for each employee. The paper will be concluded with some current trends in e-commerce, especially on the increased use of electronic markets between the companies within the same vertical market segment.

Applying digital manufacturing technology to ship production and the maritime environment Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Keywords Computer-integrated manufacturing, Shipbuilding, Process design, Simulation This paper is mainly concerned with the digital manufacturing technologies in the context of the shipbuilding industry. New concepts such as digital shipbuilding, virtual shipyard, and simulation-based design (SBD) will be explored. After reviewing the digital shipbuilding, a case study will be presented using the virtual assembly simulation system for shipbuilding (VASSS), a simulation based tool, to evaluate block erection sequence taking account of shipyard facilities, operational efficiency and equipment replacement time.

Integrated Manufacturing Systems 13/5 [2002] Abstracts & keywords # MCB UP Limited [ISSN 0957-6061]

Extends the concept of e-service to the whole phase of manufacturing. Develops an e-service platform prototype based on this concept with Java Web solution including the mobile agent broking technologies and application service provider (ASP) principle. The key point to implement the platform is to enable an open Web information service infrastructure for the whole product manufacturing chain. Inside this infrastructure, product-specific online manufacturing system can be created by means of using a kind of bidding model. All users participating in the manufacturing process are able to cooperatively finish manufacturing tasks in real time through sharing the same platform. With the help of BOM flow, the global information service flow can be controlled easily. In addition, the legacy hardware/software can also be encapsulated with aglets that are Java mobile agents. As to new ASP software packages, they can be configured simply via the plug and play mode to the e-service platform. In this way, the on-line networked manufacturing can be tested.

A survey and implementation framework for industrial-oriented Web-based applications Qingjin Peng Keywords Design, Manufacturing, Product design, Internet, Surveys Internet technology is changing the way of product development, ranging from information gathering, product managing and commerce to product development, and maintenance. In order to obtain the evidence of the extent of Web-based applications in industry, we did a survey to examine the impact and need of the Internet in product development under industrial environments. This paper provides a view of Internet-based applications in Canadian industry based on the data received from 42 Canadian small and medium sized enterprises (SMEs), which results in a solution to improve current Web-based industrial applications. A framework of an industrial-oriented Web-centred system is proposed based on the demand found from the survey. Some examples are provided to demonstrate applications of the proposed framework.

[ 279 ]

Abstracts & keywords Integrated Manufacturing Systems 13/5 [2002] 279–280

Design agility and manufacturing responsiveness on the Web J. Toussaint and K. Cheng Keywords Internet, Design, Computer-integrated manufacturing This paper is to investigate a Web-based engineering approach to enable engineers to use, share, and simulate effectively and efficiently design and manufacturing data through the World Wide Web. The enabling techniques to implement such an approach are critically discussed with an example based on a tolerancing application and two Java-based machining simulators for e-manufacturing purposes, as well as an example of real-time 3D implementation. The outcome of this research is for helping designers and manufacturing engineers to be able accurately to make the right decisions for their respective purposes.

The Internet-based virtual machining system using CORBA Sang-Hoon Kong, Jaehong Park, Young-Geun Han, Gibom Kim and Kyo-Il Lee Keywords Internet, Manufacturing, Machining, Computer-integrated manufacturing This paper presents an Internet-based virtual machining system which applies virtual manufacturing technology to machining processes of CNC machining center. The system is implemented to execute digital machining and verification, to transmit the NC code data to related machining centers after confirming the properness of virtual machining, and to manipulate the machine through the Internet. Also, this research proposes a basic structure that can monitor the status of machines via the Internet. By applying simulation techniques for machining processes, a simple manipulation and monitoring system of machining centers is realized. This entire system is constructed by adopting the latest information technology such as object-oriented method, middleware, Internet programming, and client-server structure.

[ 280 ]

Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Jason S.K. Lau, George Q. Huang and K.L. Mak Keywords Supply chain, Performance, Information, Inventory control Information sharing and coordination between buyer and vendor have been considered as useful strategies to improve supply chain performance. The debate is about what information to share and how to share most cost-effectively to maximize the mutual benefits of the supply chain as a whole and the individual business players. Proposes a systematic framework for investigating the impacts of sharing production information on the supply chain dynamic performance. This framework supports supply chain researches to study impacts of information sharing under various scenarios. Examines, under the framework, an inventory allocation problem in an arborescent distribution supply chain with two distribution channels competing for the same source of supply. Finds that the levels of benefits by sharing information vary with different players involved in the supply chain. Suggests some guidelines to balance the benefits in a supply chain in order to motivate information sharing.

Guest editorial

About the Guest Editors Dr George Huang is Associate Professor in the Department of Industrial and Manufacturing Systems Engineering, University of Hong Kong. His main research areas include collaborative product commerce, digital manufacturing. He has published extensively in these topics. Dr Huang is a Chartered Engineer, and a member of IEE, ASME, IIE, and HKIE. Professor K.L. Mak is Professor and Head of Department of Industrial and Manufacturing Systems Engineering, University of Hong Kong. His current research interests focus mainly on production and operations management, product development, and manufacturing systems design and control, and he has published extensively in these areas. Professor Mak is a Chartered Engineer, a member of IMechE and IEE, and fellow of HKIE. Professor P.Y. Jiang is Professor and Director of the CAD/CAM Institute at Xian Jiao Tong University, China. He held Humboldt and JSPS Post Doctoral Research Fellowships from 1995 to 1999 in Germany and Japan respectively. His main research interests include global manufacturing, information engineering, virtual manufacturing, CAD/CAPP/CAM. Professor Jiang is a member of ASME.

Integrated Manufacturing Systems 13/5 [2002] 281–282 # MCB UP Limited [ISSN 0957-6061]

In recent years, the importance of Internet and World Wide Web (WWW) technologies in manufacturing industries has been rising very rapidly in a global context. According to a recent study, Manufacturing Foresight 2020, conducted by UK Department of Trade and Industry (DTI), the impact of this development is deemed the most profound since the Industrial Revolution. The waving interests in the electronic commerce and electronic business (e-commerce/e-business) have spread, from the heartland (product development) to the battlefield (shop floor), of manufacturing enterprises. The number of Web applications is ever on the rise, and many practitioners are keen to try these remote systems through Web browsers to support their decision-making activities. Indeed, product design and manufacture professionals will soon be able to benefit from such remote services and support commercially available on the Internet. The practice and performance of product development and realization are expected to make immense progress. Web applications in product design and manufacture signal the beginning of a new era of the digital manufacturing enterprise. However, many loopholes are found in the development and application processes because of domain complexity and technology sophistication, thus generating new challenges to both the developers and practitioners. A simple example is the difference in the user interfaces between Web applications and traditional applications. Indeed, abundant issues need to be resolved before the full launch of digital manufacturing can come into being. This special issue has solicited seven papers from various research groups where research and practice of digital enterprises have been conducted for several years. The first paper by Dignum reports on a number of e-commerce projects at production companies in The Netherlands. The research is concerned with what is the appropriate model for introducing e-commerce in production companies, what objectives these production companies intend to achieve, and what the current trends are. The second paper is contributed by a group of Korean researchers. The research is mainly concerned with the digital

manufacturing technologies in the context of the shipbuilding industry. New concepts such as digital shipbuilding, virtual shipyard, and simulation-based design (SBD) are explored. A case study is presented using a virtual assembly simulation system for shipbuilding (VASSS) to evaluate block erection sequence taking account of shipyard facilities, operational efficiency and equipment replacement time. Provision of manufacturing services and applications over the Web are essential to form what is called a manufacturing portal. The third paper comes from a group of Chinese researchers led by Professor Jiang. They are working on the development of a manufacturing portal server platform and e-manufacturing Web applications. E-business solutions are generally beneficial to large corporations. But it is unclear how small and medium sized enterprises can benefit from the development. The fourth paper is based on a survey carried out within Canadian SMEs, aiming to identify the elements of e-business solutions that are most suitable for SMEs. The development of Web applications in product design and manufacturing has been a great challenge to all of us. Professor Cheng has been working on and supervising a number of research projects for several years. The fifth paper reports on the experience acquired from the development and the implementation of a Web-based system to support decision-making activities in bearing design and manufacturing. The issues under discussion include design information integration, remote execution, Java programming, client-server architecture, and user interface design. There is no doubt that business players involved in a supply chain or extended enterprise would benefit significantly if they shared production information and activities. However, there has been no answer to the very simple question of what production information should be shared and how they should be shared. The last paper is concerned with how e-business solutions can be used to improve the information sharing and thus its impacts on the performance.

[ 281 ]

Guest editorial Integrated Manufacturing Systems 13/5 [2002] 281–282

[ 282 ]

The sixth paper presents an Internet-based virtual machining system which applies the virtual manufacturing technology to a CNC machining centre. The system is implemented to execute digital machining and verification, to transmit the NC code data to machining centres, and to manipulate the machine through the Internet. The guest editorial team would like to thank all the authors for the time and effort in contributing their papers and in incorporating the referees’ comments in revising their manuscripts. Thanks are

especially extended to the referees in giving their valuable comments to the papers, which are most essential for this special issue to come into being. Finally, the guest editors would like to express their heartfelt thanks to Professor David Bennett, the editor-in-chief, for his advice, help, and support, to make this special issue project a success. George Huang K.L. Mak P.Y. Jiang

E-commerce in production: some experiences

Frank Dignum Institute of Information and Computing Sciences, Utrecht University, Utrecht, The Netherlands

Keywords Internet, Catalogues, Shopping, Relationship marketing, Logistics

Abstract This paper will report on a number of e-commerce projects at production companies in The Netherlands. It has been shown that there is no unique model for the introduction of e-commerce in production companies. Companies can start to use e-commerce at different ends of the organization, with different objectives and different (positive) outcomes, depending on the market situation and the state of ICT awareness within the company. Objectives might range from cutting costs on purchase procedures to increasing customer intimacy. ICT awareness can range from having one PC (personal computer) for a department with no Internet connection to a completely functioning intranet and Internet connection for each employee. The paper will be concluded with some current trends in e-commerce, especially on the increased use of electronic markets between the companies within the same vertical market segment.

Received March 2001 Revised January 2002 Accepted February 2002

Integrated Manufacturing Systems 13/5 [2002] 283–294 # MCB UP Limited [ISSN 0957-6061] [DOI 10.1108/09576060210429739]

Introduction In recent years electronic commerce (e-commerce) has been at the centre of interest of many companies and also of many researchers, as can be seen from Kalakota and Whinston (1996) and Kalakota and Robinson (2001). However, after a period of a growing world economy in which electronic commerce was advocated as one of the motors of the new economy, we are experiencing a set-back at the moment that makes many people doubt the true value of electronic commerce for the economy. In fact what we are experiencing is the maturing of a new way of doing business based on new paradigms. Although information technology (IT) is an important component of electronic commerce, the biggest mistake that many companies have made is that they expected huge benefits from just implementing new IT components that support electronic commerce, without changing their organisational structure. Electronic commerce (or electronic business as some now prefer to call it), however, is primarily a new way of relating to customers and suppliers. In general it means a transformation from a company that is geared towards production excellence or innovative excellence into a company that is geared towards customer intimacy (Treacy and Wiersema, 1997). Although IT plays an important role in this respect (think about the Internet and the WWW) it is not the only aspect of this transformation! A very similar situation is visible in traditional commerce already. There are mail-order companies that sell their products through the mail by means of (paper) catalogs. However, although this opportunity is open for every company, not The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0957-6061.htm

all companies choose to use this selling channel. In order to reap the full benefit of selling by mail the company has to have a certain structure and do business in a way that does not fit with all existing companies. For instance, these companies often give the possibility to pay for products, like clothes, in monthly instalments. This is very unusual for companies that sell their clothes through shops. Also mail-order companies have a huge department that handles returns, which is of small importance to most companies who sell their products through shop outlets. Unfortunately there is no unique, successful business model for companies that perform electronic business. The model depends (just as in traditional business) on the products (or services) of the company, the market structure, etc. In this paper we will discuss a number of aspects of electronic commerce which are of interest to most traditional companies in some form and indicate some issues that should be addressed for each aspect. For instance, electronic sales are possible in many industries. Whether the ordering should be done through an on-line catalog or based on messages depends on the specific situation. The results reported in this paper stem for a large part from the experiences in a project performed by the Eindhoven University of Technology in cooperation with 11 partners in the region of Eindhoven. (Veth, 2000). With each of these partners a specific project on electronic commerce was performed that suited best with their current situation. This reached from customer relations and feedback to designing automatic procurement procedures. Although each of these projects is interesting enough to report on separately, this paper will instead be organized around a number of recurrent topics and use the projects to illustrate the points made in each section.

[ 283 ]

Frank Dignum E-commerce in production: some experiences

Customer and supplier management

Integrated Manufacturing Systems 13/5 [2002] 283–294

One of the biggest advantages for companies to be present on the Internet and the WWW is the fact that the interactions with (potential) customers can easily be monitored and logged. If these data are used properly they provides a wealth of marketing information. Some sites are primarily financed by the companies that may use the information generated by the traffic through the sites for their own purposes. A good example of this category is travelocity (www.travelocity.com). Travelocity is a portal that builds up a community around the concept of travelling. It gives lots of independent information to attract customers. The companies that deliver the content of this site are travel agencies, airlines, hotels, etc. They use the data of all people visiting the site and therefore get far more information about the market than they would operating on their own. As advertising is very expensive and it is difficult to reach exactly the right target group, any means of defining a group of potential customers and their profiles is very valuable. Lechner et al. (2000) report on a number of articles on building on-line communities. Although the above example is very popular on the Internet, it might not seem very applicable in business to business (B2B) commerce. However, the formation of communities with similar interests also happens very often in industry. In traditional commerce this can be observed in branch organisations, usually paid by all members from the industry, that provide independent information about all kinds of new developments to their members. Many of these organisations have played a major role in promoting the use of electronic data interchange (EDI) in the past and are now giving information to their members about the consequences of electronic commerce for their industry. Although communities can be formed within branches it is also possible to form a community with the existing business customers of a company. This is illustrated by the developments at Interbrew Netherlands. This brewery has around 15,000 retail customers that purchase their products (mainly beer). Because beer is not a product that can be sold over the Internet, there is no use trying to sell directly to consumers through the WWW. Retailers also do not change their supplier very easily and therefore the Internet can also not be used to attract new retail customers. However, the Internet can be used to form a community

[ 284 ]

with the customers of Interbrew. They can get more information about products, order new supplies on-line, get on-line information about their orders and Interbrew can eventually even perform statistical analyses for them in order to provide the right supplies at the right time. In this way they are more tightly bound to Interbrew. Another example of getting into a closer relationship with the customers is given by the project performed at DAF trucks. They use the Internet connection in the first instance to get feedback from their customers on the product and the transaction performance. Traditionally, it is difficult to get feedback from users once a transaction has been completed. Usually they have little incentive to give feedback or the costs are perceived as being higher than eventual benefits. The use of the Internet makes it very easy to give feedback on an order. One can organize it as a simple WWW form that can be filled or even just clicked on. The feedback can also be acknowledged in a subsequent order of that customer and if possible it can indicate what has been done with the feedback. For example, if the delivery time was too long, the next order can be shipped more accurately or reasons for delays can be given or even a discount for repeating delays can be given. This kind of reaction relies on the fact that the feedback of a customer can be coupled with the profile of that customer and can be (automatically) processed in the departments of the company that are most affected by the feedback. This implies again that the information systems that are used internally have to be connected and synchronised, e.g. through the use of an ERP (enterprise requirement planning) system throughout the company. The advantage of the above setting is that the company gets a closer relationship with the customers. Therefore customer loyalty will increase and also the company can react more quickly and more accurately to customer demands. The disadvantage of forming on-line communities is that the customers get more power. They get more information from each other about products (both good and bad). It is also an opening up of the company to the outside world. Members of the community expect more and more up-to-date information about products, orders etc. In B2B it may lead to selecting a small number of companies with which an intimate relationship is established. Within such a relationship feedback might be given on the products and the (order) process (like at DAF trucks). This feedback can be very valuable to optimise products for

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

important clients and to optimize production planning in relation to customer orders. A different aspect arises for companies that manufacture products that are designed on demand. When products are designed for a specific order, the client usually gives some functional specifications, but does not have the knowledge to provide a complete design. For example, steel plates for vehicles. Clients know what they want to use it for, what some of the constraints are on the use, but do not know what material is best to use or which shape is optimal. This knowledge resides with the manufacturer and can be made into a product by itself. If the knowledge is made available to customers in a way that they can decide themselves what is the optimum solution for their problem, the manufacturer saves both the time of an engineer that had to determine this solution with the client, as well as being able to ask money for disclosing this knowledge. In general, the developments of IT can lead to two different types of relationship between organizations. Organizations can form hierarchical relations with a few big customers and/or suppliers. These relations are very tight and ensure high reactivity and usually efficient logistics. This means a decrease in stock and transaction costs. This development takes place in markets with only a few companies or a few very dominant companies (like in the the retail and car industries). The IT developments also support the formation of more loose relationships. Companies determine which is the best supplier for each order. This is possible if the markets get global and thus more companies compete and if information about the products is readily available and orders can be easily construed on-line. This is the case in markets with standardised products and many global suppliers (like in the computer and telecom industry) see Malone et al. (1987) for an analysis of market versus hierarchical relations.

Sales and catalogs Sales support Some of the questions to be answered when deciding how the sales function can be supported by the Internet are: . What type of product do you sell? Standardised/commodity products or specialized build on order? . What does the market look like? Is it open to new companies, how many suppliers and customers are there, are there

long-term relationships (contracts) between companies or are most transactions one-at-a-time? These questions determine for a large part the way in which products can be sold through the Internet. Of course the answer to the first question is often that products are neither completely standard nor completely customized. Many products are standard, but have a number of parameters that can be varied for each customer. The car industry forms a good example. Although a certain model of a car brand has some standard features, the customer can usually still choose the color of the car, sometimes even the type of motor and all kinds of extras. The crucial aspect actually is whether the products can be described in a standard format, such that a customer can completely determine the product he wants to purchase by indicating a combination of values for all parameters. For these cases the selling process can be more easily performed through electronic means. However, in many cases products are not standard at all and are specifically designed for a certain customer. In those cases the customer only has requirements on the functionality of the product, but the supplier determines what the product will actually look like. Although this process is more difficult to support it can still be supported in some ways. For instance, a catalog can be provided with a number of proto-products or even raw materials, which can be used as a basis from which the customized products are derived. An example is the knowledge base developed for MCB described in the previous section. The second question determines for a large part the business relations between the companies in a market and the way transactions are usually conducted. If a company only has a few (big) customers it does not make sense to sell the products through an open catalog on the WWW. Probably these customers have a strong relationship with the company and need more service than just a passive catalog that can be used by any customer. These customers usually communicate with the supplier using EDI types of communication. In EDI standard messages are used to order products. Often the products are identified through a unique product number, which indicates that this type of automatic ordering can usually only be done for standard products and is not useful for customized products. The same distinction as shown above between catalog based ordering and EDI (message) based ordering can also be seen in

[ 285 ]

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

the solutions offered by e.g. SAP. They provide a SAP online store, which is catalog based and can be connected to an internal SAP system to process orders. They provide the SAP business connector for companies that use message based order transactions. It is difficult to combine the two systems or even to migrate from one type of order processing system to the other. This implies that the choice as to whether orders will be catalog or message based has to be made when IT support for the sales process is started right away.

On-line catalogs Once a decision has been made to use an on-line catalog to support the sales process, the following questions have to be answered: . Use an on-line catalog and separate ordering system or integrate ordering with the catalog? . Sell on-line through the Internet, through private VAN or a combination? . What information should be provided to (potential) customers? Can this be done in the form of a personalized catalog? . If a catalog is used, where should this catalog be maintained? Who determines the format? Many companies start using the Internet by creating Web pages on the WWW. These Web pages give information about the company. Companies that already publish a paper catalog might decide to publish this catalog on the WWW as well. In this sense it is an extra service to the customers who already use the paper version. Ordering does not have to be coupled to the catalog yet. It can be done through an e-mail in which all information is given for the order or can be by phone or fax. Starting like this familiarizes a company with publishing its data on the WWW. It can also experiment with all types of searching facilities before getting into the complex situation of automatic ordering. A decision that also should be made early is whether automatic orders will be made through a private network connection (value added network) or through the Internet. The advantages of using a VAN are the security of the data and the reliability of the network. However, by definition, these networks are closed and therefore only available for already existing customers. Suppose the company decided to use an online catalog that is visible to both existing as well as potential customers. Then the next decision is what information should be included in the catalog. If there are many companies that supply the same type of

[ 286 ]

products and the products are quite standardised it is more likely that customers want to compare products on price, and price information becomes important for online selling. Most companies, however, try to avoid giving price information to potential customers. They will only provide this information for existing customers. The main reason is that in B2B transactions the services supporting the product are often more important than the price of the product. A dealer in office supplies stated it as follows: our customers do not pay for the pen or pencil, but for the service of having a pen or pencil at the right time on their desk. For example, timely delivery, logistic services, etc. become an integral part of the product offered. It is therefore difficult for customers to compare products of different suppliers based on the price for the physical part only. Another reason not to offer price information in B2B transactions is that prices may depend on long-term contracts, quantities ordered and constraints on delivery. In our experience the price information is mostly given to existing customers and personalized for each customer. Mainly, it becomes a service to connect order information to the accounting systems of the customer or a service to compare alternative products from the same supplier. For example, a client might want to compare the prices of bolts made of different types of steel in order to decide what is the best choice. Besides determining which information should be available in the catalog, the company also has to decide how the information should be available. Many catalogs still follow a traditional format where products are ordered according to categories that are determined by the supplier. This leads to trees that can be searched by the customer. However, the modern techniques make it very easy to implement different structures by creating links between products on all levels in the tree. In principle, customers will have a different view on the product than the supplier. For example instead of first distinguishing between six and eight sided bolts and only later looking at the length or the material of the bolt, they might be primarily interested in the length of the bolt and have no idea whether they want a six or eight sided bolt. The success of an on-line catalog is largely determined by how good the search mechanism connects with the way that the customer wants to find a product. So, this depends on the information contained in the

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

catalog and the indexes on the products. A new development in this area is the personalization of the catalog. The catalog can be completely customized for big customers that only see the subset of products that are of interest for them, and maybe in the way (including format) that is most suitable for their situation. This can all be done based on the same basic information of the general catalog. This ensures that all the information is updated whenever the basic catalog is updated as well. A simpler and less labor and cost intensive customization of the catalog is based on keeping track of the customer profile. This profile can be used to first show the products that the customer has shown interest in more often. It can also be used to suggest alternatives or other products based on customers with similar profiles. See Amazon.com for a good example. The last issue to be discussed is the matter of where the catalog should reside. In the first instance it seems logical that the catalog resides at the site of the supplier. However, large companies that have many different suppliers and sometimes already use electronic procurement products like Ariba or Commerce-one might want the supplier to deliver the information about its products in a certain (standard) format. There are also industries where the catalogs are combined in portals, which are maintained by branch organizations. The latter often happens in industries where standardisation is of prime importance and there are no dominant parties in the industry. This is the case, for example, in the Dutch installation tools industry. Where the catalog is hosted is mainly determined by the power structure in the particular market. The party that has most power will usually host the catalog and/or determine its format. Whenever catalogs of different suppliers have to be combined, like in the case when the catalogs reside with the customer or at an intermediate party, issues on data integration and accuracy come up. When the integrated catalog resides with the customer, the customer usually dictates the format in which the data should be provided. We will describe this situation in more detail in the next section on electronic procurement. In portals such as VerticalNet for machine tools, the catalogs of the suppliers are not integrated at all. The only feature that this portal offers is a uniform format in which information is provided. However, this format is still very free and mainly contains free text fields that describe different aspects of the products and their suppliers. The

portal contains a search function that allows a user to search the catalogs based on keywords. The suppliers are responsible for keeping their catalog up to date. Although this seems rather limited one should realise that the integration of independent databases is a notoriously difficult problem. The main difficulty lies in the fact that the semantics of database fields with the same names and/or syntax may differ in the databases. For example, a supplier address in one database might refer to the address to which an order should be sent, while in an other database it refers to the general address of the company (maybe of the head office), which can be quite different. Integration of the data can at this moment only be done manually. However, there are some tools that might be used to assist and partly automate this task. One of these tools is Concept Base which describes both the data as well as the conceptual model in the same format and which allows the writing rules to map concepts on to each other. Concept Base (Jarke et al., 1995) is used in the MeMo project (Dignum, 1998) in which a broker for B2B electronic commerce is developed. What is required to use this tool is that for each supplier a conceptual model of its catalog is provided. This can be made by experts in a few days. Given these conceptual models, the integration rules can be written easily in cooperation with the suppliers of the catalogs. Another advantage of using a tool like Concept Base is that the actual data can reside with the supplier. The integrated catalog provides a view on these data. This means that whenever suppliers update their catalog it is immediately visible in the integrated catalog. The only major operation comes from a change in the format of the local catalog (e.g. adding a feature to products). In that case the conceptual model has to be adjusted manually and a reintegration performed.

Integration Another major issue when using on-line catalogs is the integration of the catalog and ordering process with the back office. We distinguish the following points: . coupling of the catalog with internal databases; . order information and user profiling; and . integration with the customer’s back office Not all people realise that once a catalog is on-line it should be available 24 hours per day and seven days per week. This means that updating the information in the catalog should be done while it is on-line (or with a

[ 287 ]

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

very short interruption). It also implies that customers will contact the company 24 hours a day. Often this implies that some organizational changes have to be made to capture orders and requests at any time of the day and week. The second point deals with capturing the order of the customer and relaying it to the order processing system (usually an ERP system). Ideally, the customer can check the availability of the ordered items before the items are actually ordered. If this is not possible, the client should get a feedback as soon as possible. In the case of Borstlap (a producer of all types of fastening products such as screws, bolts etc.) many customers do not order directly but accumulate orders over a week or month. The main reason is that these products are much cheaper to transport in bulk and therefore it is cheaper for a customer to order a large quantity at one time rather than many small packages. It means that ‘‘shopping baskets’’ containing potential orders of the customer are kept between sessions. They serve two purposes. First, it is easier for the client to add to his order than having to keep track of all items himself. Second, Borstlap could use the information to predict order quantities and therefore ensure availabilities more accurately. The last issue that is not often addressed is the integration of the ordering process with back office of the customer. Often the order is made from a browser and all information contained in the order is not accessible for the customer after the order has been placed. It means that the customer has to insert the order in its own system again. Some companies will now send an e-mail confirming an order in a pre-specified format (often using XML), which can be used to insert the order also in the procurement system of the customer.

Electronic procurement In the previous section we have discussed some issues with regard to electronic sales and catalogs, from the point of view of a supplier. In this section we will review the transaction from the customer side. First we have to distinguish different functions in the procurement process. There is a strategic procurement function in which decisions on supplier selection are made. In the tactic procurement function decisions are made on contracts and contract renewal. Finally, in the operational procurement function the ordering, delivery and payment process is performed. It is primarily the operational

[ 288 ]

function that is automated in electronic procurement systems. So, these systems assume that there is a contract with a supplier that specifies the price and delivery conditions. They do, therefore, not offer any negotiation facilities, but only offer order and payment functions and sometimes tracking functions. The type of products that are most suitable to pilot electronic purchasing on are the ones that have low risk and low influence on the company results. These are the so-called routine products, according to Kraljic (1983). Office supplies form a good example. These products only account for a small part of the purchase value, but usually cause a major part of the procurement costs. The main goal in electronic procurement for these products will therefore be cost reduction. For the products that have a low risk in supply but a high impact on company results, the so-called leverage products, a company might look at using electronic markets. For these products, the price/quality ratio is more important than transaction costs. A good example of these products includes commodities that are part of the end product of a company. For example, most of the components of PCs are of this type. We will come back to electronic markets in the next section. The products that carry both a high risk in supply and have a high impact on the company results (the strategic products) can also be purchased electronically. However, these products are usually purchased directly through an EDI connection from the ERP system. A good example is the car parts for a car manufacturer. The delivery of these parts has to be in line with the company’s own production, while there is only one (or a few) possible supplier of that part. The last type of products are the bottleneck products that have a low impact on company results and a high supply risk. The fastening products of Borstlap are a good example of this category. There are only a few suppliers in this market, but the products have a low impact on the end product of the customers (like car manufacturers). For this type of product no readily available solutions are on the market. Most companies experiment with combinations of EDI and Web-based ordering. In the remainder of this section we will concentrate on the electronic procurement of routine products (like office supplies) for which most existing procurement systems are meant.

Costs and efficiency One of the benefits of electronic procurement is the reduction in transaction costs. In a

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

company like Oce Netherlands traditional orders will cost between $40 and $60 for handling the complete procedure. This is a rather standard amount in Western Europe and the USA. Although this amount looks large at first sight it is not that high if one realizes it includes the search of the product information, the typing in of an order, checking it, sending it, receiving the confirmation, etc. By automating the purchase procedure a company can reduce the costs to around $20 per order. This seems quite impressive, but a few remarks are in order at this place. First it should be noted that the costs incurred per order are, for a large part, labor costs. These are significantly lower outside Europe and the USA. Even though a company might achieve the same relative cost reduction all over the world, the absolute benefits become smaller when the labor costs diminish. In relatively small companies where orders may be at most once each working day, the cost reduction will be between $5,000 and $10,000 per year. Although this seems an interesting amount for a small company, this cost reduction can only be achieved when the whole procurement process has been automated. Systems readily available on the market cost between $100,000 and $1,000,000, meaning that a company would take at least ten years to break even. It is clear, therefore, that the benefits of cost reduction only arise for large companies that process hundreds of orders each week (or even day). However, there are at least two more benefits of electronic procurement that are worth discussing also for small companies. The first is reduction of errors. The fact that the whole process is automated means that the input can be checked for consistency and on flagrant mistakes. For instance, the name of a supplier can be taken from a fixed database, which also includes the other data for that supplier. Therefore these facts do not have to be typed anymore and errors are prevented in name of the supplier, address, bank account etc. Also the description of the product can be made unique. Often this is done already by giving products unique identifiers. However, retyping these numbers inevitably leads to mistakes. Using an automated procedure the product data are also input from the electronic catalog straight away, without human intervention. This also prevents mistakes in indicating quantities, which can occur if the quantities in which the product is sold by the supplier are not the quantities in which the customer normally thinks about that product. For example, ‘‘one stapler’’ might mean for the supplier ‘‘one box of 20 staplers’’.

The second benefit that electronic procurement offers is that of a reduced lead-time. In a paper administration an order form has to physically circulate through the company for authorisations, checks on budget, accounting etc. before it is sent to the supplier by fax or mail. Each step can easily take a day and therefore a lead-time of ten days for order processing is very common. Using electronic procurement the time that is needed for the electronic form to be passed through the network can be reduced to around ten minutes. This reduction of the lead-time of procurement can lead to significant changes for the company. Given this short lead-time it might decide to considerably reduce the stock of office supplies that it had to keep. A final benefit, which can also be stated as being an efficiency benefit, is that electronic procurement facilitates the bundling of the procurement. This leads to a better negotiation position with potential suppliers, because volumes are bundled and therefore discounts based on quantities can often be negotiated. The bundling of orders can also lead to logistics benefits. We will discuss this item in the next sub-section.

Logistics Logistics form an important aspect of any business transaction. The fact that a transaction is performed electronically might not seem to matter too much. Although this is true if one looks just at a single transaction, new opportunities and challenges exist when most transactions are performed electronically. Let us first look at a development that is taking place at many large companies, the shifting of the stock from the company to the supplier. Due to the rather long lead times in the ordering process, traditionally a company keeps a considerable stock of routine products like office supplies. Once a week (or month) new supplies are ordered to replenish the stock. Due to the dramatically reduced lead times of the order process, and therefore the possible quick reaction of the supplier, this stock can be reduced to almost zero. Whenever a product is needed, it can be delivered within a few days by the supplier to the person who needs it, thus eliminating the need to keep an expensive stock. Of course this also depends on a quick response of the supplier to an order. This is largely enforced simply by the competition in this type of market. Because routine products are often commodities that can, in principle, be supplied by many companies, the suppliers have to compete on more than just the price of the product. Both ASPA (supplier

[ 289 ]

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

[ 290 ]

of office supplies) as well as Borstlap (supplying fastening products) indicated that they did not compete on price but on logistics, which was indicated as being far more important for their customers. So, for both companies a quick reaction to an order is important. A consequence of this decision to distinguish oneself on the basis of logistics is that instead of regular large transports in which many orders are bundled, the companies will deliver smaller packages at a higher frequency. In order to keep efficiency this leads again to a new distribution process. Sometimes this may again lead to the outsourcing of the logistics to a specialised logistics company that is able to combine transports of several companies to the same destination. The actual developments in this field depend on the actual volume and geographic spread of the orders that a company has to process. For ASPA, which has a very high volume and wide geographic spread, it lead to establishing its own distribution centre and logistics department. Companies that have less high volumes might do better to outsource the logistics for efficiency reasons. There are also some logistics consequences for the customer. First of all, if every order is directly (and decentralised) sent to the supplier this might lead to high handling and transport costs. Therefore, companies try to bundle the orders as much as possible. However, the procurement should then keep track where different parts of an assembled order have to be delivered internally. This leads to the concept of ‘‘cross-docking’’, where deliveries of different suppliers are re-assembled into packages for different internal departments. An important aspect becomes the just-in-time delivery of the products and the coordination of the deliveries. If deliveries are not well coordinated or arrive too early, more warehouse space is needed than is available! Another aspect that is often neglected is that of wrapping of the products. The wrapping should be done according to the quantities that are ordered by the individual departments. If a number of orders are bundled into one large order which is packed by the supplier as one package, the distribution centre of the customer has to split up that package and repack the parts for internal transport to the departments again. Ideally, the supplier (or logistics company) supplies the products packaged in the way of the customer already. This reduces the costs for rewrapping and wrapping material considerably. For large companies like the RABO bank, even universities, this can lead to considerable cost benefits.

A final issue we will discuss concerning logistics is that of the on-line ordering of perishable products. This is illustrated by the case of on-line retailing where consumers can order their groceries on-line from a supermarket. The logistics of this form of electronic procurement are the critical success factor. Traditionally the consumers take care of the logistics by going to the shop, picking their own order and transporting it home. When a consumer orders on-line the supermarket has to take care of the order picking and either has to transport the products to the consumer or provide storage for the complete order. The costs of these operations are usually relatively high, while the consumer is usually not willing to pay (much) extra for the service. No satisfactory solution to this problem has been found yet.

Procedures and legal issues Given the developments sketched in the previous sections, can something be said about the procurement procedures and their changes? Three general points have become clear from the case study at the RABO bank where protocols have been designed for electronic procurement: 1 Roles and associated authorisations play an important role in the procurement procedure and should therefore be modeled in the protocols. 2 The protocol should provide for all possible exceptions and deviations from the default process. 3 It is important to agree on standard message formats such that the same format can be used for all suppliers. We will start our discussion with the point on the importance of roles and authorisations. In general, the procurement process starts with the selection of a product by the customer. After he composes his order it usually has to be approved by the person responsible for the budget. After this approval, the order can be transmitted. Some choices should be made here right away. One might take the most liberal point of view and let the supplier determine which products are available for the departments to order and offer the same selection to all persons in the company. This leads to a big variety of products on offer, but also demands a careful approval procedure for each order. On the other extreme the company might make its own selection of the products of the supplier to offer in its internal procurement catalog. It might also personalise this catalog such that a person can only compose an order for which he has the authority (or budget). This would eliminate much of the approval

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

procedure. However, it means a more complicated procedure to determine the person who can order a certain type of product. In practice a mixture of the two extremes is chosen. At the RABO bank, the bank determines itself which products are offered for the procurement. But the same assortment of products is available for all persons within the bank. The procurement system that ARIBA offers contains many possibilities to assign authority to departments and persons in order to facilitate the approval procedure. In principle, everyone can compose any order. But an order is only sent to a supplier if it has been digitally signed by an authorised person. It should be noted that the above procedure only discusses the operational procurement function. For the tactical (and strategic) functions separate procedures and (more important) authorities have to be established. These procedures will determine how products are added to and/or deleted from the internal catalog. The second general point seems almost to be trivial. However, it is surprising how many exceptions (of different types) are normally handled by the persons of the procurement department. The default procedure for the procurement is rather simple. First a request for an order is made, then the order is authorised, after the order is confirmed it is sent to the supplier, who sends an order acknowledgement. Later, a delivery notification is sent, the delivery is received and payment is made. However, at almost every step several alternatives and exceptions can be constructed. For instance, it might be the case that the supplier only has part of the order available. Therefore he might decide to split the order into two parts. The part containing products that can be delivered right away and a part which will be delivered at a later date (which might not even be known yet). Including this possibility into the protocol has quite some consequences. One of them is that it must be possible to refer to separate order lines and/ or products within an order. The different delivery dates should also be noted in the acceptance procedure of the customer such that a delivery is not rejected because it is incomplete! Another subtle point is that the procedure in general requires action from both parties in the transaction. However, it is not always sure that the other party will send the right message, in the right format at the right time. Therefore mechanisms should be included to capture these situations. For example, the procedure might require the supplier to send

an order confirmation after an availibility check. It might be that a supplier gets two orders at the same time. He performs an availibility check and would have enough supply for each order separately (but not for both). Now the supplier sends an order confirmation to both customers and then tries to allocate the products to the orders. In the last order he will find out that the stock is not big enough anymore. In a correct procedure it should be possible for the supplier to send a second message to the customer to withdraw the confirmation or to adjust it. This implies that the customer should be able to process such a message at almost any time during the process. Of course the customer may also decide to cancel the order at almost any point in time and the supplier should be able to react adequately to that event. The third general point that was already briefly mentioned is that an adequate standard message format should be designed. Although certain aspects are quite straightforward some others are less so. It might be handy to have unique identifiers for orderlines within an order. Also one might want to have a place for two product identifiers. One is used by the customer, and another one is used by the supplier. One might also want to include the possibility for repeating orders. That is, products that will have to be delivered repeatedly (every week or month). Ideally a company would like to use the same format for its messages in dealing with all its suppliers. Usually this leads to taking a union of all the fields that are required for each supplier and thus creating a large message that is half empty. (This is the situation in EDI standards like EDIFACT.) However, with the rise of XML one can now more readily define a flexible format. The format itself is included in the message and can be used by the supplier to determine the content. Peat and Webber (1997) give some recent developments in this area. The last item that we want to discuss in the light of procurement procedures is the legal issue concerning electronic procurement. If, in an ideal case, the whole procedure is handled electronically, it becomes very difficult to prove that the transaction actually took place. This might be beneficial for a company if it wants to avoid paying value added tax (VAT) on the transactions, but it might also cause problems in case of disputes over the transaction. Therefore companies have to check very carefully how the procedures can be designed in a way such that the legal status of the transaction is assured. Under the Dutch law, this means

[ 291 ]

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

that all electronic orders are compiled and printed once a week. This compiled record will serve as the basis for VAT payment. One should, however, carefully check the national legislation and find out what requirements follow from the law. Especially when the transactions cross national borders this might prove a difficult task, because the countries involved might require different legal documents and might accept different types of evidence. The main conclusion we can draw from our projects regarding this point is that efficiency and robustness are prime economical drivers for the design of the procurement procedure. The influence of the legal issues might be of great influence on the actual implementation of these procedures.

Electronic markets In the two sections above we have discussed transactions from both a supplier as well as a buyer perspective. We have seen that for both sides catalogs are used. For suppliers it means an extension of their advertising and sometimes just another sales channel. The buyers use catalogs to integrate the product supplies of all their suppliers in order to offer a uniform format to all internal customers. We can also see these two forms as instances of electronic markets. Electronic markets can be divided into four types: 1 Web sites. Connect one supplier with one customer. 2 Sales portal. Connect many suppliers with one customer. 3 Procurement portal. Connect many customers with one supplier. 4 Exchange. Connect many suppliers with many customers. Previously we have seen examples of Web sites, but have not discussed procurement portals. The reason is that none of the companies involved in our project provided this form of market. In B2C (business to consumer) this type of market is well known through the auctions at Internet sites such as e-Bay. In B2B this form of market is less often used. We have also discussed the use of sales portals, where the portal itself is organised by the procurement department of the customer. We have not discussed exchange markets yet, because this form of market is (usually) not organised by the customer or the supplier, but by an independent intermediary. At first people had the idea that Web sites, which model one-to-one relationships between customers and suppliers, would become the prevailing market model on the

[ 292 ]

Internet. The underlying idea was that IT would support many functions that traditionally were performed by intermediaries, therewith eliminating the need for intermediaries on the Internet. Customers would deal directly with the suppliers in a one-to-one format. Although this actually happens to some extent in e.g. the tourist sector where consumers can now directly order tickets from airlines and reserve rooms at hotels, the elimination of intermediaries is still not common in B2B transactions. Besides bringing supply and demand together (electronic) markets have a number of important functions: . they often decide on admission of buyers and/or sellers; . they might determine the quality of the products (this happens often in markets of perishable products like the fish market and the flower market); . they provide standardised formats for product specification, contracts etc.; . they provide protocols to arrive at contracts; . they provide logistics facilities; and . they provide payment facilities. These functions are also important for electronic transactions. Especially, because electronic market places are often opening up global markets with many unknown participants, ensuring trust and security become important aspects of the role of the electronic market. Therefore the party that organizes the market is of the utmost importance and should inspire trust in all parties trading on the market. Due to this fact, there are more examples of sales and procurement portals than of exchanges. Exchanges should in general be organized by independent parties (neither supplier nor customer in the market) that are trusted by all parties involved. In some cases this role can be fulfilled by a branch organisation, but also, e.g. financial institutes and Internet service providers try to take up this role. However, it is not yet clear which will be the most successful in establishing exchanges. On the other hand, the sales portal already has taken hold. Sales portals are for some part implemented as procurement systems for big companies. By organizing the procurement in this way, they can standardise supply information and have centralised control on procurement, while the actual buying of products can take place decentralised. A pre-condition for this situation is that the company has enough buying power to force the suppliers to trade through its portal.

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

Another popular way in which sales portals are realised is by closed markets where a number of companies bundle their procurement and thus establish leverage against the suppliers. A good example is Covisint (www.covisint.com), the virtual market place for the car industry. In these markets competitors cooperate in their procurement, because for a large part they use the same components for the products they produce. Their competition only arises in the way they add value to the components during assembly, etc. of the products. Another type of market uses tendering (reverse auctions). Auctions are very popular on the Internet, because the protocol to reach a contract is very simple (fixed message orders and only a few types of simple messages are exchanged). In this case the customer provides a (standard) product specification and all suppliers can provide a bid to supply the product. This type of auction is mainly used to reduce the costs of purchasing routine products. It should not be used to purchase products for which a close connection with the supplier is necessary, because for each tender a different supplier may come out as the winner. Finally, we want to give one example of an exchange market that is set up in the insurance branch. The reason for setting up this exchange was mainly to avoid unnecessary transaction costs and VAT costs. Many insurance companies also invest huge amounts of money in many different financial products and derivatives. In their trade they often exchange products between themselves. Instead of trading everything through the official stock markets they decided to exchange some of these products through a private exchange. At the end of the day, the positions of all parties are cleared and the differences are traded through the official market. Because the volumes that these companies trade are huge, the use of a private exchange that actually functions as a clearing house is very beneficial.

Conclusions In this paper we have discussed our experiences of a number of projects on electronic commerce in different types of industry. Had there been a uniform model to perform electronic commerce (even in B2B or in industry) we could have discussed our results in the light of such a framework. However, just as in traditional commerce, there is not one business model that fits all companies across all industries. Different aspects of electronic commerce are important

in different degrees depending on the type of products that are traded and the type of market structure existing in a particular industry. Because all business transactions at least include a buying and a selling party we decided to structure our experiences into a sales part and a procurement part, connecting the two in the previous section on the electronic markets themselves. The experiences that we had all point out that the type of relationships that a company has (or wants to have!) with its suppliers and customers is the most important determinant of how to start electronic commerce. If a company has relations with only a few suppliers and there are not many more potential suppliers of the same products, the company wants to bind those suppliers and form tight relations with them. This can be done by the introduction of a feedback system for individual suppliers (like at DAF Trucks) or through incorporating supplier catalogs in a procurement system (like at the RABO bank and the mental health care centre in Eindhoven). In these cases it would not make much sense to start an electronic auction or tendering system for procurement. If the company has only a few (big) customers or a fixed set of customers it might want to tighten the relationship with those customers. This can be done through providing them with extra information and giving them the opportunities to order online (as was done at Interbrew Netherlands). Also one might enhance the service level in order to provide support by making the knowledge that is needed to select the right product available through the WWW (a first step in this direction has been taken by the MCB steel company). In the B2C market this can be seen at many sites of financial institutions and insurance companies that are now offering on-line support to choose the product that fits your situation best (see e.g. www.egg.com and www.mrfinch.nl). Companies that supply products that fall in the category of routine products are likely to feel a stiffening of the competition. The use of electronic commerce makes it possible for the customers to reach more suppliers and get more information to compare the products. One way to distinguish oneself is by adding services to the products such that price comparison becomes more difficult and also the relationship with the customer becomes tighter. This can be done e.g. by offering logistics services and guaranteeing fast and accurate delivery. This has been done by both ASPA for office supplies as well as by Borstlap, providing fastening products. From the other side, a company that wants to reduce procurement costs might reduce

[ 293 ]

Frank Dignum E-commerce in production: some experiences Integrated Manufacturing Systems 13/5 [2002] 283–294

the transaction costs of procurement considerably by installing an automated (and centralised) procurement system. This will reduce the labor costs involved, but also serve to bundle strategic procurement functions while decentralising operational procurement functions. This can lead to better contracts with suppliers, while at the same time leading to logistics benefits. Some examples of this approach are RABO bank, Oce, the mental health care centre of Eindhoven and the town hall of Eindhoven. Throughout the paper and our experience it became clear that if a company wants to experiment with electronic commerce a good start is with routine products such as office supplies. These products have less strategic impact and therefore the company is less likely to be hurt if the project fails. It also became clear that a company can only gain benefits from electronic commerce if it restructures its organisation such that the relationships with suppliers and customers become central for its every day functioning.

The author would like to thank all the students and staff members participating in the strategic electronic commerce project ‘‘e-motion in e-business’’ conducted through the Centre for Electronic Business Research amd Applications (CEBRA) at the Eindhoven University of References Technology. Without their Dignum, F. (1998), ‘‘Mediating and monitoring work and numerous electronic commerce’’, Competing in the discussions this article could not have been written. Information Society, Genova, pp. 209-19.

[ 294 ]

Jarke, M., Gallersdo¨rfer, R., Jeusfeld, M.A., Staudt, M. and Eherer, S. (1995), ‘‘ConceptBase – a deductive object base for meta data management’’, Journal of Intelligent Information Systems, Special issue on advances in deductive object-oriented databases, Vol. 4 No. 2, pp. 167-92. Kalakota, R. and Robinson, M. (2001), e-Business 2.0, Addison Wesley, Boston, MA. Kalakota, R. and Whinston, A. (1996), Frontiers of Electronic Commerce, Addison-Wesley, Reading, MA. Kraljic, P. (1983), ‘‘Purchasing must become supply management’’, Harvard Business Review, September. Lechner, U., Stanoevska-Slabeva, K. and Tan, Y.-H. (2000), ‘‘Communities and platforms’’, special issue of Electronic Markets, Vol. 10 No. 4. Malone, T., Yates, J. and Benjamin, R. (1987), ‘‘Electronic markets and electronic hierarchies’’, Communications of the ACM, Vol. 30 No. 6, pp. 484-97. Peat, B. and Webber, D. (1997), Introducing XML/ EDI . . . , available at: http://www.xmledigroup.org/xmledigroup/start.htm Treacy, M. and Wiersema, F. (1997), The Discipline of Market Leaders, Perseus Press, Cambridge, MA. Veth, T. (2000), E-motion in E-business, CEBRA, Eindhoven.

Applying digital manufacturing technology to ship production and the maritime environment

Hongtae Kim Korea Research Institute of Ships and Ocean Engineering, KORDI, Daejeon, Korea Jong-Kap Lee Korea Research Institute of Ships and Ocean Engineering, KORDI, Daejeon, Korea Jin-Hyoung Park Korea Research Institute of Ships and Ocean Engineering, KORDI, Daejeon, Korea Beom-Jin Park Korea Research Institute of Ships and Ocean Engineering, KORDI, Daejeon, Korea Dong-Sik Jang Department of Industrial System and Information Engineering, Korea University, Seoul, Korea Keywords Computer-integrated manufacturing, Shipbuilding, Process design, Simulation

Abstract This paper is mainly concerned with the digital manufacturing technologies in the context of the shipbuilding industry. New concepts such as digital shipbuilding, virtual shipyard, and simulation-based design (SBD) will be explored. After reviewing the digital shipbuilding, a case study will be presented using the virtual assembly simulation system for shipbuilding (VASSS), a simulation based tool, to evaluate block erection sequence taking account of shipyard facilities, operational efficiency and equipment replacement time.

Received June 2001 Revised January 2002 Accepted February 2002

Integrated Manufacturing Systems 13/5 [2002] 295–305 # MCB UP Limited [ISSN 0957-6061] [DOI 10.1108/09576060210429748]

Introduction In recent years, global competition among companies and diversification of customer requirements have led to very short product life cycle and in order to adjust to these changes companies needs a new paradigm in the manufacturing environment. Information technologies (IT) centered on the Internet in the area of shipbuilding and marine engineering further incur the needs to increase the flexibility of the organization, the dispersion of work processes, and the use of out sourcing, as well as the globalization of related markets. In the near future, electronic commerce (e-commerce) and concurrent engineering (CE) based on CALS/EC and the Internet will be a integral part of the environment and upon these changes, ship design and production will become a computer supported cooperative work of many dispersed and specialized groups. As the means of active response to these environmental changes, many new concepts such as digital shipbuilding, virtual shipyard, and simulation based design (SBD) are appearing. In this paper, the concept and current status of digital manufacturing in general manufacturing industry will be reviewed. Then, related technologies, area of application and methods of digital shipbuilding in shipbuilding and marine industries are presented. In addition, virtual assembly simulation system for shipbuilding (VASSS), a tool for crane operability and block erection simulation in virtual dock based on 3D product model, will be introduced. The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0957-6061.htm

Digital manufacturing The manufacturing system related technology began from automation, advanced through flexible manufacturing systems and computer integrated manufacturing, and recently to implement an intelligent manufacturing system (Angster, 1997). As the manufacturing system related technology has reached the stage of intelligent manufacturing system, new concepts such as digital manufacturing, virtual manufacturing, and virtual engineering are appearing. Virtual manufacturing technology takes account of diverse fields in the area of engineering, and is called by many different names according to its applications and methodology. Digital manufacturing is the technology to provide rapid and efficient product development and production by composing an accurate and integrated computer model of physical and logical elements in the manufacturing system, and utilizing various computer technologies such as 3D CAD and simulation as the means of detecting errors in earlier stages and performing efficient decision making over the whole process of manufacturing. In other words, compared to the conventional manufacturing process of design, build, test, and fix, the virtual manufacturing process is design, verify, release and build (Shukla et al., 1996; Kraftcheck et al., 1997). The field of automobile production is where digital manufacturing technology is actively adopted, and in the USA, GM, Ford, and Diamler Chrysler are actively developing and applying virtual manufacturing and digital mock-up (DMU) technology as a next generation technology. In the case of Daimler Chrysler, virtual

[ 295 ]

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment Integrated Manufacturing Systems 13/5 [2002] 295–305

engineering technology is widely used, and in 1998 many types of automobiles were designed, tested, and prepared for manufacturing by computer. Most companies in the automobile industry in the USA have already finished applying DMU technology, and are actively applying virtual manufacturing technology in order to reduce cost, improve processes, create high added value, and improve competitiveness by reducing development and manufacturing time. Automobile companies in Japan and Europe are also seeking to reduce new automobile development processes and cost by applying DMU technology, sharing through standardization, and applying virtual manufacturing using digital data. Visual and virtual communication (VCOMM) and computer aided simultaneous engineering (CASE) project of Toyota are good examples.

Digital shipbuilding The shipbuilding process, starting from order to delivery, can be divided into design stage and manufacturing stage. Design stage can be further divided into contract design performed for the negotiation with ship owner, basic design to meet the requirements of ship owner, and manufacturing design performed in functional aspect. On the other hand, the manufacturing process includes pre-processing, processing, assembly, precedence outfitting, painting, precedence block erection, block erection, outfitting, etc. and these processes occur in a very complex pattern over a long period of time. Figure 1 shows the flow of each process composing the shipbuilding process. These shipbuilding processes are different from other typical manufacturing processes in the following characteristics: . ship type and form is very diverse and it is difficult to standardize since the design process is done according to the user’s requests; . material procurement and manufacturing begins while the design stage is not complete, so engineering changes and materials replacement are expected in the manufacturing stage; . shipbuilding is a labor intensive industry that is very difficult to mechanize and automate, so a lot of qualitative information is processed; . while materials are big and heavy, required accuracy is high and structure is complex, so it is difficult to standardize manufacturing process;

[ 296 ]

.

ships with different specifications are built at the same time, and a lot of information is required for management of each ship.

Because of the above reasons, building prototype for verification of product validity and quality assurance is practically impossible in the sense of manufacturing cost and time. Therefore, in order to increase efficiency in shipbuilding, extraction of detailed design and manufacturing information is required, and such information needs to be exchanged and integrated with simulation based manufacturing technology. Digital shipbuilding, in a wide sense, can be defined as the process of performing the whole shipbuilding process from conceptual design to operation and maintenance in computer model and simulation. It can also, in a production-oriented aspect, be defined as a process of modeling shipbuilding process and implementing the whole manufacturing process using an integrated database. Digital shipbuilding, similar to digital manufacturing in general manufacturing industry, is a manufacturing methodology to verify and solve various trial and error using computers and existing information technologies, to support efficient decision making by providing accurate information to those who need it at the right time, and thus to reduce time and cost. Generally, the digital shipbuilding approach can be divided into design centered; production centered and control centered approach. In other words, the best results can be achieved by establishing appropriate digital shipbuilding approaches. Digital shipbuilding provides an integrated environment for the whole shipbuilding process and eventually can be used as a fundamental technology for a virtual shipyard (Baum and Ramakrishnan, 1997; Behning et al., 1997). Basic technologies for digital shipbuilding are information technologies such as simulation, CAD, and visualization. In more concrete aspect, issues in applying digital shipbuilding to various engineering fields in the manufacturing area are as follows (Kim, 1998). For practical use of virtual prototype and SBD concept, there are a few issues to solve both in technical and cultural aspects. Representative issues in the technical aspect are high performance computing, high performance visualization and high-speed networking. These are needed for displaying complex graphic images at very high refreshing ratio (more than 30 frames per

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment

Figure 1 Flow of shipbuilding process

Integrated Manufacturing Systems 13/5 [2002] 295–305

second), they should be integrated as a part of the design and manufacturing system. Technologies such as high performance computing, high performance visualization and high-speed networking are rapidly advancing and they will be highly used in the near future. However, the most serious and difficult issues are integration and standardization. Integrating CAD/CAM/CAE is a very important element in SBD where it should overcome the whole spectrum of ship design engineering. In the ship design process more than 50 different technical areas should be integrated, Conventional CAE programs should be improved or extended for direct relationship with product model, and the functionality of the product model itself should also be supplemented. Standard for software and data exchange is also a very important issue. Especially the standard for different CAD systems to exchange their data such as standard for the exchange of product data (STEP) is a necessary element for SBD. The necessity and merit of a standard-based system is already proved in other areas, and use of standard language, open architecture of computer system, and advancement of networking have already solved a fair number of problems. In addition, though it is slow, ship STEP is also continuously developed.

Current research works and application examples Recently, many government funded researches have been done to turn the originally labor intensive and experience centered shipbuilding industry into a knowledge based and technology intensive industry with advancement of information technology based on computer technology. The USA is running a simulation based design (SBD) program in order to develop a design system/environment that can reduce the cost of new system design and developments, reduce development time, and verify and reduce the risks centered on the Navy’s Defense Advanced Research Projects Agency (DARPA) (Jones and Hankinson, 1994). Starting from 1996, phase I of the SBD program achieved establishment of environment for implementing SBD, and in phase II, SBD architecture that supports verification of phase I concept is being developed (Cardner, 1998; Fast, 1996). DARPA has tried to develop a few prototypes for applying the SBD concept in feasibility study and follow-up researches. Good examples are operation simulation of LPD-17, a next generation carrier, NSSN submarine development of General Dynamics Electric Boat Division, and Mobile Offshore Base of Gulf Coast Region Marine Technology Center (GCRMTC). The technology used in these development processes is being commercialized.

[ 297 ]

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment Integrated Manufacturing Systems 13/5 [2002] 295–305

[ 298 ]

In addition, Bath Iron Works (BIW) performed simulation of crane usage, floating dry-dock usage, dock and pier usage, installation and removal of production equipment, emergency vehicle movement and routes through yard, and personnel emergency evacuation routes from ship based on the results of LPD-17 development (Hugan, 2000). BIW is also trying to change the environment of the shipyard under digital manufacturing concepts for the production of LDP-19 based on previous experiences. BIW is currently introducing the concept of sharing manufacturing, engineering and operation information in virtual environment through electronic facility mockup since manufacturing planning and engineering related data are hard to acquire in the early stage under current shipbuilding design and manufacturing process. On the other hand, Virtual Reality Laboratory (VRL) in University of Michigan is doing research about applying virtual reality such as immersive virtual reality and augmented reality to industries (Beier, 2000). These researches include structural walk-through model, accident simulation and training simulation, and projects related to virtual prototyping and virtual reality are ship motion simulation, and virtual simulation of the shipbuilding process. Traditionally, Europe has been in the lead of shipbuilding system technology, and naval/marine related organizations conducting inter or international projects through European Strategic Program for Research and Development in Information Technology (ESPRIT) project. Representative of such projects is management and reuse of information over time (MARITIME) project led by Norwegian ship registration office (DNV) with many European naval related organizations and run as a part of ESPRIT III project. The main subject is to develop an infrastructure of next generation naval system based on ship product model, and an aim to lead naval system technology in the world through internal standardization and commercialization of the project results. On the other hand, the Department of Ship and Marine Technology in the University of Strathclyde is conducting research about computer technology application and interface to human factors as a means to achieve the objective of shipyards such as user requirement, competitiveness of ship, efficiency of cost and safety, under the rapidly changing environment of the shipbuilding industry (Vassalos et al., 2001). A main project includes ‘‘Sub-sea navigation

of remotely operated vehicle (ROV)’’ and ‘‘Evacuation simulation of Ro-Ro ferry ship’’. Japan is trying at government level to continue to keep current technology and competitiveness while turning the shipbuilding industry into a futuristic industry. Under the lead of Ship and Ocean Foundation (SOF), computer integrated manufacturing for shipbuilding (CIMS) project has begun since the mid 1980s and succeeded by general product model environment (GPME) project in 1996 for acquiring the technology for putting ship CIM model to practical use. Recently, based on GPME, advanced CIM, which is related to knowledge sharing technology, and LINKS project for implementing virtual shipyard under CALS concept are completed Shipbuilding industry in Korea is the top in the world in the amount of building, but in quality of technology, it is still behind other countries. Korea has supported the computerized ship design and production system (CSDP) project led by Korea Research Institute of Ships and Ocean Engineering (KRISO) to acquire ship CIM base technology, since 1996, ship manufacturing system integration technology development project for acquiring application technology is completed, and recently preparing for ship CALS/EC development.

Simulation based manufacturing for digital shipbuilding Up to now, researches related to digital shipbuilding and its application are very incomplete, and design demonstration of a constructed ship is partly done for the use of sales or assembly simulation of welding robots. Researches aiming at virtual shipyard in an integrated and overall aspect are yet to be done. The reasons for this is because new concept ships and large marine structures include thousands of different parts, therefore the performance and cost of hardware and software that can process such a vast amount of information can be difficult to achieve under irregular work environment. To achieve digital shipbuilding under integrated approach, the most important factor is successfully solving the planning problem. Planning in ship manufacturing process is the process of devising a method to minimize the effects of design change and delay, to support appropriate manufacturing methods, and a design making process for maximizing usage of resources. To sum it all, in order to accurately plan a shipbuilding

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment Integrated Manufacturing Systems 13/5 [2002] 295–305

process, the technology for implementing dynamic simulation is required. For dynamic simulation, definition of the whole shipyard facilities and manpower resources, and also of work and process for shipbuilding are required. The modeling of simulation is done after these definitions and to model and analyze manufacturing flow. Such a developed simulation model can analyze problems in product and process, and optimize the overall design process. Generally, development process of ship or marine structure is done in the order of conceptual design, basic design, detailed design, production design, building, maintenance, and operation. Table I summarizes detailed areas for digital shipbuilding. Previously mentioned digital shipbuilding related technologies aim to develop a system that can implement and verify the whole product life cycle including design, manufacturing, and maintenance under virtual computer environment based on 3D CAD system. For practical application of areas summarized in Table I, functional modeling of design and operational resources related to a ship’s functionalities based on 3D virtual ship prototype and process modeling of manufacturing process and related plan and manufacturing resources as shown in Figure 2 should be performed. The results of such modeling and information of a virtual ship should be shared by shipyard, ship registration office, engineering companies, ship owner, and marine transport companies under a distributed network environment.

Under the distributed network environment, function simulation, process simulation and safety simulation can be done. Function simulation includes verification of main functions of a ship such as the evaluation of general arrangement (GA), various equipments, and cargo loading/unloading. Process simulation includes evaluation of manufacturing process such as the evaluation of process and resource planning, and manufacturing automation equipment. Safety simulation became center of focus recently, and includes safety of ship and passengers’ analysis such as evacuation analysis and risk analysis.

Virtual assembly simulation system for shipbuilding Korea Research Institute of Ships and Ocean Engineering has been developing VASSS from 1998 to 2000 to perform crane operation verification and block erection simulation in a virtual dock based on 3D product model in development stage of a ship. VASSS was developed for dry dock in S shipyard. The following are overview and implementations of VASSS.

System configuration Generally in shipbuilding, establishing a manufacturing plan has an important role of connecting design and manufacturing, but due to the characteristics of shipbuilding industry, a manufacturing plan cannot be created and relayed at an appropriate time. In order to solve this problem, VASSS enabled simulation of block erection process plan items such as location of block erection

Table I Application areas of digital shipbuilding Step

Application area

Step

Application area

Collaborative design

Shipyard, class and ship owner

Production engineering

Design for assembly Design for manufacturing Design for maintenance

Concept design

Interface with 3D CAD system Visualization Immersion using virtual reality Availability of operation Manning requirement studies

Construction

Scheduling and resource planning Decision making of production Strategy Robotic off-line programming NC program inspection

Detail design

General arrangement studies Space allocation optimization Equipment selection optimization Evacuation route analysis

Training and maintenance

Crew training Training of maintenance Maintenance/training of equipment Safety training

Process planning

Planning of assembly sequence Worker and robot simulation Evaluation of equipment and facility Inspection of NC part and process

Operation

Accident simulation Ship motion simulation Cargo loading/unloading ROV operation simulation [ 299 ]

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment

Figure 2 Concept of simulation based shipbuilding process

Integrated Manufacturing Systems 13/5 [2002] 295–305

decision and erection sequence evaluation, taking account of information and ability of equipment from 3D CAD system. In other words, VASSS can evaluate block erection sequence by taking account of product data related to block and equipment such as goliath crane. Block erection process in shipbuilding is the last stage, but it is the first stage and basis in the flow of a manufacturing plan. Block erection process is the process of constructing a ship by assembling blocks transported to dry dock by crane. VASSS first constructs ship assembly workcell based on product data from 3D CAD system and block erection sequence from day plan system as shown in Figure 3. Then through process modeling, such as equipment usage pattern, generates path and sequence that can optimize collision free, erection time and availability. In this research, appropriate tools for system development are selected according to the scope and application area. Delmia’s ENVISION is used for virtual prototyping tool, and QUEST is used for process simulation. In addition, Window NT 4.0 is used for O/S. and Visual Basic, Visual Fortran and Visual C++ are used for application development.

[ 300 ]

Modeling Geometric information for block is entered in IGES data format, and various equipments such as Goliath crane and dock layout are modeled by device modeling from information extracted from appropriate blueprint. Device modeling of work environment is to model equipment and facilities except product related model from CAD system. Facilities and equipment for constructing virtual dock are dock, gate, goliath crane and various equipment. Modeling of workers and operators is done using ENVISION/ERGO by varying standard human data. Such human models are used in human walk through, turn-round, interaction with virtual objects, and collision detection.

Graphic user interface Most CAD systems and simulation systems have graphic user interface (GUI) appropriate for their systems to increase flexibility and work efficiency of systems. Generally, GUI is used for visualizing input, flexible connection with each module and visualization of intermediate and final results. In addition, it is possible to verify functions and efficiencies of the system to some degree while designing GUI. In this research, GUI is designed and used for user convenience and to verify functions in each system development step. GUI of

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment

Figure 3 Configurations of the VASSS

Integrated Manufacturing Systems 13/5 [2002] 295–305

virtual dock support simulation system prototype is implemented using graphic simulation language (GSL) and command line interpreter (CLI) functions provided by ENVISION. Menu is defined by each command or functions required in simulation, and top level command becomes top-level menu. Each sub command belonging to a command is composed with

pull-down menu. Figure 4 shows hierarchy chart of VASSS prototype menu. [FILE] menu shows the functionalities to define and manage workcell files and device fines required for virtual dock support simulation prototype system. Functionalities of [GEOMETRY] menu are to read existing hull, block, crane, and dock geometry for virtual dock support simulation. It is composed of sub menus such as [Ship Load],

Figure 4 Hierarchy chart of VASSC prototype menu

[ 301 ]

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment Integrated Manufacturing Systems 13/5 [2002] 295–305

[Block Load], [Crane Load], [Dock Load], and [Equipment Load]. Figure 5 shows the screen when a sub menu of [GEOMETRY] is selected. The functionality of [VIEW] menu is to change views showing simulation process. There are four sub menus of [ISO], [Top], [Front], [Right], and four-window view. In addition, there are other sub menus of [Wire], [Flat], [Smooth] and [Transparent] for showing model lines in different methods according to simulation characteristic. [ENVIRONMENT] menu is for defining environmental variables to describe workcell and define each device. Sub menus are [Block], [Crane], [Dock] and [Misc.]. Figure 6 shows [ENVIRONMENT] menu and its sub

Figure 5 Example of [GEOMETRY] menu

Figure 6 Example of [ENVIRONMENT] menu

menus. [PROCESS] menu loads day plan and process plan data required for simulation process and set the plans. [Sequence] sub menu loads block erection sequence from day plan system, and confirms for final sequence. [Pattern] sub menu is used for storing standard patterns for equipment in database and is used for block erection attributes. Based on previous block erection methods, 12 standard patterns for equipment are implemented as shown in Table II. These patterns should reflect shipyard dependent facilities and environment, and also time required for standard pattern for equipment should be considered. Figure 7 shows standard pattern for equipment when [Pattern] sub menu is selected. [SIMULATION] menu has the functionalities to operate workcells made of device, program and motion, and analyze the operation. It can be divided to continuous running and step running according to running type and step size can be set according to user request. In addition, the operation can be exported to AVI movie for presentation and its result can be visualized by movie. Figure 8 shows sequence chart of block erection through [Show] sub menu, and actual operation by [Run] sub menu. In [SIMULATION] menu, user can simulate workcell according to predefined scenario and read evaluation results of block erection sequence, time and crane usage. The results of virtual dock support simulation are block erection sequence, time, goliath crane usage, equipment replacement time, and evaluation of goliath crane evaluation.

Simulation In workcell implementation for assembly simulation, many children parts are attached to a parent part in hierarchical structure. Kinematics and dynamic implementation of the device makes it actually operate and detect collision with other workloads, workers or devices. It also includes motion control, setting path, GSL implementation, and collision check. GSL is ENVISION specific interactive programming language for operating devices. When block erection process is optimized, trajectory path is created, where overall sequence time is estimated using Gantt chart. Previous movement speed of crane is input and sequence is optimized based on the speed. Auto trajectory function is used for creating a collision free path. Auto trajectory function is used for creating collision free path between two tag points or the whole

[ 302 ]

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment

Table II Standard patterns for equipment of goliath crane

Integrated Manufacturing Systems 13/5 [2002] 295–305

Type

Upper trolley 150T loader 500T loader Beam Sheave

A

.

Lower trolley 200T loader 400T loader

.

Sheave

No. of goliath crane

. .

2

B-1

. .

.

B-2

. . . .

. .

C-1

. .

.

1

C-2

. . . .

. .

2

. . .

. .

2

D

.

E

.

. .

1 2

2

.

F

. .

. .

2

G

.

.

1

H

.

I J Figure 7 Example of [PROCESS] menu

Figure 8 Example of [SIMULATION] menu

.

. . .

1

.

.

.

1

.

1

trajectory. Before using this function, parts or devices that can collide with the device should be defined in collision queue, and after collision queue, assembly trajectory can be created. In order to use auto trajectory function, the user should determine if it is for two tag points or for all trajectories. In addition, options for deciding either transportation or rotation path should be predetermined. If rotation is selected, computational load is exponentially increased, so in most cases, only optimized path for transportation is used. Figure 9 shows the process of creating a collision free path. In creating block erection path, step size is very important since part movement and collision detection occurs at every step. If this step size is too small, a huge amount of time and computer memory is used for arbitrary solutions, and if it is too large, a solution may not exist since the accuracy is too low or collision may escape from two tag points. Trajectory verification function is used for in this case. If the verification of block erection process is perfectly done, actual simulation is performed to verify restraint conditions during motion process. Various methods such as Gantt chart, fly, fly-through, walkthrough, and view port are used to view simulation result. In addition, collision during goliath crane motion can be verified as a process of optimizing block erection sequence by goliath crane. Figure 10 shows the process of optimizing block erection sequence through the above process. Evaluation of block erection sequence is don by VASSS, and main functions are as follows:

[ 303 ]

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment

.

real-time visualization of block erection

.

process; selecting optimal path of goliath crane in

.

block erection process; calculating optimal time for block

.

erection process; visualization of block erection sequence

Integrated Manufacturing Systems 13/5 [2002] 295–305

. .

in diagrams; evaluation of block erection methods; evaluation of block erection sequence.

Figure 9 Creation of auto collision free path

Evaluation of block erection method Block erection method is the most important factor in block erection sequence and they are horizontal method, vertical method, pyramid method and multi-point method. Horizontal method starts from keel block to the direction bow and stern, and the order of construction is from ship’s bottom through traverse bulkhead, and longitudinal bulkhead to shell plating. Vertical method is faster in upper direction and from ship’s bottom to upper deck is constructed as one unit. Pyramid method is between horizontal and vertical method, and the sequence is from ship’s bottom through bulkhead block and shell block to upper deck. Multi-point method uses more than two keel blocks in construction. This paper used container ship for evaluation of the block erection methods. Table III shows the evaluation results for block erection methods using goliath crane. Lead time here considers movement speed of goliath crane, equipment replacement time, and block fastening time. As shown in Table III, horizontal and vertical method are almost identical in lead time and multi-point method takes the least time.

Conclusion

Figure 10 Optimal path of erection process

Shipbuilding industry is a key industry aiming to build a marine power country of the twenty-first century, and should be maintained and advanced in information driven society of future. For this, innovation in conventional design and manufacturing process, and information/automation technology should be supplemented. Current shipbuilding industries promote next generation shipbuilding systems based on CAD/CAM systems. Naturally, these systems automate conventional processes and increase productivity and quality, but they are limited in introducing fundamental changes in conventional processes or accepting rapidly advancing technologies efficiently. Recently, simulation based manufacturing which can model and simulate the whole process of manufacturing products,

Table III Simulation results for evaluation of erection method Method

[ 304 ]

Lead time (min) Remark

Horizontal Vertical Pyramid

1665.958 1666.276 1658.323

Multi-point

1623.345

Goliath crane 2 Total hoisting capacity: 450 ton Travelling speed: 30m/min

Hongtae Kim, Jong-Kap Lee, Jin-Hyoung Park, Beom-Jin Park and Dong-Sik Jang Applying digital manufacturing technology to ship production and the maritime environment Integrated Manufacturing Systems 13/5 [2002] 295–305

including design based on product model in computer, is the center of attention. This simulation based manufacturing environment is called virtual manufacturing, digital manufacturing, or virtual factory, and it is implemented in concepts such as virtual shipyard, and digital shipbuilding. In order to develop practical systems based on the contents of this paper, basically, product model based 3D CAD system and product data management (PDM) technology is needed. Furthermore, technologies for the process of block erection, assembly, processing, cutting and testing should also be developed with application development and operational verification. In the future, by applying virtual simulation technology to design, modeling, analysis, simulation, manufacturing, testing, and information system under simulation based manufacturing environment, the basis for a concurrent engineering system can be built, and with CALS/EC, it will be a principal factor affecting the competitiveness of the shipbuilding industry.

References Angster, S. (1997), ‘‘VEDAM: virtual environments for design and manufacturing’’, VR News, Vol. 6 No. 5, pp. 16-29. Baum, S.T. and Ramakrishnan, R. (1997), ‘‘Applying 3D product modeling technology to shipbuilding’’, Marine Technology, Vol. 34 No. 1, pp. 56-65. Behning, A., Cary, T. and Wittmeyer, J. (1997), ‘‘Simulation and visualization opportunities in the ship production and maritime

environment’’, Proceedings of 1997 Ship Production Symposium, SNAME. Beier, K. (2000), ‘‘Web-based virtual reality in design and manufacturing applications’’, Proceedings of COMPIT’2000, Germany. Cardner, J. (1998), Simulation of Mobile Offshore Base, Project Report, GCRMTC. Fast, K. (1996), ‘‘EVS at electric boat’’, Proceedings of 1996 Deneb’s User Group, Michigan, MI. Hugan, J.C. (2000), ‘‘Using simulation to evaluate cargo ship design on the LPD17 program’’, Proceedings of the 2000 Winter Simulation Conference, Orlando, FL. Jones, G. and Hankinson, T. (1994), ‘‘Simulation based design for ship design and acquisition’’, Proceedings of the 9th International Conference on Computer Applications in Shipbuilding, Germany. Kim, H. (1998), ‘‘Collaborative virtual prototyping technology for engineering simulation’’, Journal of Ships and Ocean Engineering, Vol 27, pp. 10-22 (Korean). Kraftcheck, J., Dani, T. and Gadh, R. (1997), ‘‘State of the art in virtual design and manufacturing’’, VR News, Vol. 6 No. 4, pp. 16-22. Shukla, C., Vazquez, M. and Chen, F.F. (1996), ‘‘Virtual manufacturing: overview’’, Computers and Industrial Engineering, Vol. 31 No. 1/2, pp. 79-82. Vassalos, D., Kim, H., Christiansen, G. and Majumder, J. (2001), ‘‘A mesoscopic model for passenger evacuation simulation in a virtual ship environment and performance-based evaluation’’, Proceedings of Conference on Pedestrian and Evacuation Dynamics, Germany.

[ 305 ]

ASP-driven e-service platform for Web-based online manufacturing Pingyu Jiang CAD/CAM Institute, School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi, China Guanghui Zhou CAD/CAM Institute, School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi, China Yong Liu CAD/CAM Institute, School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi, China Keywords Internet, Computer-integrated manufacturing, Applications software

Abstract Extends the concept of e-service to the whole phase of manufacturing. Develops an e-service platform prototype based on this concept with Java Web solution including the mobile agent broking technologies and application service provider (ASP) principle. The key point to implement the platform is to enable an open Web information service infrastructure for the whole product manufacturing chain. Inside this infrastructure, productspecific online manufacturing system can be created by means of using a kind of bidding model. All users participating in the manufacturing process are able to cooperatively finish manufacturing tasks in real time through sharing the same platform. With the help of BOM flow, the global information service flow can be controlled easily. In addition, the legacy hardware/software can also be encapsulated with aglets that are Java mobile agents. As to new ASP software packages, they can be configured simply via the plug and play mode to the e-service platform. In this way, the on-line networked manufacturing can be tested.

Received April 2001 Revised December 2001 Accepted February 2002

Integrated Manufacturing Systems 13/5 [2002] 306–318 # MCB UP Limited [ISSN 0957-6061] [DOI 10.1108/09576060210429757]

[ 306 ]

.

Introduction The rapid progress of information and network technologies is now changing the situation of the worldwide market competition. Globalization is becoming a trend for both present and future industrial practice. One of the factors to go to success under this competitive environment is the ability to quickly respond to market and provide the high-quality and lowcost product manufacturing without the limitation of working sites. Unfortunately, the traditional manufacturing modes/enterprises cannot satisfy these requirements well. So they have to face such a fatal revolution, which often deals with changing their running modes, re-configuring their resources and organizational structures, and so on. One solution to meet the needs of this revolution is to establish a novel Web-based online manufacturing mode that is market-driven and has a fast responsive capability to market. In fact, a Web-based online manufacturing mode possesses some key natures like globalization, agility, customization, digitalization, etc. A lot of researches in this area are being undertaken. However, what are key points to implement a Web-based online manufacturing is still under investigation. In our opinion, at least the following issues should be emphasized: . The concept of information service should go through the whole manufacturing chain. Manufacturing process may be equal to the correspondent service process. . Application service provider (ASP) may be used as a kind of philosophy to implement the infrastructure of Web-based online manufacturing. The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0957-6061.htm

.

.

.

Executing, controlling and scheduling manufacturing tasks in Web-based online manufacturing systems should work in a full-scale and explicit mode and can be tracked easily. Collaboration is always important to enhance the capability of online manufacturing. Encapsulating the legacy hardware/ software and configuring them on Web would solve some application bottlenecks due to the lack of new bottom-level manufacturing resources support. The security mechanism should be always emphasized in developing the Web-based application.

Starting from the above points, we develop an ASP-driven e-service platform for Web-based online manufacturing. Here, manufacturing is only defined in a narrow range, not including design, sale, and recycle processes. One of the most important contributions in this paper is a conceptual extension of e-service, from after-manufacturing to in-manufacturing. In addition, our research emphasizes the use of ASP logic to implement a Web-based infrastructure and the correspondent platform tools. We do not focus on developing some concrete enabling systems such as ASP-based CAD, CAPP, PDM and ERP, but integrate such systems into the e-service platform.

Related work As mentioned above, the researches on Web-based online manufacturing are being given more attention in both academic and industrial institutions in recent years. A strategic project, entitled next generation manufacturing (NGM) and led by Massachusetts Institute of Technology, refers that a corporation based on NGM is a part of global corporations. The global corporations are a cluster of institutions that can collaboratively produce products and provide

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

services by sharing the information, knowledge and resources. The manufacturing activities are optimized into the whole and the exact information can be transmitted to the exact place at the exact time by using the integrated, inter-operative information systems. Here, the importance of the information service mechanism for Web-based manufacturing is addressed (Jordan et al., 1999; Agile Forum and MIT, 1997). TEAM project provides a new approach that the agility and globalization of manufacturing enterprises can be implemented through an integrated framework for flexibly configuring the advanced manufacturing techniques. In this project, the integrated framework and networked manufacturing system based on the agent techniques are two key points (TEAM, 1998). The goal of the GlobalMan21 project is to set up e-commerce processes, methods and key enabling techniques for the manufacturing environment in the twenty-first century. The intelligent tele-design and manufacturing environment is one of its focuses (Japanese Domestic Summary Report, 1999). The CyberCut project by Paul Wright et al. emphasizes developing a Web-based manufacturing architecture to quickly design and manufacture products over the Internet. Under this architecture, Web-CAD and Web-CAPP are developed. At the same time, CNC machine tools can be controlled remotely (CyberCut, 1999). Cheng et al. (2001) also put forward a Web-based manufacturing environment. On the basis of analyzing the functions of existing and standalone product design and manufacturing systems, they use Java solution and accomplish the information interaction and share between sub-systems. Huang and Mak (1999) and Huang et al. (2000) have done a series of work for DFX and product development based on broking agent and workflow technologies. With their help, information integration and share in the phase of product development can be reached in the manner of DFX. From the view of digital enterprises, Lee and Lau (1998) and Digital Factory (2001) have worked for networked manufacturing and brought forward the methodology of digital factory. Based on multi-agent method and the broking technology of CORBA, ManAge system by Heikkila et al. (2001) and CBMAS by Yang et al. (2000) are developed (see also Cheng, 2000). Authors have also done much for tele-design and manufacturing and the integrated framework for Web-based manufacturing since 1996. The main achievement is ‘‘TeleRP’’, which is an Internet Web-based solution for remote rapid

prototyping service and maintenance (Jiang and Fukuda, 2001).

Overall architecture of e-service platform Definition of basic concepts Before describing the overall architecture of the e-service platform, we have to define some basic concepts used in this paper. Manufacturing task is defined as a description in which planning, assigning, ordering, machining or assembling operations for parts/assemblies/product are specified. Especially, a product may be fabricated through triggering a series of manufacturing tasks. There are more than one mapping relations between the product tree and such tasks depending on the selection of manufacturing modes. Manufacturing chain is defined as a group of time-related directed block diagrams that consist of a series of triggered manufacturing tasks. The generation of a manufacturing chain depends on triggering the feasible manufacturing tasks. Customer is defined as a kind of platform user who locates in any place, submits manufacturing tasks and plans to finish them through the platform. Manufacturing site is defined as a kind of geographically distributive place where various existing manufacturing systems are installed. Provider is defined as a kind of platform user who locates in a manufacturing site and provides the manufacturing service if necessary. Online manufacturing system is defined as a temporary configured manufacturing system for finishing manufacturing tasks submitted by customers. It is a combination of different manufacturing sites where the whole or partial facilities are contributed to form the manufacturing system. Bill of materials (BOM) flow is defined as a group of correlative data depending on the product structure configuration. These correlative data, like process plans, are created dynamically during the different phases of manufacturing processes.

Overall architecture and running logic The overall structure of the e-service platform for Web-based online manufacturing can be seen in Figure 1. A three-tier ‘‘browser/server/database’’ architecture is used to host this platform.

[ 307 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing

Figure 1 Architecture of e-service platform for Web-based online manufacturing

Integrated Manufacturing Systems 13/5 [2002] 306–318

In client sides, there are two kinds of users, that is, customers who submit manufacturing tasks to the platform and providers who register the manufacturing site for getting the order of potential manufacturing tasks from the platform. Through the correspondent Web portal, customers and providers can interact with each other. Besides Web portals, in server side, a series of ASP-based tools are configured to enable the functions of e-service platform under the support of a cluster of Java-enabled Web servers and Web databases. Typically, above server-side platform tools include: 1 Tool for modeling and scheduling the activity-based manufacturing chain. 2 Bidding tool for manufacturing tasks. 3 Tools for managing online manufacturing processes: . managing tool for manufacturing tasks; . customized visual monitoring tool for manufacturing processes. 4 Tool for manufacturing collaboration. 5 Tools for manufacturing resources management: . tool for the encapsulation and management of legacy manufacturing software/hardware; . tool for customer relationship management; and . tool for manufacturing site management.

[ 308 ]

The tool for modeling and scheduling the activity-based manufacturing chain is the core of enabling the e-service platform. To generate the correlative manufacturing data concerning BOM and form the BOM flow, a two-layer activity model is used for building a dynamic manufacturing chain that is connected with manufacturing history. The scheduling operation is also based on this model. In addition, querying and managing the manufacturing history through the manufacturing chain is also included in this tool. The bidding tool of manufacturing tasks is used to choose suitable manufacturing sites to form an online manufacturing system for special manufacturing tasks by means of the negotiation between customers and providers. The VRML model of manufacturing site is used for a visual evaluation supported by virtual tele-presence method. Managing and tracking the contracts are also emphasized in this tool. Tools for managing online manufacturing processes mainly deal with managing manufacturing tasks and implementing the customized visual monitoring during manufacturing processes. The first tool is used for maintaining and managing all manufacturing queues concerning different manufacturing sites. Actually, each manufacturing site can contain one or more manufacturing queues. Queue

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

operations such as adding or deleting a manufacturing task depend on time constraints. The second tool is just used for viewing the current manufacturing status under the control of customer privileges and managing the correspondent messages. There are two ways to reach ‘‘seeing’’ actions. One is to ‘‘push’’ the manufacturing information in the form of multimedia to the correspondent customers. Another is to use Web cameras for the visualization of remote online manufacturing process. The tool for manufacturing collaboration is used for manufacturing collaboration between customers and providers. A whiteboard-like collaborative space together with video clips from Web cameras, table loaders and various part geometric model loaders can be utilized synchronously to finish any remote discussion on Web if necessary. One of the important features is this tool can capture the history of collaborations as references or data mining sources in upcoming applications. Tools for manufacturing resources management include three tools, that is, the encapsulation and management of software/hardware, customer relationship management, and manufacturing site management. The tool for the encapsulation and management of manufacturing software/hardware is used for integrating existing traditional or novel ASP-based software into the e-service platform. Here, ASP-based software can be integrated into the platform only by using a very simple registration operation. For traditional software packages, the tool provides various templates of mobile agents to encapsulate them and to let the software work in the manner of plug-in. Here, the encapsulation of hardware like CNC machine tools is realized by encapsulating their front-end software. As to management functions of this tool, they mainly deal with the maintenance of mobile agent templates and encapsulated software registration, etc. The tool for customer relationship management is used for registering, querying and maintaining the data of customers. The tool for manufacturing site management is used for registering, querying, maintaining the manufacturing sites in four levels, which are: 1 plant level; 2 shop-flow level; 3 manufacturing cell level; and 4 individual equipment level. .

In database side, all dynamic and static data related to both the process/tasks and the product/assemblies/parts can be stored, queried, and managed. The buffer functions

are also provided for the exchange and sharing of data. Since a Web database is identified with IP address, it can be installed in any place. By using the above tools, a Web-based online manufacturing can be configured and used in following steps: 1 customer analyzes the product manufacturing nature and correlates assemblies/parts to different types, that is, outsourcing, and in-house and out-house machining; 2 customer does a plan to form a set of manufacturing tasks that can be finished via the platform; 3 bidding activity between the customer and potential providers runs through the platform; 4 an online manufacturing system is configured according to the bidding result; 5 manufacturing tasks are scheduled and executed via above tools of the platform.

Some key enabling technologies Web-based infrastructure As shown in Figure 2, in our research, Web-based Java solution is selected to enable the infrastructure of the e-service platform. Here, Internet computing model, Web portals, server-side configuration, database, security strategies are key points. Since the running logic of the platform is ASP-based, the above key points have to meet with the demands of ASP logic too. Under this case, we select Java applet-servlet-pair and IBM aglets programming models as the master models of Internet computing. The development of Web portals uses HTML/JSP techniques. The Web database appears in distributive mode. It provides a huge pool to support all platform tools to generate, exchange, and store manufacturing data. It also locates in the server-side. We use SQL-driven server and JDBC package to implement it. Because this infrastructure is built up on the Internet and TCP/IP adopts a quite loose security strategy, Internet computing must be conducted under the control of additional security strategies. Accordingly, the security strategy is mainly concerned with: . collection and confirmation of client’s certificates in different levels; . client’s access security consideration to Web servers, directories and files inside the servers, including individual Web pages; and

[ 309 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing

Figure 2 A Java Web solution to enable the e-service platform

Integrated Manufacturing Systems 13/5 [2002] 306–318

.

security encrypted data transaction over the Internet, which are only used for data inside applet-servlet pairs.

.

.

The server-side configuration includes the usage of a cluster of Java-enabled Web servers, database engines and aglet servers, the setup of directory structures that host Java applet-servlet-pair and IBM aglets, integrating method to put all together, etc.

Activity-based modeling and scheduling for manufacturing chain Mapping product online manufacturing tasks into the mainline of ‘‘virtual’’ manufacturing chain As mentioned above, there are three types of parts/assemblies existing in the configuration structure of a product, that is, outsourcing, in-house and out-house-machining parts/assemblies. They can be produced or ordered by using Web-based online manufacturing systems. In order to describe a manufacturing chain of parts/assemblies in the above online manufacturing systems, we define the following types of manufacturing tasks: . planning-type tasks in product, assembly and part levels, which decide what kinds of parts/assemblies will be fabricated in the online manufacturing system; . ordering-type tasks in assembly and part levels, which execute outsourcing operations;

[ 310 ]

.

assigning-type tasks like bidding, which decide how to configure the online manufacturing system; machining-type tasks, which execute machining operations; and assembly-type tasks, which execute assembling operations.

In fact, the product configuration structure can be described as a tree. Leaves in the tree express parts and other nodes stand for assemblies. Since different types of manufacturing tasks are able to correspond with relevant leaves or nodes of the tree, it is possible to plan potential manufacturing tasks according to this tree to pre-define an initial mainline/outline of a ‘‘virtual’’ manufacturing chain in which tasks are not triggered.

Activity model A manufacturing task is executed and used as an element of the manufacturing chain only when it is triggered. We define the whole procedure to run a triggered manufacturing task as an activity. It is a pre-defined activity when the manufacturing task comes from the above planned mainline. Otherwise, it is a stochastic activity for some special manufacturing processes like handling facility faults, modifying unsuitable manufacturing operations, etc. Furthermore, a pre-defined activity can be decomposed into a cycle of three sub-activities, that is, ‘‘do’’, ‘‘coordinate’’ and

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

‘‘redo’’ and a stochastic activity into a cycle of ‘‘coordinate’’ and ‘‘(re)do’’. In this way, an activity model including the manufacturing collaboration is introduced. Figure 3 shows the landscape of this activity model in the form of serial pre-defined activities. It should be pointed out that there also exists the parallel relationship among pre-defined activities.

‘‘Request-to-response’’-driven scheduling The purpose of scheduling is to trigger manufacturing tasks, to make them become activities and to form the manufacturing chain gradually. By means of the manufacturing chain, we can view the full-scale control flow of the online manufacturing. From the angle of information correlation, furthermore, the BOM flow can be browsed with the help of the manufacturing chain. Manufacturing tasks can be triggered in two ways, from the planned task mainline and from the stochastic task insertion. ‘‘Request-to-response’’ mode, which corresponds with the Web working principle, is used as a kind of scheduling mechanism for activating an activity. The context of the activity can be related to a series of ‘‘do’’ sub-activities if this activity is a pre-defined one. In the same way, a cycle of sub-activities also runs under the ‘‘request-to-response’’ mode.

Bidding for the configuration of Web-based online manufacturing system Creating manufacturing resource models It is important to build up manufacturing resource models for Web-based online manufacturing systems. In the research, manufacturing resources mainly refer to describing manufacturing sites, which are made up of facilities, FMC, FMS and work floor, and their manufacturing capabilities. To model a real manufacturing site, VRML-oriented facility blocks and their combination are used for presenting the correspondent real site on Web. This virtual site can be operated remotely in the form of tele-presence. Figure 4 shows us this. As to manufacturing capability descriptions, there are two levels, that is, dynamic facility level and static site level. Correlating to each facility of the above virtual site, there exist a series of descriptive templates of form-feature-based machining capability. In the template, the site-related machining parameter table, machining precisions, and possible machining cost per hour are included. When a machining-type manufacturing task comes, the template is instanced. So the data preparation for dynamic manufacturing capability evaluation can be implemented in this way. For the site level, some general parameters like status of delivery on time, service quality after manufacturing, the whole processing precision, manufacturing

Figure 3 Activity model for constructing a manufacturing chain

[ 311 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

managing ability, skill of technicians, productivity rate, facility running time per week, etc, are combined together for the synthetic description of the site capability. The site evaluation can be reached by using a fuzzy weight-added method developed by authors previously (Wang, 2000).

Bidding model for manufacturing system configuration Finishing the machining-type manufacturing tasks on Web needs a networked manufacturing system. In fact, a configurable online manufacturing system may consist of the different combination of several manufacturing sites. In such a site selected, the whole resources or only a part of them are used for configuring the above online manufacturing system. Making this selection depends on evaluating the facility-related and site-related manufacturing capabilities mentioned above, that is, the usage of manufacturing resources. Here, the selection procedure can be implemented in the form of bidding. According to the requirement of machining-type manufacturing tasks submitted by a customer, the e-service platform is used as a kind of agent to invite bids over the Internet. When potential manufacturing sites receive the information and find out some of them can satisfy the requirements, relevant sites will submit their own bids together with publishing their manufacturing capabilities and what resources will be involved. When receiving these bid documents, the customer has a comparison by using the e-service platform and decides which one will be selected to form an online manufacturing system. As

Figure 4 The constructing logic for virtual manufacturing systems

[ 312 ]

soon as the sites are chosen, the customer will sign bargains with relevant sites. In Figure 5, an implementing logic for bidding procedure is illustrated.

Encapsulation for legacy manufacturing hardware and non-ASP software based on mobile agents Despite the fact that more and more new hardware (such as networked NC devices) and software (such as ASP-based software) will be available in the near future, legacy hardware and software will remain in existence and have great influence for some time. Therefore, how to take advantage of the current existing legacy hardware/software and integrate them into Web-based online manufacturing systems is becoming a bottleneck. In this paper, we adopt mobile agent technologies to encapsulate current legacy hardware/software and to implement the functions of plug and play configuration on this platform.

What is a mobile agent? Mobile agents or mobile objects have their own behaviors, data, running status and itineraries. They can be dispatched from one computer and transported to a remote computer for execution. Arriving at the remote computer, they present their credentials and obtain access to local service and data. Compared with traditional agents, mobile agents have some benefits like object passing, autonomous, local interaction, disconnected operation, parallel execution, and so on. Figure 6 illustrates the work mechanism of mobile agents. At present, there exist some developing environments or

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing

Figure 5 The bid logic based on mobile agents

Integrated Manufacturing Systems 13/5 [2002] 306–318

Figure 6 The working mechanism of mobile agents

[ 313 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

developing tools for mobile agents, such as JAT, JAM, Voyager and Aglets. In this paper, we select Aglets Development Kit offered by IBM Japanese corporation to develop corresponding mobile agents.

Creating the encapsulating model for legacy hardware/software In general, legacy hardware/software systems include existing CAX systems, CNC manufacturing devices, and even manufacturing sites. Referring to these software/hardware systems to be encapsulated, we have developed a series of mobile ‘‘aglet templates’’. These ‘‘aglet templates’’ rely on the contents and formats of information exchange, that is, the input/output relationship of the hardware/software systems. When an ‘‘aglet template’’ is instantiated to generate a real aglet, the corresponding encapsulating interface of a remote legacy software/hardware system is only implemented in the server side and this software/hardware system is just plugged into the e-service platform. Running this aglet via dispatching to the remote place, where the existing software/hardware system works, makes the encapsulation play. In this way, a plug-and-play function is reached and existing resources can be shared via the platform. Figure 7 shows the encapsulating mechanisms for Solidworks, a commercial

CAD tool, with ‘‘Solidworks aglet template’’. The whole running procedure includes the steps listed as follows: 1 A user starts the ‘‘aglet template’’ for Solidworks encapsulation through a Web browser. Here, we must point out the current implementing framework is not based on a real B/S architecture, but a pseudo B/S architecture. Using this architecture, we can directly start corresponding aglet templates under the control of unified Web browser interfaces locally. 2 Through the aglet server, an aglet oriented from ‘‘Solidworks aglet template’’ can be generated and executed. 3 The user selects the corresponding initial information related to product/ assemblies/parts to be viewed or designed, and dispatches a slave aglet containing the design information to a destination aglet server. 4 The remote aglet server receives the dispatched slave aglet and gets the corresponding design information. 5 The remote operator browses the design information and starts local Solidworks software. 6 The remote operator uses local Solidworks software to implement the design task based on design information. 7 The dispatched aglet obtains the design results from Solidworks software operator.

Figure 7 Encapsulating mechanism for Solidworks with aglet template

[ 314 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

8 The aglet retracts the final design results from the destination aglet server. The user evaluates the design results and decides whether or not saving or feed-backing them.

Management for Web-based online manufacturing process Manufacturing queue management It is important to manage the manufacturing queues in different levels under the control of correspondent access permissions. In our research, site-level and facility-level queues are set up. Here, site-level queues are activity-based and facility-level queues are both activity- and sub-activity-based. According to the customer’s delivery requirements, each site or each facility uses three kinds of queue, that is, ‘‘express’’ queue, ‘‘urgent’’ queue, and ‘‘normal’’ queue, to host triggered manufacturing tasks. A method to manage these queues has been developed in our previous research (Jiang and Fukuda, 2001). Here, what we need to emphasize is that: . only providers can query and modify their own queues; and . customers can only browse the contents of queues which belong to themselves.

Customized visual monitoring The customized visual monitoring means a customer can access his/her own manufacturing tasks in queues through Web cameras, visual manufacturing status ‘‘push’’, and collaborative tools, only when the customer’s tasks are being handled. Here, the purpose of using Web cameras to ‘‘see’’ something and getting ‘‘pushed’’ visual manufacturing status is to collect the information. The purpose of using the collaborative tool among the customer and correspondent providers is to implement the information feedback and to make sure that online manufacturing can go smoothly by means of the customer’s join.

Collaboration and information sharing technology Synchronous collaboration based on manufacturing tasks Although we can use legacy collaborative tools like NetMeeting for information share in the form of pseudo B/S architecture, we still develop a novel synchronously collaborative tool in the real B/S mode. This tool is based on java media framework (JMF) and consists of a series of general ‘‘applet-servlet’’ pairs which integrate sound and video, and include whiteboard and online chat to support manufacturing collaborations. Here, the whiteboard can be used for displaying VRML geometric models

besides general images and for marking some comments on line. Because of the platform-independent natures of Java and B/S mode, this collaborative tool overcomes the shortages of legacy collaborative tools and can communicate and interact with one another by downloading to any operation system. Figure 8 shows the collaboration mechanism based on process planning for a valve part.

Privilege management for collaborative process Privilege management plays an important role in enabling the manufacturing collaborative process on this platform. It influences the system security and the function realization. In general, privilege management deals with the password check, new client registrations, and so on. Referring to the relationship among manufacturing tasks, product structure configuration, and relevant users, the working privileges can be assigned to each user. By using a Web database, all users’ privileges can be managed.

Case study Based on the above ideas, we develop a prototype of ASP-driven e-service platform for Web-based online manufacturing. Figure 9 is a snapshot of the Web portal. Currently, this system includes six functional modules that are: 1 manufacturing task planning; 2 bidding for configuring the Web-based online manufacturing system; 3 online manufacturing process management including task submission, queue handling, visual monitoring, etc; 4 encapsulation of manufacturing resources; 5 process planning for parts; and 6 synchronously collaborative tool. Here, activity-based scheduling for manufacturing chain is still under development. These modules can also be divided into several sub-functional modules. All modules and sub-modules on the platform are embodied with a service tree. The boughs represent the modules and the branches represent the sub-modules. Through clicking corresponding branch node with the mouse, we can enter the corresponding functional modules and implement corresponding functions. Using shaft and valve machining as an example, the customer can enter the platform from the Web portal shown in Figure 9 and

[ 315 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing

Figure 8 Collaboration mechanism based on process planning for a valve part

Integrated Manufacturing Systems 13/5 [2002] 306–318

Figure 9 The Web portal for e-service platform

[ 316 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

selects four manufacturing tasks, that is, planning, bidding, shaft machining, and valve machining. Through planning, the customer decides to send both parts for out-house machining. Then a bidding activity is executed based on the logic shown in Figure 5. During the bidding procedure, shown in Figure 4, a virtual manufacturing site on the Web can be viewed. As soon as an online manufacturing system is configured, the shaft and valve machining tasks can be triggered and executed. During the executive procedure, for example, when the provider does the process planning, he/she has to get the part’s geometric data from CAD system. Since the CAD system is not installed in the same place with the manufacturing site, the encapsulation-based CAD call is necessary. Figure 7 shows this. When the process plan is created, a ‘‘coordinate’’ sub-activity for evaluation can be selected and finished with the help of using a Webbased collaborative tool shown in Figure 8. As to online manufacturing process management and visual monitoring, the module shown in Figure 10 can be used. In this way, the online manufacturing of shaft and valve parts via e-service platform is realized.

Concluding remarks On the basis of analyzing the current researches for Web-based manufacturing technologies, this paper presents and builds up a prototype of ASP-driven e-service platform for Web-based online manufacturing. Using this prototype, we can fully take advantage of networked technologies to construct the online manufacturing systems from the point of ASP philosophy and e-service concept. Furthermore, with the activity scheduling mechanism, we can openly control to run the Web-based online manufacturing system in full-scale mode. Using the bid mechanism, we can implement dynamic allocation for manufacturing tasks with the help of the tele-presence of virtual manufacturing system on Web. With the broking technologies such as mobile agents, we can implement the encapsulation and the integration for legacy hardware/software to this platform. At present, this system is still under development and a number of further problems need to be solved. Some of them are as follows: . the evaluation for dynamic manufacturing capabilities is still under

Figure 10 The snapshot of online manufacturing process control and management tool

[ 317 ]

Pingyu Jiang, Guanghui Zhou and Yong Liu ASP-driven e-service platform for Web-based online manufacturing Integrated Manufacturing Systems 13/5 [2002] 306–318

.

.

research. In this prototype, we mainly adopt static evaluating indexes to evaluate corresponding manufacturing capabilities; bid mechanism needs to improve because its algorithm is not perfect currently; and activity-based scheduling model needs to be studied further.

It is expected that these issues will be solved in the near future. Also, we hope that this platform would provide a new research idea for the Web-based online manufacturing researches and establish a fundamental for agility manufacturing on Web.

References

The research is under the support of National 863 High-Tech Research and Development Program of China (Project No. 2001AA415230), University Key Teacher Foundation of China Ministry of Education, and Doctorate Foundation of Xi’an Jiaotong University (Project No. DFXJU200013). The authors hereby thank them for their financial aid.

[ 318 ]

Agile Forum and MIT (1997), Next-generation Manufacturing: A Framework for Action, Agile Form and MIT, Boston, MA. Cheng, T. (2000), ‘‘Research on some key issues in distributed networked manufacturing theory and technology’’, doctoral dissertation, Huazhong University of Science of Technology (in Chinese). Cheng, K., Pan, P.Y. and Harrison, D.K. (2001), ‘‘Web-based design and manufacturing support systems: implementation perspectives’’, International Journal of Computer Integrated Manufacturing, Vol. 14 No. 1, pp. 14-27. CyberCut (1999), Homepage, available at: http:// cybercut.berkely.edu Digital Factory (2001), available at: www.mf. polyu.edu.hk/DF/, The Hong Kong Polytechnic University. Heikkila¨, T., Kollingbaum, M, Valckenaers, P. and Bluemink, G.-J. (2001), ‘‘An agent architecture for manufacturing control: manage’’, Computers in Industry, Vol. 46, pp. 315-31.

Huang, G.Q. and Mak, K.L. (1999), ‘‘Design for manufacturing and assembly on the Internet’’, Computers in Industry, Vol. 38, pp. 17-30. Huang, G.Q., Huang, J. and Mak, K.L. (2000), ‘‘Agent-based workflow management in collaborative product development on the Internet’’, Computer-aided Design, Vol. 32, pp. 133-44. Japanese Domestic Summary Report (1999), ‘‘GlobalMan21: enterprise integration for global manufacturing for twenty-first century’’, Japanese Domestic Summary Report. Jiang, P. and Fukuda, S. (2001), ‘‘TeleRP – an Internet Web-based solution for remote rapid prototyping service and maintenance’’, International Journal of Computer Integrated Manufacturing, Vol. 14 No. 1, pp. 83-94. Jordan, J. and Michel, F. (1999), ‘‘Next generation manufacturing (NGM)’’, CASA/SME Blue Book Series, May. Lee, W.B. and Lau, H.C.W. (1999), ‘‘Multi-agent modeling of dispersed manufacturing networks’’, Expert Systems with Applications, Vol. 16, pp. 297-306. TEAM (1998), ‘‘TEAM Executive Summary’’, available at: http://cewww.eng.ornl.gov/ team/executive.html or http:// www.etsummit.osti.gov/oakridge/focus/ team.html Wang, J. (2000), ‘‘Building and using a Web-based database for telemanufacturing ability evaluation’’, BE Thesis, Xi’an Jiaotong University, Xi’an, July (in Chinese). Yang, S. et al. (2000), ‘‘Network manufacturing and enterprise integration’’, China Mechnical Engineering, Vol. 11 No. 1-2, pp. 45-8 (in Chinese).

A survey and implementation framework for industrial-oriented Web-based applications

Qingjin Peng Department of Mechanical and Industrial Engineering, The University of Manitoba, Winnipeg, Manitoba, Canada

Keywords Design, Manufacturing, Product design, Internet, Surveys

Abstract Internet technology is changing the way of product development, ranging from information gathering, product managing and commerce to product development, and maintenance. In order to obtain the evidence of the extent of Web-based applications in industry, we did a survey to examine the impact and need of the Internet in product development under industrial environments. This paper provides a view of Internet-based applications in Canadian industry based on the data received from 42 Canadian small and medium sized enterprises (SMEs), which results in a solution to improve current Web-based industrial applications. A framework of an industrial-oriented Web-centred system is proposed based on the demand found from the survey. Some examples are provided to demonstrate applications of the proposed framework.

Received August 2001 Revised January 2002 Accepted February 2002

Integrated Manufacturing Systems 13/5 [2002] 319–327 # MCB UP Limited [ISSN 0957-6061] [DOI 10.1108/09576060210429766]

Introduction and background The Internet is significantly affecting all aspects of product development, and the Internet is revolutionizing the whole product lifecycle process. It has improved and enhanced industrial capability in design and manufacturing. It is clearly evident that the Internet provides numerous solutions to manufacturing more efficiently and economically. This is an area where rapid progress is currently being made. The distributed, collaborated, and interactive product development processes are major features of Web-based industrial applications. Using the Internet in industry efficiently and cost-effectively has been highlighted by researchers. There is a variety of research and development that has been reported in Web-based applications, ranging from information management, product design and manufacturing to product improvement, and maintenance. A number of demonstrations of the Internet-based systems are available on the Web. The most common use of Web-based technology in industrial applications is currently in collaborative product design, supply chain management, and training (Huang and Mak, 2001). For examples, a virtual design studio (VDS) is presented in Tay and Chen (2001) to integrate the existing information technologies into the design process and to explore the possibility of extending of the collaborative design systems over the Internet. A networked CAD/CAM framework is proposed in Ping and Yu (2001), it expects to provide engineers with a handy tool to implement concurrent engineering. The framework’s architecture would be either built on the Internet or on an intranet of a company. Using this framework and the STEP translator, CAD/CAM users can share The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0957-6061.htm

their design files in a convenient way. An industrial case is illustrated in Richir et al. (2001), and it shows that a ‘‘digital engineering design process’’ allows quicker innovation in a more creative way and favours direct commercialization of industrial products. A WWW-based knowledge server is described in Rodgers et al. (2001), which supports designers in the early stages of their design decision-making activities. It aims to address knowledge management issues in modern design enterprises by providing effective and efficient mechanisms for the capture, storage and management of concept design knowledge. As a result of the Web-based incorporating collaborative technology, changes in the design process are observed in Nidamarthi et al. (2001). These changes include being able to capture design rationale and exchange ideas analogously with traditional methods, and being able to simulate the assembly sequence in the iterative design phase. It concluded that collaborative environments can be effective in helping teams overcome the problems associated with conflicting schedules, heterogeneous computing environments and data, and can be effective in facilitating consensus-based decision making necessary for collaborative design. In the application of manufacturing process, an Internet-based system is proposed in Su and Amin (2001) to collaborate over the Internet for the purpose of integration in the gear design and manufacture; the study of a threedimensional (3D) virtual prototype is described in Kuutti et al. (2001), intended for usability testing and concept validation over the Internet. The results indicate that 3D virtual reality prototypes can be used for remote usability testing and design evaluation. It is mentioned in Adamczyk and Kociolek (2001) that Internet global network to support distributed manufacturing is a new philosophy of manufacturing in which distributed participants of a machining

[ 319 ]

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications Integrated Manufacturing Systems 13/5 [2002] 319–327

process can communicate with each other interactively. The supply chain management and distributed manufacturing over the Internet in the automotive industry is discussed in Schaub and Krauss (2001). The object-oriented database fits seamlessly into the Internet/intranet approach and provides potential to realise virtual enterprises with minimum efforts and expenditures. Another area is Internet-based training, which can be used for a flexible, efficient alternative to instructor-led training (Kamens, 2001). However, most of the above-mentioned research is developed in the academic institute or commercial company. There is no evidence to show the extent of Web-based applications in industry, and the strong points and weaknesses in current Web-based industrial applications. In order to find out the state of Web-based applications in the manufacturing industry, e.g. to what extent industry has applied the Internet-based techniques in their business, we did a survey to examine the impact and need of the Internet in product development. We investigated Internet-based applications in the manufacturing industry. The questionnaire was sent to 120 SMEs. This paper provides a view of Internetbased applications in Canadian industry based on the data received from 42 Canadian SMEs, and suggested methods of facilitating the adoption of Internet-related technologies in manufacturing firms.

The question design and questionnaire distribution The structure of the questionnaire is shown in Figure 1. Design of the question is based on a product cycle, from market research, product design, manufacturing to after sale support. There are sub-questions in each part. A total of 88 questions are included in the questionnaire. The questionnaire begins by asking participants if they currently conduct business over the Internet. If the answer is ‘‘no’’, they are asked whether they are familiar with the potential of the Internet for business; if they intend to utilize it in the future; and finally, what concerns have kept them from utilizing the Internet thus far. They are also asked whether they wish to find out more about the possibilities the Internet offers to production before the survey concludes. If the answer to the first question of the questionnaire is ‘‘yes’’, the participants proceed directly to check off the parts of the product cycle, for which they utilize the Internet, and are instructed to

[ 320 ]

complete only those sections of the survey that deal with these parts. The survey questionnaires were distributed to 120 Canadian SMEs of manufacturing companies by four methods of distribution. These are post, e-mail, telephone and fax. The most fruitful method of distribution proved to be the regular mailing system. In light of the focus of this research, this is quite ironic; 42 per cent of the questionnaires distributed by this method were returned completely. The postal system has proven itself to be quite efficient and effective. An envelope in hand seems to leave a stronger impression than an e-mail, which can be easily deleted. Initially it was hoped that e-mail would be the primary means of distribution. It is the least costly method, and much faster than regular mail. It also helps in that the responder does not have to deal with any physical paperwork. It seems, however, that the e-mail surveys resemble junk e-mail and were likely deleted on all but one occasion. It is also likely that in the wake of the recent e-mail virus outbreaks, executives feared the worst when they received any unsolicited messages. This method was discontinued after 23 surveys were distributed and only a single response was received. The most effective means of survey distribution would have been the telephone. When the responder completes the survey over the phone, the results are in hand immediately. There is also less of a chance that the survey will not be completed, because even if the responder would be momentarily busy, he/she is asked for a mailing address or a fax number. When the survey does arrive, they will feel somewhat compelled to complete and return it. The telephone was underutilized, because of the time cost to both the surveyor and the responder, especially when both would be at their respective places of work during business hours. Also, the surveyor would accumulate significant long distance charges since most companies queried were outside the local area. The final method used was fax. Only a single questionnaire was distributed by this method, and that was previously solicited though the telephone. The reasons why this method was underutilized are similar to the telephone in terms of long distance charges.

The findings and discussion of the survey A brief look at the data from the survey can give us an overview of the way Canadian manufacturing firms make use of the Web.

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications

Figure 1 Questionnaire structure

Integrated Manufacturing Systems 13/5 [2002] 319–327

The results of the survey, broken down on a percentage basis, are shown in Figure 2. A closer look at the individual parts of Internet-based applications shows a more in-depth understanding. It can reveal the weaknesses and strengths within the individual application area. Utilizing today’s latest Web, Internet-based industrial applications have been called e-manufacturing or collaborative

Figure 2 An overview of Internet-based utilization from the survey

manufacturing. Figure 2 points to a generally low usage of the Internet for design purposes. Of the companies surveyed, 22.7 per cent do conduct some conceptual design with the help of Internet-related technologies. Nearly all respondents request e-mail feedback from customers to aid conceptual design, and many contract out the process to specialized design firms. This seems to follow an emerging trend of contracting out more and more of the portions of product development as the design process becomes complex. About one-quarter of the firms do use search engines to find out the requirements of customers, but perhaps at this point most have a fairly good idea, and further research is redundant. Online product research is rarely used. Only a few organizations use the Internet in the final part of product design, detailed design. About one-fifth of the companies use the Internet to contract out the process. Virtual reality (VR) and 3D modelling are in fairly common use among these select organizations. The usage of process design follows detailed design in the questionnaire. Although sample size is very small, data suggests that two-thirds of the companies

[ 321 ]

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications Integrated Manufacturing Systems 13/5 [2002] 319–327

[ 322 ]

that do use Internet-related technology for process design contract out the entire process, and half of them search for the design firms online. These findings seem to reinforce the trend of contracting out more and more of the design process as it becomes more involved. Just-in-time (JIT) manufacturing and supplier communication is conducted over the Internet by 40.9 per cent of the companies surveyed. It shares third place with intranet and Internet in terms of utilization. More than three-quarters of the organizations which responded to this section, and onethird of all organizations surveyed consider themselves JIT in Web. A further one-third of these seek out their JIT partners (suppliers), submit electronic tenders, and provide electronic confirmation of shipments received via the Internet. More than onequarter of all respondents to this section maintain online JIT schedules. Currently, 90.9 per cent of all firms surveyed have established an online presence with a WWW site; 22.73 per cent of respondents have provided the means for customers to order merchandise online. Of all respondents, only 4.55 per cent had made a secure means of payment online. The numbers demonstrate that there is a lack of awareness in the industry when it comes to selling online. There are methods of online advertising, however, that do seem to be neglected. Few manufacturing companies try to attract potential investors online, and even fewer hold online promotions to generate publicity. Online mailing lists also seem somewhat underutilized. This may be due to reluctance on the part of customers to be a part of mailing lists in general. The lack of interest in hiring opportunity spotters is not surprising. A total of 29.5 per cent of companies provide some type of after sales support online. This figure seems low considering the potential of the Internet to provide product support, but customers seem to prefer to talk with company representatives directly. Those firms that do provide sales support via the Internet have about a third customers’ orders online. All together, 38.6 per cent of respondents have either the intranet or Internet, 31.82 per cent supporting the intranet and 18.18 per cent supporting the Internet. The results suggest that the primary uses of company intranet are to aid the generation of product ideas, and in advertising. They are also used for market research and process design purposes. The Internet seems to be needed for advertising, sales support, and, occasionally,

for market research and to aid the JIT manufacturing process. All companies who do not use the Internet in their business intend to utilize the Internet in the future, most within the next two years. They have three main concerns that have prevented them from using the Internet so far: 1 security; 2 lack of technical know-how; and 3 lack of funding. In summary, the areas of high utilization generally include market research and advertising and selling, while the design stages show poor utilization; 9.1 per cent of the respondents claimed that they do not use the Internet to conduct business at all. This scenario reflects a narrow vision in industry, focusing purely on selling, rather than the full potential of the Internet, especially when it comes to improving product performance. Another surprising area of weakness is providing online support. Only 29.5 per cent of the respondents provide after sale support online. This figure is surprising considering that a Web page containing answers to frequently asked questions, or an automated e-mail reply system can be used to provide product support 24 hours per day, seven days a week to a customer who does not even have to leave home to get it. The convenience to customers and the savings to the company would be significant. Anecdotal evidence suggests that the reason for this trend may be customer driven. Customers seem to prefer talking directly to a live sales rep or technician, instead of logging on to a Web page that may not have the right answers. As well, since customers often do not use the same technical terms as the professionals who manufacture the product, any e-mail messages may be misunderstood, requiring further communication. The following conclusions from the above survey and discussion provide insight to improve the effectiveness of the Internet-based technique in industry: . Methods of education appear to be needed to raise awareness of the potential of the techniques described herein, especially for product design purposes. . Other areas where there is lower than expected deployment of the Web-based technology include after sale support and selling online. . In order to increase the usage of the technologies in question, it may bring about greater awareness of the potential of the Internet-related technology to aid manufacturing by increasing the deployment of such techniques for SMEs.

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications Integrated Manufacturing Systems 13/5 [2002] 319–327

.

.

The primary area that needs to be addressed is product design. Education may be needed to raise awareness of the possibilities of the Internet and related technologies, such as the Web-based VR and 3D virtual prototyping. This would certainly create an environment of expertise about the effectiveness of emerging technologies for different business purposes.

Even if the previous recommendations are followed, and indeed they prove successful, the smaller firms surveyed may still not be able to afford some of the technology discussed herein, especially when it comes to virtual design and 3D prototyping. To make some of these more costly techniques available to even small companies, technology centres should complement the expensive consulting firms operating today. At these centres clients could pay to use the technology themselves, and the service could be tied in with the aforementioned short courses to give the clients the necessary background on how to go about it. This would be of great benefit to clients that would, for example, only need to design a product once every two years. For them, purchasing the technology and hiring the qualified personnel would not be worth the returns since the technique is sparsely utilised. Based on this research, it may be argued that Web-based technology has been widely applied in industry. A significant percentage of companies involved do employ them. However, considering the cost and complexity of these technologies, and since very few companies are using any Internet-related techniques to facilitate detailed or process design, the low usage for Web-based tools in design is a more fundamental problem, and needs to be addressed on a more basic level. As well, perhaps the issue of cost and complexity should be addressed, making these techniques available to companies that could not otherwise afford them. It is critical to develop a user-friendly interface to support Web-based industrial applications. There are also strengths and weaknesses when it comes to the utilization of specific technologies and techniques. This is demonstrated by the simple common use of the Internet, such as e-mail and search engines, and the much more rare employment of the product design tools. This variance may be due to the relative maturity of the technology, its usefulness, and the degree of public’s awareness of them. A good interactive industrial-oriented interface for Web-based applications is critical to support

collaborative, data sharing, and product design and manufacturing. Great attention should be paid to the support of product design, earlier feedback of product performance from users, and the use of interactive visualizations to interpret and explore engineering data. Using VRML-based 3D displays of individual design requests can provide capabilities for examining data that exceed those of traditional Web-based analysis tools. A VRML-based 3D discrete event simulation tool is being developed by the author to support virtual industrial activities. The majority of real systems are complex, with many boundary conditions. The discrete event simulation can be used to model such systems, and to maximize the utilisation of resources.

The proposed Internet-based industrial application framework As shown in the survey and discussion, there is a need to develop general infrastructures that support flexible and effective cooperation among enterprises, with special focus on the needs of SMEs. The implementation of the approaches and associated systems is not an easy task because design and manufacturing are normally sophisticated engineering operations (Chang et al., 2001). A Web-based design and manufacturing support system may need to be supported by design information integration, remote execution of the systems, use of Java programming, client-server architecture, open computing and user interface design, and the humancomputer interaction. Some ideas and approaches from current research and development projects can contribute to the development of basic models and supporting infrastructures, such as using a knowledge management system to support the linking of artefacts to processes, flexible interaction and hypermedia services (Tiwana and Ramesh, 2001). For example, a systematic approach to the performance in the Web-centred industrial application is developed in Akarte et al. (2001) using the analytical hierarchy process, which enables the combination of tangible and intangible criteria and checking the consistency of decision making to improve manufacturing companies’ competitive position. Using the Internet technology, the virtual enterprise can be facilitated by the Web-based system and applications. It opens up a domain for building a global, network-centric, and spatially distributed industrial application, but it needs to be

[ 323 ]

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications Integrated Manufacturing Systems 13/5 [2002] 319–327

reliable and perform well. It is critical to develop an industrial-orientated Web application system for SMEs, especially in the design and manufacturing areas. The framework of an industrial-oriented Web-based platform, as shown in Figure 3, is proposed by the author. It can support product design, analysis, and the manufacturing process via the Internet. The software and hardware, as resources, are linked to the system server. The data, which can support the design and manufacturing, are collected and stored in a database. A Web browser-based user interface has been developed to link users to the system for applications. To build an industrial-oriented Web-centred application system, there is a need for a sound methodology, a disciplined and repeatable process, better development tools, and a set of good guidelines in order to minimize risks, and enhance Web site maintainability and quality. Therefore, a great deal of attention is paid to the architectural design, earlier feedback on product performance from users, and the use of interactive visualizations to provide users with abilities to interpret and explore Web-based engineering data. By using VRML-based 3D displays of individual design requests, the proposed system can provide capabilities for examining data that exceed those of traditional Web-based analysis tools. The user interface is designed with the intention of making it easy for non-specialists to use. The user interface can work on any Web browser that supports Java. With this simple user interface, one

user can shape the design tools from the Web browser with the visual feed back. A VRML-based 3D discrete event simulation system is used to support virtual industry activities. The majority of real systems are complex, with many boundary conditions. The discrete event simulation can model such systems to maximize the utilisation of shop floor resources. The framework of the industrial-oriented Web system can provide the following features: . uniform interface for easy integration of different design and manufacturing processes into the system’s framework; . intuitive used interface and adequate feedback; . low cost and easy for maintenance; . easily extendable to add more complex functionality.

Some examples During the development of the project, different set-ups of the system were tested in different application areas. Following are some examples to support design and manufacturing by the proposed framework.

The common mechanism design The common mechanisms, such as gears, springs, belt-drives, and bearings, are basic mechanical elements to support mechanical motion in engineering products. Some processes to select and design these mechanisms are complicated and rely on knowledge and experience. In order to reduce the demand for specialist and experience in

Figure 3 The framework of an industrial-oriented Web-based platform

[ 324 ]

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications Integrated Manufacturing Systems 13/5 [2002] 319–327

the design, a Web-based mechanism design system can support the interactive design process, the sharing database and design knowledge. A V-belt design process is shown in Figure 4. After design requirements are entered, the design data can be retrieved from the database, the design process can be done, and design parameters can be produced interactively.

Material requirements planning (MRP) MRP is a basic component in manufacturing planning and control process. It consists of logical procedures for managing inventories of component assemblies, sub-assemblies, parts, and raw materials in a manufacturing environment (Singh, 1996). The MRP system is used to determine how many of each item in the bill of materials must be manufactured or purchased and when. The traditional MRP system is not a simple-to-use and cost-efficient manufacturing tool because it is not platform independent, and can not be integrated with the supply chains. It can never respond quickly to competitive market demand. The proposed Web-based MRP system provides a dynamic and interactive production tool for SMEs, it includes a master production schedule, bill of materials, current inventory status and lead time of manufacturing parts, and schedule receipts. Figure 5 shows an example of the MRP processing.

Design for assembly (DFA) A Web-based DFA tool is developed as a part of the proposed framework in order to

Figure 4 A Web-based V-belt design process

improve the fact that products are presently designed by commercial CAD/CAM systems with little consideration for the product assembly environment. The goal is to reduce the time-consuming redesign, and the delay of the assembly process at the product manufacturing level. The DFA tool consists of a design for assembly modeling, evaluation, and process visualization. While a user performs an assembly operation online, the system is expected to automatically capture the user’s intent and automatically add assembly constraints to the product model. The knowledge-based reasoning mating constraints from a constraint-based model are used, and VRML-based 3D modelling is used to display a 3D assembly model in Web browsers. A DFA example is shown in Figure 6. The assemble time and cost are reduced by using the DFA analysis tool provided, the data is shown in Table I. The DFA model consists of product objects, subassembly objects, part objects and connector objects. The assembly-related information is organized by a hierarchically structured product data model in three levels. The top level is a table form that summarizes the product and associated subassembly components. This level defines the nature of the assembly between the parts. The link between mating features, such as cylindrical fit or block-slot fit is defined at the second level. On the third level, the shape of individual parts, such as block, cylinder, cone, and sphere, is defined. The platform is designed to support three basic components in a manufacturing system: resource, process and inventory. It can significantly improve each component in the way it is managed, leading to new challenges while ultimately promising to provide value. The likely future is collaborative design and manufacturing management that promises to make, for the first time, the dream of virtual integration a reality. Firms can improve their performance within the enterprise by Internet-enabled activities with an expanded physical scope that includes producers and users with an expanded functional scope that includes product design, manufacturing and marketing.

Conclusions There are high expectations on Internet technology to provide an adequate information infrastructure to support industrial collaboration. This paper examines Web-based industrial applications

[ 325 ]

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications

Figure 5 An example of the MRP processing

Integrated Manufacturing Systems 13/5 [2002] 319–327

Figure 6 The virtual design for assembly and virtual prototyping used in the analysis of an industrial cylinder: (a) original model; (b) improved model

Table I The improvement of proposed design Design change Combine the base eye, base cap and barrel eliminating all weld times Eliminate the base port and design the flow tube to enter the base of the barrel Combine the rod with the rod eye eliminating welding Totals [ 326 ]

Time savings (hrs)

Assembly cost savings ($)

1.000

72.00

0.375 0.250 1.625

18.00 18.00 108.00

through a questionnaire survey. The issues that emerged from the survey, the potentials and challenges of the Web-based technology, are identified and discussed in the Web-based product development. This paper also attempts to address issues with the objective of showing how SMEs can make use of the Internet technology in their product development. The proposed framework is described with its main components based on the demand found from the industrial survey. The approach and examples of information sharing and management are described. The implemented mechanisms to support industrial communications over the Internet are presented in this paper. Some examples are presented to illustrate the product design and manufacture via the virtual enterprise network. The realization and implementation of the industrialoriented Web-centred system are based on following understandings. Internet technology can provide the infrastructure for a powerful exchange of knowledge between companies and their customers. It is about forming a global community, and building better collaboration for the benefit of all. The SMEs require a reliable infrastructure connecting transparently to the global network of enterprises. Beyond information communication security, an operating

Qingjin Peng A survey and implementation framework for industrial-oriented Web-based applications Integrated Manufacturing Systems 13/5 [2002] 319–327

reliable infrastructure to manage information communication is required. The application of Web-centred technology is the key for SMEs to enter into the business-to-business electronic commerce. The quick development of information technologies and the enterprises’ tendency to concentrate on their core competencies, aiming to become agile to survive in the competitive actual market, increase the number of co-operations among enterprises. Virtual enterprises are an appropriate alternative and competitive advantage to SMEs.

References

Special thanks to John Molnar for his contribution in the survey of Internet-based industrial applications, and to Ariel Lazaro, Shawn Roch, Anthony Agtarap, and Sook Ling Tan for their work in the development of Webbased applications.

Adamczyk, Z. and Kociolek, K. (2001), ‘‘CAD/CAM technological environment creation as an interactive application on the Web’’, Journal of Materials Processing Technology, Vol. 109 No. 3, February, pp. 222-8. Akarte, M.M., Surendra, N.V., Ravi, B. and Rangaraj, N. (2001), ‘‘Web based casting supplier evaluation using analytical hierarchy process’’, Journal of the Operational Research Society, Vol. 52 No. 5, pp. 511-22. Cheng, K., Pan, P.Y. and Harrison, D.K. (2001), ‘‘Web-based design and manufacturing support systems: implementation perspectives’’, International Journal of Computer Integrated Manufacturing, Vol. 14 No. 1, pp. 14-27. Huang, G.Q. and Mak, K.L. (2001), ‘‘Issues in the development and implementation of Web applications for product design and manufacture’’, International Journal of Computer Integrated Manufacturing, Vol. 14 No. 1, pp. 125-35. Kamens, H. (2001), ‘‘E-learning: training operators via the Internet’’, Circuits Assembly, Vol. 12 No. 2, February, pp. 70-4. Kuutti, K., Battarbee, K., Sade, S., Mattelmaki, T., Keinonen, T., Teirikko, T. and Tornberg, A. (2001), ‘‘Virtual prototypes in usability testing’’, Proceedings of the 34th Annual Hawaii International Conference on System Sciences, IEEE Comput. Soc., Los Alamitos, CA, USA, p. 7. Nidamarthi, S., Allen, R.H. and Sriram, R.D. (2001), ‘‘Observations from supplementing the

traditional design process via Internet-based collaboration tools’’, International Journal of Computer Integrated Manufacturing, Vol. 14 No. 1, January-February, pp. 95-107. Ping, Y. and Yu, C. (2001), ‘‘A data exchange framework for networked CAD/CAM’’, Computers in Industry, Vol. 44 No. 2, March, pp. 131-40. Richir, S., Taravel, B. and Samier, H. (2001), ‘‘Information networks and technological innovation for industrial products’’, International Journal of Technology Management, Vol. 21 No. 3-4, pp. 420-7. Rodgers, P.A., Caldwell, N., Clarkson, PJ. and Huxor, A.P. (2001), ‘‘The management of concept design knowledge in modern product development organizations’’, International Journal of Computer Integrated Manufacturing, Vol. 14 No. 1, January-February, pp. 108-15. Schaub, G. and Krauss, E. (2001), ‘‘Supply chain management and distributed manufacturing in the automotive industry’’, E-business and Virtual Enterprises, Managing Business to Business Cooperation, IFIP TC5/WG5.3 (Second IFIP Working Conference on Infrastructures for Virtual Organizations: Managing Cooperation in Virtual Organizations and Electronic Business towards Smart Organizations), Kluwer Academic Publishers, Norwell, MA, pp. 251-8. Singh, N. (1996), Systems Approach to Computer-integrated Design and Manufacturing, John Wiley & Sons, New York, NY. Su, D. and Amin, N. (2001), ‘‘A CGI-based approach for remotely executing a large program for integration of design and manufacture over the Internet’’, International Journal of Computer Integrated Manufacturing, Vol. 14 No. 1, January-February, pp. 55-65. Tay, F. and Chen, M. (2001), ‘‘A shared multimedia design environment for concurrent engineering over the Internet’’, Concurrent Engineering: Research and Applications, Vol. 9 No. 1, March, pp. 55-63. Tiwana, A. and Ramesh, B. (2001), ‘‘A design knowledge management system to support collaborative information product evolution’’, Decision-support-systems, Vol. 31 No. 2, pp. 241-62.

[ 327 ]

Design agility and manufacturing responsiveness on the Web

J. Toussaint School of Engineering, Leeds Metropolitan University, Leeds, UK K. Cheng School of Engineering, Leeds Metropolitan University, Leeds, UK

Keywords Internet, Design, Computer-integrated manufacturing

.

Introduction

Over the last few years, the Internet has been extensively used by companies to Abstract This paper is to investigate a Web- communicate with their customers, to based engineering approach to advertise their products, and even to offer enable engineers to use, share, on-line transactions. To compete effectively, and simulate effectively and efficiently design and however, these companies are having to manufacturing data through the implement solutions which will allow them World Wide Web. The enabling to interactively communicate information techniques to implement such an related to product design, development and approach are critically discussed with an example based on a manufacturing within their own tolerancing application and two infrastructures. Internet-based software Java-based machining simulators solutions could offer scalability, easier for e-manufacturing purposes, as well as an example of real-time 3D implementation, and compatibility across implementation. The outcome of diverse information technology platforms this research is for helping and thus reduce incremental infrastructure designers and manufacturing engineers to be able accurately to investments. Furthermore, this would allow make the right decisions for their companies to cut product development and respective purposes. manufacturing costs, and greatly increase collaboration among partners internally and externally. Moreover, it has been widely agreed that the business success of manufacturing companies depends on their ability to Received March 2001 identify the needs of customers and to Revised November 2001 Accepted February 2002 quickly create products that meet these needs and can be produced at low costs, and with The authors would like to thank Mr Arnau Ferna`ndez the shortest delivery time. Achieving these goals is not solely a marketing problem, nor Escude´, Mr Jordi Anton Lopez, and Mr Jean Fabre is it solely a design problem or a for their assistance in manufacturing problem. It is rather a programming the part of the product development problem involving all systems illustrated. these functions: . speed; . agility; . responsiveness; and . knowledge.

Integrated Manufacturing Systems 13/5 [2002] 328–339 # MCB UP Limited [ISSN 0957-6061] [DOI 10.1108/09576060210429784]

[ 328 ]

These are the watchwords for success in this new business environment, as they act on: . bringing products to market faster; . responding faster to customers; The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0957-6061.htm

.

avoiding production problems related to inventory and materials; and reducing inventory and production costs.

Therefore, communication and collaboration across companies and their partners become vital. Internet/intranet/extranet are the key for enhancing their design agility and manufacturing responsiveness and thus their competiveness. In this paper a Web-based engineering approach is presented for developing design and manufacturing support systems, as well as the enabling techniques for its implementation. The interactive applications presented, being part of a major Web-based system dedicated to bearings, will also involve the implementation of client/server technology, and database integration, as well as extensive programming in HTML, JavaScript, and Java (Cheng et al., 2000). Particular attention was paid on creating and implementing a real-time interactive interface, which will lead the users directly into the application. The paper concludes with a discussion on the potential of the approach and particularly on possible link with a CAD/CAM application. The system aims to be a useful design tool, being able to quickly and responsively provide design and manufacturing data in a highly interactive way.

Design agility Engineers have always managed to achieve creative and inventive solutions to design problems. It does not matter where the problem is from, or in which particular area. Wherever there are people there are problems needing a solution. This last sentence has been true for hundreds of years and is still true even with the technology advances, because the number and the complexity of new products or solutions never ceases to increase. However, the design process is still the same, and even if

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

computerised tools, such as CAD/CAM applications, have helped designers to achieve their goals, the process can be itemised as follows (Ulrich and Eppinger, 1995): 1 Problem recognition: . need for creating a new product; . modifying an existing product to improve its features. 2 Investigation, by defining: . the function of the product; . its appearance; . the material(s) required; . its construction; and . the related safety issues. 3 Development of alternative solutions by producing a number of solutions, using sketches without being too detailed or accurate. The more ideas found, the more likely to end up with a good solution. 4 Choosing a solution, by considering: . the ability of the company to manufacture the product; . the material(s) availability; . the time needed to build each solution; . the cost of each solution. 5 Producing detailed design layout: . the overall dimensions of the product; . the detailed dimensions; . the material to be used; . how the product will be made; . if needed, what finish will be required. 6 Elaborating models and prototypes: . make the idea easier to understand; . first working version of the designer’s solution; . help in comparing competing solutions. 7 Testing and evaluating: . Does the product work? . Does it meet the initial need or requirement? . Will modifications improve the solution? 8 Manufacturing. The design process model above has been established by making the assumption that the product is marketable and that it can be manufactured. It is now going to be demonstrated that with the introduction and the availability of new technologies such as the Internet, the way new products are introduced has changed product designers’ work processes. Designers not only need to know what the product is, but they also need a proactive view of forthcoming changes due to the market evolution in order to minimise changes redundancy. Now changes come from both the demand side, driven by such needs as improving price/performance/ functionality or creating features and

variants for a market opportunity, and the supply side, driven by components shortage, quality issues, or part going end-of-life. In fact, the rest of the ‘‘old’’ design process has also been modified, and with the usage of other technologies such as virtual reality (VR) (Summers and Butler, 1999), each step within the process could respectively be re-written as follows: 1 Problem recognition: . need for creating more products better and faster; . need for creating product variants to reach a bigger range of customers; . changes now come from both the demand and the supply sides. 2 Investigation, by defining: . the functionality of the product; . the appearance, which could now be required or defined on the Web by the future users; . the material(s), which could also now be required or defined on the Web by the future users; . the construction, with a direct link to a virtual factory, the methods required to manufacture the product could be agreed or decided without the intervention of a third-party; . the related safety issues 3 Development of alternative solutions: . with the availability of CAD/CAM tools, designers are now able to sketch a far bigger number of alternative designs under an electronic format; . it is also possible to reuse previous solutions (partially or entirely) by just reloading the previous electronic data. 4 Choosing a solution, by considering: . within the virtual factory environment, designers are now able to plan in advance the way the product is going to be manufactured; . the material(s) could now be retrieved with an up-to-date dedicated database, or ordered in advance if needed; . the time needed to build each solution could be directly known by using e-manufacturing simulators, allowing also testing several manufacturing solutions; . the cost of each solution could now be obtained immediately with a better communication with the purchasing and sales departments over the Web. 5 Elaborating models and prototypes: . the CAD/CAM model could be easily converted into a highly portable VR model, which could also be published on the Web, for customers to review it;

[ 329 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

the VR product could also be implemented in its virtual working environment to appreciate all its functionalities; . the VR models could be sent over the Web and allow a much quicker and handier access in comparing competing designs; . another operation could be directly added to this stage, which is the edition of VR procedures (using several medias such as VR models or video clips) that could be dedicated to the customers or even for a maintenance purpose. 6 Testing and evaluating: . this last stage could now be merged with the previous step, as VR and Internet allow delivering real-time interactive 3D graphics in unprecedented quality and speed, as well as fully animated 3D scenes, ranging from simple sequences of motions all the way to complete interactive 3D presentations. .

These improvements in the design agility would possibly be achieved thanks to a network architecture allowing all the different departments within a company to be linked together, but also feedback available from the customers over the Internet. In addition, the design team would need to be also in contact with the sales and the purchasing departments by sharing a common database over the Web, which could radically change their working methods (Anderson and Pine, 1997). With information such as costs and selling prices, designers could compare, with great accuracy and almost in real-time, the benefits and costs of various designs, while these designs were still on ‘‘the drawing board’’. Added to this, such granularity of cost data, if accompanied by knowledge of customer desires, could permit designers to design products for eversmaller better-defined market segments. The design team could be represented in Figure 1, as opposed to the ‘‘old’’ linear layout. In fact, the concept ‘‘design agility’’ could easily be associated to the concept ‘‘design for supply chain’’ or ‘‘design for manufacture’’, as it describes an approach to design where supply chain issues are given strong consideration during the design phase. Companies and partners collaborate over their intranets or over the Internet by sharing design information, in particular: 1 Engineering data: . material information; . product drawings; . CAD geometry/models.

[ 330 ]

2 VR models: . product as a VR object, either on its own, or integrated within its future working environment/assembly; . realistic rendering of the finished VR prototype. Any partner who fabricates, assembles, tests, and/or procures components for the product, participates in review of the product design and configuration. Making information flow quickly and accurately between all team members is a fundamental requirement to make the whole process go forward. Problems of delivering product information across the enterprise drive manufacturing costs way up. The agility to communicate product data and engineering changes instantly to employees, suppliers, and customers is critical in modern manufacturing industry.

Manufacturing responsiveness It has been demonstrated previously that the Internet could change the well-established rules of design teams, by allowing real-time connection to the rest of the company and therefore by decreasing the new product introduction (NPI) process in matters of time and costs. In the same way, advances in technology and growing customers’ expectations for feature-rich products delivered quickly and at competitive prices could modify manufacturing processes by the use of the Internet to improve global responsiveness. These technologies could become inevitable in areas such as: 1 Collaboration: . by increasing responsiveness from both the manufacturing and the design side; . by increasing information availability; . by enhancing mass customisation. 2 Communication: . by reducing transfer times; . by reducing data navigation iterations. 3 Content management: . by providing consistent and accurate product definition; . by providing helpful product viewing (2D drawings, as well as 3D virtual models). Figure 2 shows a simple input-output model of manufacturing system. The inputs include: . raw materials; . purchased components; . employee’s efforts; . energy; and . equipment.

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

The outputs include finished goods and waste. Manufacturing costs is the sum of all the expenditures for the inputs of the system and for disposal of the wastes produced by the system. Because of the competition, there is no hiding place for inefficiency, high production costs, or outmoded products. This is the reason why ‘‘manufacturing responsiveness’’ has to be closely associated with ‘‘product design’’ by improving their communication model via Internet/intranet or extranet connection. Manufacturing cost is a key determinant for the business success of a product. Economically, a successful design is therefore about ensuring high product quality while minimising manufacturing cost. Some component parts may be costly because the designers do not understand the following issues of the production process:

Figure 1 Design team with a new role within the company’s architecture

Figure 2 A simple input-output model of manufacturing system

1 The capabilities: . Is the product ‘‘manufacturable’’? . What kinds of machines are available? . Is the needed labour available? 2 The costs drivers: . the cost of any purchased materials; . the cost of assembly labour; . equipment occupying time needed. 3 The constraints of the product: . Are the tolerances and accuracy possible to reach? . Could the product be redesigned to avoid costly assembly operations? It is clear that the manufacturing department has a deep knowledge of all the previous issues, which is probably not the same case for the design team. Therefore, being able to share on a real-time basis all the information of a product – design and manufacturing data – would improve the feedback between the two departments, and so improve the overall process responsiveness. It has previously been demonstrated how manufacturing responsiveness could be applied to linking the design and the manufacturing departments. Furthermore, there are four fundamental approaches that a manufacturer can adopt to become more responsive to the Internet-driven changes, and they are listed as follows (Kidd, 1994): 1 Manufacturers must change the way they plan their inventories, by building in as much flexibility as possible in terms of: . inventory of production goods; . scheduling of assembly processes; . warehousing of finished goods; . delivering products to the marketplace. 2 Manufacturers have to postpone final assembly by: . developing a product line that can be custom-configured rapidly; . pre-assembling core products up to the point of customisation; . minimising the differences between individual items; . having the assembly processes readily adaptable to short-notice changes in demand. 3 Enhancing the flexibility of the actual manufacturing plant by: . implementing a pseudo build-to-order environment; . reducing cycle times; . increasing the production process frequency; . aligning production lines more closely to actual customer sales; . producing an inventory level of finished stock in accordance with the actual daily demand.

[ 331 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web

Figure 3 Four steps for JDBC/ODBC integration

Integrated Manufacturing Systems 13/5 [2002] 328–339

4 Implementing a true build-to-order environment by: . producing small quantities on highly flexible and configurable lines; . switching equipment designed for low variability to highly flexible equipment for reducing batch sizes. In the long run, products are going to be made where it is the cheapest and most efficient to make them, and every player in the market will have to compete with the most efficient manufacturer. The concept ‘‘manufacturing responsiveness’’ describes the approach of implementing a manufacturing model coupling virtual manufacturing with real-time integration of product design and operational processes, rather than the traditional ‘‘engineer first, manufacture later’’ model (Groover, 1999) by: . establishing instantaneous communication throughout an integrated network via the Internet; . allowing parts procurement while engineers are still designing product, revising part orders as new engineering changes order is created;

Figure 4 GUI developed in Java

.

achieving phenomenal time to market results; e.g. from product concept to assembled prototype.

Finally, in order to coordinate their manufacturing processes and to aggregate their purchasing processes, leading manufacturers could use Internet technology to link their technical data for their teams to operate faster and smarter than the old vertically integrated manufacturers of the past. This has resulted in improving responsiveness to changing conditions on the supply or demand sides, and in drastically reducing the time it takes to ramp up production to meet demand. To conclude, it has previously been shown that new technologies, particularly the Internet, would give another aspect and other challenges to the well-established design and manufacturing processes. In order to meet customers’ expectations which are becoming harder and quicker to reach, companies must now adopt new strategies to address these challenges and so remain competitive. These strategies exploit the immediate exchange of information across technical and organisational boundaries to achieve benefits. The end result of these benefits is that companies could now bring their product to market more efficiently while at the same time achieving higher levels of customers’ satisfaction.

Case studies Java/database tolerancing and fitting application Java database connectivity (JDBC) is an API (application programming interface) for linking Java programs to databases (Zukowski, 1998). JDBC is very similar to Microsoft’s open database connectivity (ODBC) standard. JDBC-compliant database applications are not tied to a specific database vendor, and so can be used with Oracle, Sybase, Informix, mySQL, Microsoft SQL, FoxPro, and Microsoft Access, etc. As with ODBC, a vendor-specific driver is used to link JDBC applications to the actual database. JDBC is likely to be critical to industry acceptance of Java as a corporate

[ 332 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

Table I Data retrieval process Task

Code/action

User enters nominal diameter and chooses grade for Housing and Shaft. (Ex. 52 H7 g6) Button ‘‘Display Results’’ pressed

String ... = textfield.getText(); //get input String ... = choicelist.getSelectedItem(); //get item

Desired table is accessed and data are retrieved Data processing and calculations Data displayed in the GUI

String query = ‘‘SELECT * FROM your_table’’; statement = connection.createStatement(); resultSet = statement1.executeQuery( query ); displayResultSet( resultSet ); statement1.close(); First loop is launched to retrieve the column, and then inside it, a second loop is started to get the row, based on a number of loops. When these two conditions are met, the program reads the datum from the cell All retrieved data from the several tables are processed and analysed through several mathematical operations Data are finally formatted and then displayed in their appropriate text fields

client/server development tool. So, in order to show this relationship between Java and a database, this case study is going to describe all the required ingredients to fulfil the goals. The example is based on a very simple Java application/Microsoft Access database (Taylor, 1997) to simply connect, get particular data from a particular table/database, compute, and finally output the desired result. These four steps can be easily represented by the diagram in Figure 3. To demonstrate the process described above, a Java standalone application has been developed and aims to retrieve tolerances data (function of an input diameter) in a Microsoft Access database. The data are then processed and displayed in

Figure 5 The screenshot of a working application

a GUI (graphical user interface) as shown in Figure 4. When the application is running, the user must enter a nominal size/diameter (in mm) and then must select a grade for the desired housing and shaft assembly. So far, the database is still not connected with the Java application, and this will be done only when the user will hit the ‘‘display results’’ button. After performing this action, the database is accessed, as well as all the tables specified in the code, and the data can then be retrieved through several loops executed by the program. Even if this process seems quite simple, it involves, nevertheless, a particular way of retrieving the data. Table I illustrates the data retrieval process in the Java code. Figure 5 shows the screenshot of the working application with all the tasks completed. As described previously, this example is based on a standalone Java application. In order to be able to have the same features but across a network (Internet or intranets), this application must be converted into an applet or a Java server page. The MS Access database will also be converted to a mySQL database, allowing SQL statements and queries and therefore more dedicated to a network implementation.

Java-based simulators for milling/turning operations The Web-based system also features two Java-based simulators for manufacturing operations: 1 milling process; and 2 turning process. Figures 6 and 7 show a screenshot for each of these simulators. Compared with the tolerancing application described above, these two simulators are not related to a database. Instead, all the needed

[ 333 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web

Figure 6 Turning process simulator

Integrated Manufacturing Systems 13/5 [2002] 328–339

Figure 7 Milling process simulator

[ 334 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

values have been entered in arrays within the program, and those values are accessed and computed depending on the desired calculation. The working process of these applets is as follows: . first the user connects to the server by entering (or selecting) the URL of the desired simulator; . the Web page and the embedded applet are then downloaded in the cache memory on the user’s local machine; . when a calculation is processed, only the Java source code in the cache memory is accessed, without any further connection with the server. This process has obviously some advantages, but also some disadvantages including (Ladd et al., 1999): 1 Known advantages: . Java applets are highly customisable and could be applied to any particular manufacturing process; . the embedded Java code is downloaded once and can be run repeatedly; . the processing time of a calculation is really quick, as it is done on the user’s local machine; . no more client/server connections needed; . applets are very small sized application and do not require a lot a network bandwidth. 2 Known disadvantages: . no automatic up-to-date data availability; . if a value within the Java code has to be created or modified, the source code has to be re-compiled, and re-published on the Web server to become accessible again. In fact, these simulators can offer an alternative solution to the Java/database model. They require far less difficult implementation on the server side, while still having all the advantages related to Java applets. They could run ideally within e-manufacturing environment where most of the calculations values/ formulas are well known and defined, providing then a solid network-based application available on any computing platform. Finally, the further development related to these two applets is to link them with a real-time 3D environment, allowing on-line information/data processing related to on-line simulations. This would obviously give a far better understanding of the processes concerned, as well as perhaps introducing on-line machine

programming or remote manufacturing management.

Further discussion on the implementation Java programming Sun Microsystems coined Java as: A simple, object-oriented, network-savvy, interpreted, robust, secure, architecture neutral, portable, high-performance, multithreaded, dynamic language.

Java is a technology enabling easily building distributed applications which are executed by multiple computers across a network. The state of the art in network programming, Java has expanded the Internet’s role from an arena for communications to a network on which fully-fledged applications run. Its breakthrough technology allows businesses to deploy full-scale transaction services that deliver real-time, interactive information over the Internet. The reasons so much attention has been paid to Java are summarised as follows (Deitel and Deitel, 1999): . allow developers to write robust and reliable programs; . to build applications on almost any platform and run those applications on any other supported platforms without recompiling the code; and . to distribute applications over an untrusted network in a trusted fashion. In particular, Java programs can be embedded into Web documents, turning static pages into applications that run on the user’s computer. No longer is online documentation limited to articles, like a printed book. With Java, the documentation can include simulations, working models, and even specialised tools. This means Java has the potential to change the function of the Internet, much as the Web has changed the way people access the Internet. In other words, not only will the network provide information, it will also serve as an operating system. Moreover, Java allows object-oriented programming (OOP) for professional applications development (Budd, 1998). OOP is a way to write software that is reusable, extensible, and maintainable. Java is an object-oriented language, and in this way it appears as being very suitable for SME’s projects as well as for companies at a much bigger scale. In fact, the Java-OOP structure becomes a redoubtable weapon when software subsystems are reused from one project to the next. Those projects could be

[ 335 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

associated with the idea of programming in the small or in the large. The first concept is established with the following attributes: . Code is developed by a single programmer, or perhaps by a very small collection of programmers. A single individual can understand all aspects of a project, from top to bottom, beginning to end. . The major problem in the application development process is the design and development of algorithms for dealing with the problem at hand. Programming in the large on the other hand, characterises software/application projects with the following features:

Figure 8 Three-tiered architecture

.

.

The software system is developed by a large team of programmers. Individuals involved in the specification or design of the system may differ from those involved in the coding of individual components, who may differ as well from those involved in the integration of various components in the final product. No single individual can be considered responsible for the entire project or even necessarily understand all aspects of the project. The major problem in the software development process is the management of details and the communication of information between diverse portions of the project.

Finally, OOP languages, like Java, support encapsulation, inheritance, and polymorphism. Encapsulation separates the interface from the implementation by hiding data within the object and making those data accessible via methods (or functions). Subobjects inherit the methods and variables of their parent objects, making it easy to reuse the functionality in the parent object. Polymorphism allows creating generic, reusable code that will work with a wide range of different objects. Because Java supports all these features, Java code is likely to be integrated in an e-manufacturing environment, particularly to facilitate collaborative work.

Client/server architecture

Figure 9 Exploded view post animation

[ 336 ]

In today’s computing world, client/server technology has found a place in most corporations and seems unavoidable in an e-manufacturing application. The biggest benefit of this technology is that the processing load is shared between the client and the server (Minasi et al., 2000). A client can be any program (GUI application, Telnet, Java-applet, and so on) that requests services from a server application. Server applications are extensively used in an e-manufacturing environment, as they include database server applications, application servers, communication (FTP, Telnet, Web) servers, and more. In these shared applications, there is always an important factor, which is how to program applets and applications that can perform multiple activities in parallel. Java includes capabilities to enable multithreaded applications (i.e. applications that can specify that multiple activities are to occur in parallel). This makes Java better prepared to deal with the more sophisticated multimedia, network-based, multiprocessor-based applications that will be introduced in the new millennium (Orfali et al., 1994).

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

If a single-threaded application has one thread of execution running at all times, it means that this application can do only one task at a time, which is probably the last wanted result when talking about client/server applications (Cha et al., 1999). On the other hand, a multithreaded application can have several threads of execution running independently and, more important, simultaneously. Multithreading is commonly used to perform the following functions: . Maintaining user interface responsiveness. For instance, this will keep the user interface running at nearly full speed while another task is in progress. . Waiting for wake-up call. This will allow to have a routine waiting for a specified time when placed in a sleeping thread. . Simple multitasking. Multithreading allows to run multiple instances of a process quite easily. . Building multi-user applications. Server applications wait for requests to arrive and then establish conversations with the requester (client(s)). . Multiprocessing. By breaking an application into different threads, it is possible to make the best use of processing power. Every item in this list applies to Internet and embedded-systems applications. Another architecture, called ‘‘three-tiered application’’ can add an additional layer to the standard client/server model and being developed. The processing of the application

Figure 10 Enabling transparency on elements

is split between two portions of code. One portion processes on the client and the other on the server. The third remaining component of the application could be for example the database server as shown in Figure 8. With three-tiered applications, a client-side application performs a portion of the processing. An additional server-side portion of the application operates on the server and communicates with the application partition that resides on the client. Another approach can be considered as a four-tiered approach. In the case of firewall machines, there is a restriction on making a direct socket connection from the applet to the application server. In this case the HTTP/WWW server is used as an intermediary between the application server. The application server interacts with the database, returns data to the HTTP server, which in turn returns the data to the client machine. This approach provides an additional layer of security that many sites require for their Internet access.

Interactivity via ‘‘cult3d’’ technology Cult3D technology is a software-only, multi-platform rendering engine that delivers real-time interactive 3D graphics of unprecedented quality and speed. Cult3D allows fully animated 3D scenes, ranging from simple sequences of motions all the way to complete interactive 3D presentations, without any special hardware support. The Cult3D viewer plug-in is optimised to run on a wide range of computers and configurations, including low-end systems with low-bandwidth connections – such as a first-generation Pentium PCs or Power Macintosh computers – without sacrificing responsiveness or quality. With built-in compression and streaming capabilities, the Cult3D file format is small and efficient and results in faster downloads and progressive viewing. Cult3D’s event-driven architecture gives you the ability to set up all the criteria for what your object can or can not do depending on what the end-user does or does not do within your defined scene. In addition, you can use the traditional timeline animation for a more complex animation sequence. By supporting both of these modes, Cult3D gives you the ability to create almost any presentation, while still keeping the authoring process simple. Cult3D can also include Java code to control the presentation and that way enables complex interactivity and animations such as in a game or build a calculator in Cult3D that actually works like in real life. Using Cult3D’s scripting capabilities advanced presentations can be created, enabling the Cult3D object to

[ 337 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

Figure 11 Fluid animation ‘‘in’’

communicate with the Web page or a database. Whether the final goal is educational, commercial, or technical, most compelling Cult3D worlds have certain characteristics in common: . A Cult3D world is immersive. The user enters the 3D scene on the computer screen and explores it as s/he would explore part of the real world. Each person can chart a different course through this world.

.

.

.

The user, not the computer, controls the experience. The local browser allows the user to explore the Cult3D scene/world in any way s/he decides. A Cult3D world is highly interactive. Objects in the world can respond to one another and the external events caused by the user. A Cult3D world blends 2D and 3D objects, animation, and multimedia effects into a single medium.

Figures 9-12 show some screen-shots representing the virtual assembly of a bearing-housing unit. Because of the advantages and the features described above, Cult3D appears to be a very suitable tool for sharing 3D models across the WWW, in conjunction with e-manufacturing concepts. The further developments in this area will be focused on linking such models with manufacturing simulators in order to link a process to its 3D representation in a highly interactive way.

Concluding remarks

Figure 12 Fluid animation ‘‘out’’

This paper presents an approach to implementing design agility and manufacturing responsiveness by using emerging techniques available today, such as virtual reality and the Internet. The approach explores various changes that will occur in the design and manufacturing areas with the integration of these new technologies. Application examples are then described including Java-based simulation, Java/database integration, and real-time 3D visualisation, etc. Further development is going to focus on enhancing the users’ interactivity in the 3D area so as to allow real-time simulations operating in a parametric format over the Internet.

References Anderson, D.M. and Pine II, B.J. (1997), Agile Product Development for Mass Customisation, Irwin Professional Publishing, Chicago, IL. Budd, T. (1998), Understanding Object-oriented Programming with Java, Addison-Wesley Longman, Reading, MA. Cha, J., Lee, S. and Jeoun, H. (1999), ‘‘A remote home automation system based on client/server architecture of Internet and VRML with Java’’, Proceedings of the 1999 ASME Design Engineering Technical Conferences, Las Vegas, NV, 12-15 September. Cheng, K., Pan, P.Y. and Harrison, D.K. (2000), ‘‘Internet as a tool with application to agile manufacturing: a Web-based engineering approach and its implementation’’,

[ 338 ]

J. Toussaint and K. Cheng Design agility and manufacturing responsiveness on the Web Integrated Manufacturing Systems 13/5 [2002] 328–339

International Journal of Production Research, Vol. 38 No. 12, pp. 2743-59. Deitel, H.M. and Deitel, P.J. (1999), Java: How to Program, Prentice-Hall, Englewood Cliffs, NJ. Groover, M.P. (1999), Fundamentals of Modern Manufacturing, John Wiley & Sons, New York, NY. Kidd, P.T. (1994), Agile Manufacturing, Addison-Wesley Publishing, Reading, MA. Ladd, E. et al. (1999), Programmation Internet, HTML 4, XML et Java 2, CampusPress, Paris. Minasi, M., Anderson, C., Smith, B. and Toombs, D. (2000), Mastering Windows 2000 Server, Sybex Corporation, San Francisco, CA.

Orfali, R., Harkey, D. and Edwards, J. (1994), Essential Client/Server Survival Guide, John Wiley & Sons, New York, NY. Summers, J. and Butler, A. (1999), ‘‘Development of a feature based design system using virtual reality,’’ Proceedings of the 1999 ASME Design Engineering Technical Conferences, Las Vegas, NV, 12-15 September. Taylor, A. (1997), JDBC: Developer’s Resource, Informix Press, White Plains, NY. Ulrich, K.T. and Eppinger, S.D. (1995), Product Design and Development, McGraw-Hill, Maidenhead. Zukowski, J. (1998), Mastering Java 1.2, Sybex Corporation, San Francisco, CA.

[ 339 ]

The Internet-based virtual machining system using CORBA Sang-Hoon Kong Automatic Control Research Center, Seoul National University, Seoul, Korea Jaehong Park Management Consulting Services, PriceWaterhouseCoopers, Samsung-dong, Korea Young-Geun Han School of Industrial and Systems Engineering, Myongji University, Yongin, Korea Gibom Kim Department of Manufacturing Automation, Seoul National University of Technology, Seoul, Korea Kyo-Il Lee

Keywords Internet, Manufacturing, Machining, Computer-integrated manufacturing

School of Mechanical and Aerospace Engineering, Seoul National University, Seoul, Korea

Abstract This paper presents an Internetbased virtual machining system which applies virtual manufacturing technology to machining processes of CNC machining center. The system is implemented to execute digital machining and verification, to transmit the NC code data to related machining centers after confirming the properness of virtual machining, and to manipulate the machine through the Internet. Also, this research proposes a basic structure that can monitor the status of machines via the Internet. By applying simulation techniques for machining processes, a simple manipulation and monitoring system of machining centers is realized. This entire system is constructed by adopting the latest information technology such as object-oriented method, middleware, Internet programming, and client-server structure.

Received March 2001 Revised February 2002 Accepted February 2002

Integrated Manufacturing Systems 13/5 [2002] 340–344 # MCB UP Limited [ISSN 0957-6061] [DOI 10.1108/09576060210429793]

[ 340 ]

Introduction As manufacturing systems are globalized and become worldwide, rapid development and production are essentially requisite for competitiveness. A new paradigm is necessary for shortening the time and expenses related to planning, development and manufacturing. Lately, the application of virtual manufacturing system has become usual because simulation of design, experiment and analysis is performed virtually in computer prior to real production. It can help developers, suppliers and customers to accomplish interactively the entire product life cycle, which consists of design, production planning, control and manufacturing processes. Virtual verification of design, experiment, verification, test and system operation also are performed under virtual manufacturing system environment. It can support accurate and prompt human decision making. Tarn et al. developed the remote controlled Puma robot that is manipulated via the Internet (Fitzpatrick, 1999). In this system, the robot is controlled with a joystick in a remote place, a special algorithm is adopted to solve the time-delay problem. This solution is a very cutting-edge technology, especially in an uncertain environment because there is a lot of traffic on the Internet and users often experience network failure or The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0957-6061.htm

data loss. Kimura et al. used VRML and Java language to monitor the three-axis milling machine controlled by PC-based controller (Goncharenko et al., 1998). VRML language is used to make the machining scene appear in a remote user’s monitor and Java language is applied to animate the VRML scene. CCD camera is also used to show the real machining process. This research shows the capacity of VRML and Java to be used in remote monitoring and controlling. Leong et al. developed a remote monitoring system of rapid prototyping machine (Luo et al., 1999). Users connect to homepage and upload STL file for sample product, then Web server creates appropriate support and tool path. This system shows the machining process using a CCD camera because of RP machine’s process feature. The RP machine is controlled by a PC-based controller and application programming interface provided with a controller is used to control the machine. In this paper, an an Internet-based virtual machining system is developed. It applies virtual manufacturing technology to the machining process. The developed system is composed of NC file analysis module, virtual machining simulation module, machine monitoring module and machine control module. This system is made in client-server structure and operates in Internet environment. To utilize machining or production applications, each module is integrated with CORBA which provides platform and language independence properties for systems (Hoque, 1999). CORBA also enables developers to expand and to manage with ease. The virtual machining system helps engineers to produce virtual product and to verify manufacturability via the Internet. It also enables users to transmit

Sang-Hoon Kong, Jaehong Park, Young-Geun Han, Gibom Kim and Kyo-Il Lee The Internet-based virtual machining system using CORBA

manufacturing data to machine and to

Integrated Manufacturing Systems 13/5 [2002] 340–344

Internet-based virtual machining system

control manufacturing resources in a remote place.

An Internet-based machining system serves cutting simulation of NC code, verification of machining result and machine controlling function. CORBA (common object request broker architecture) is introduced to integrate distributed computer resources and VRML language is used to make scene of machining process.

The developed system is composed of a virtual machining module and remote control module. In virtual machining module, users transfer NC code file to simulation server and then simulation server analyzes NC code file and performs machining simulation. Remote control module downloads NC code file into machining center and performs real machining process. It also serves real-time monitoring services. Figure 1 shows the architecture of the developed system. Dashed polygon is a computer and shaded box is implemented program. Client computer has three Java applets called NC file manager, VRML manager and machine manager. Simulation

Figure 1 System architecture

Figure 2 Three-tier in virtual machining system

[ 341 ]

Sang-Hoon Kong, Jaehong Park, Young-Geun Han, Gibom Kim and Kyo-Il Lee The Internet-based virtual machining system using CORBA

server analyzes NC code and performs

Integrated Manufacturing Systems 13/5 [2002] 340–344

there is any other machine to be added to the

simulation for NC code verification. Machine server is connected to a machining center. It interfaces between the machine and client. If system, only the server program is coded and linked to ORB. So this structure provides convenience for the addition or exclusion of production resources. Figure 2 shows the system from the viewpoint of three-tier architecture. Between client side and machine side, application

Figure 4 Tool path simulation

[ 342 ]

Figure 3 represents implemented system which is composed of a virtual machining part and a remote control and monitoring part. In the virtual machining part are simulation server and two client applets, which are NC file server and VRML server. Remote control and monitoring part consists of machining center, machine server and machine manager applet. The machining center is controlled by a PC-based controller installed on the machine server computer and the controller is connected to machine server software by Win32 API.

process like simulation server or machine

Virtual machining

server offers expansion characteristic and

The user connects to the virtual machining system and transmits NC file. Because the simulation server needs more information like tool diameter, height of material, etc. it requires the data of user. After being provided with the necessary information, server analyzes the NC file and creates several VRML files to represent analysis results. Analysis results are stored in database and translated to client applet that is NC file manager. VRML files are stored in the simulation server system. Results stored in database are used by VRML manager applet that shows machining results to user in tool path style or solid style. VRML manager applet simulates tool path and detects collisions between tool and material. To show clients tool path, the applet receives VRML string with CORBA and represents the scene with Java EAI methodology. IndexdLineSet node in VRML is used to show tool path in line style and ElevationGrid node which is similar to z-map method which is utilized to display tool path in solid style. Simulation server calculates the value of each node. Figure 4 represents tool path simulation scene and Figure 5 represents collision

program efficiency.

Figure 3 Implemented system

Implementation

Sang-Hoon Kong, Jaehong Park, Young-Geun Han, Gibom Kim and Kyo-Il Lee The Internet-based virtual machining system using CORBA

detection simulation.

Integrated Manufacturing Systems 13/5 [2002] 340–344

center and control manufacturing resources.

Monitoring and control of machine After simulating the machining process, users can download NC file to the machining If a user is not permitted to control resource, he can only monitor manufacturing processes executed by another user.

Machine monitoring and control part also recommends appropriate machining centers for analyzed NC code. Machine is recommended by size of parts, size of table, tool diameter, and much other information that can be extracted from database which has manufacturing resource information and NC code analysis results. In this case study, we selected a small machining center, Benchman 4000

Figure 5 Collision detection

Figure 6 Real-time monitoring of machining process

[ 343 ]

Sang-Hoon Kong, Jaehong Park, Young-Geun Han, Gibom Kim and Kyo-Il Lee The Internet-based virtual machining system using CORBA Integrated Manufacturing Systems 13/5 [2002] 340–344

manufactured by Light Machine, Ltd, for the machining. Because the control program of this machine is Windows-based application, it provides machining information for the machine server that is connected to the controller with message hooking method (see Figure 6). The controller of Benchman 4000 machining center is installed in ISA slot and the control program is installed in the machine control computer system. Despite the controller system being PC-NC style, machining center control is very restrictive because the API of control program or library for interface is not provided. Therefore, we control the machining center by calling message function that transmits Windows message to the machining center control program. Calling message function uses Win32 API to catch Windows message and is called through CORBA. In this method, the architecture of the remote control system is implemented in an Internet environment. More sophisticated control is implemented by using openarchitecture control program.

Suggestion for machining center monitoring In this system, real-time monitoring is impossible because the controller of Benchman 4000 is closed. So we figure out a way to monitor the status of the machining center. After reading the file which has the location of the tool, the server sends the location to the client. If open control program and API for programming are supplied, other information about the machine besides tool location and direction is available with the architecture provided in this paper. In monitoring system, the data changed in the machine should be presented to the client without any request of user. A unique method for the client to supply server with the interface that is called by server is necessary. CORBA support this function, Callback service or Event Service. Callback

[ 344 ]

service activates CORBA objects available in client side and sends reference of activated objects to server. Server can call the function in client side in this way.

Conclusion In this paper, virtual manufacturing and verification for machining center are performed in an Internet environment. Virtual machining process is calculated in server system with parameters provided by users and sent back to users via the Internet. Virtual reality technology is used to show the machining process to users. Internet-based virtual machining system executes the cutting operation simulation and enables users to verify machining. Also users can monitor and control machining centers located in various places with just a VRML-enabled Web browser. To provide developers with flexibility, expansibility and simpleness, many computer resources are connected in three-tier architecture with the use of CORBA.

References Fitzpatrick, T. (1999), ‘‘Live remote control of a robot via the Internet’’, IEEE Robotics & Automation Magazine, September, pp. 7-8. Goncharenko, I., Kimura, F. and Mori, K. (1998), ‘‘Web-based user interface for machine tool monitoring and control’’, Proceedings of the 31st CIRP International Seminar on Manufacturing Systems, May, pp. 448-53. Hoque, R. (1999), CORBA 3, IDG Books. Luo, R., Lee, W.Z., Chou, J.H. and Leong, H.T. (1999), ‘‘Tele-control of rapid prototyping machine via Internet for automated telemanufacturing’’, Proceedings of the 1999 IEEE International Conference on Robotics and Automation, May, pp. 2203-8.

Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Jason S.K. Lau Department of Industrial and Manufacturing Systems Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong George Q. Huang Department of Industrial and Manufacturing Systems Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong K.L. Mak Department of Industrial and Manufacturing Systems Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong

Keywords Supply chain, Performance, Information, Inventory control

Abstract Information sharing and coordination between buyer and vendor have been considered as useful strategies to improve supply chain performance. The debate is about what information to share and how to share most costeffectively to maximize the mutual benefits of the supply chain as a whole and the individual business players. Proposes a systematic framework for investigating the impacts of sharing production information on the supply chain dynamic performance. This framework supports supply chain researches to study impacts of information sharing under various scenarios. Examines, under the framework, an inventory allocation problem in an arborescent distribution supply chain with two distribution channels competing for the same source of supply. Finds that the levels of benefits by sharing information vary with different players involved in the supply chain. Suggests some guidelines to balance the benefits in a supply chain in order to motivate information sharing.

Received August 2001 Revised February 2002 Accepted February 2002

Integrated Manufacturing Systems 13/5 [2002] 345–358 # MCB UP Limited [ISSN 0957-6061] [DOI 10.1108/09576060210429801

Introduction Supply chain dynamics has been studied for more than three decades. Since Forrester (1961) discovered the fluctuation and amplification of demand from the downstream to upstream of the supply chain, there has been a lot of literature analysing this phenomenon (e.g. Towill, 1991; Wikner et al., 1991; Towill et al., 1992; Wu and Meixell, 1998; Helo, 2000). This effect can be readily illustrated in the well-known ‘‘beer game’’ in which one can observe the amplification of demand signal and fluctuation of inventory level along a supply chain which consists of customer, retailer, wholesaler, distributor and factory (Sterman, 1989; Simchi-Levi et al., 2000). This effect is obviously undesirable as it exacerbates the supply chain costs (e.g. stock holding, backlog, late delivery, under/over resource utilization etc.). The source of such fluctuation and amplification of order and inventory is mainly due to the lack of sharing of production information between enterprises in the supply chain. (e.g. Lee et al., 1997; Simchi-Levi et al., 2000). These factors lead to distortion of actual demand information and cause unnecessary wastes. Lee et al. (1997) have studied this phenomenon extensively and termed it as ‘‘bullwhip effect’’. From the studies of bullwhip effect (e.g. Lee et al., 1997; Metters, 1997), one of the remedies is to share information along the supply chain. It has been reported that the benefit of information sharing is significant, especially in reducing the bullwhip effect (e.g. Lee et al., 1997; Cachon and Fisher, 2000; Lee at al., 2000) The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0957-6061.htm

and supply chain costs (e.g. Swaminathan et al., 1997; Gavirneni et al., 1996; Tan, 1999). By using the shared information, each supply chain entity can make better decisions on ordering, capacity allocation and production/material planning so that the supply chain dynamics can be optimized. Information sharing, however, may not be beneficial to some supply chain entities due to high adoption cost of joining the interorganizational information system, unreliable and imprecise information (e.g. Swaminathan et al., 1997; Cohen, 2000), and different operational condition of each firm (e.g. Dong and Xu, 2001). Zhao and Xie (2002) have recently found that, by using simulation study, sharing information may hurt some supply chain members under most conditions. National Research Council (2000) reported that there are some barriers (e.g. expensive technology investment, personnel training, lack of mutual trust etc.) which hinder small and medium sized enterprises (SMEs). Singer (1999) also reported that, based on an industrial case, sharing information may lead to loss of business. It is, therefore, necessary to investigate the benefits of information sharing that can be gained by each firm and the whole supply chain before implementing. Regarding technology of implementing information sharing, electronic data interchange (EDI) has been employed as a major tool of information sharing for many years (e.g. Davis and O’Sullivan, 1998; Strader et al., 1998; Lee at al., 2000; Bhatt, 2001; Furst and Schmidt, 2001; Warkentin et al., 2001). As Internet and e-commerce technology continue to evolve, there has been much literature studying how such technology can improve supply chain performance, especially on information

[ 345 ]

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

sharing (e.g. Davis and O’Sullivan, 1998; Strader et al., 1998; Conway, 2000; Graham and Hardaker, 2000; Croom, 2001; Kehoe and Boughton, 2001; Warkentin et al., 2001). Given a wide spectrum of information technologies (e.g. Internet, extensible mark-up language (XML), common object request broker architecture (CORBA)), it is unclear which technology is most suitable for enabling the share of production information in the supply chain. The above brief review of the related literature has highlighted the necessity and significance of sharing production information in order to remedy the so-called bullwhip effect and its associated impacts. However, the literature generally does not address adequately what exactly production information the enterprises in the supply chain should share, and to what extent they should share. Having recognized these limitations in the literature, a research project has been initiated to address these research questions. The overall objective of this project is to develop a framework for investigating the impacts of sharing production information on the supply chain dynamic performance. The proposed framework will be used to carry out a series of simulations from several perspectives such as resource allocation, material requirement planning and rescheduling. Techniques of experimental design will be applied to analyse the results to derive insightful findings. Guidelines will be established to assist industrialists to set up efficient and effective supply chain management systems. The research will take full advantage of the emerging information and telecommunication technologies, and take into consideration industrial needs. This paper reports on a Web-based implementation of a conceptual framework proposed from this research project (Huang et al., 2001) and some early findings from initial investigations.

Web-based simulation portal A conceptual systematic framework has been developed from our early research for investigating the impacts of sharing production information on the supply chain dynamics (Huang et al., 2001). Based on this proposed framework, further development has been accomplished to establish a simulation portal on the World Wide Web (Web or WWW). The resulting prototype simulation portal is diagrammatically shown in Figure 1.

[ 346 ]

Generally speaking, this simulation portal provides a suite of online facilities in a step-by-step fashion for analysts to design and conduct simulation experiments and afterwards to analyse the results and draw implications. The key steps of using the simulation portal follow those of the conceptual framework. They will be explained in detail in the subsequent section one by one. The purposes are threefold: 1 to serve as the test-bed for implementing and executing the experimental simulations; 2 to provide a Web-based test-bed or portal open to the entire research community to share the experimental data and results; and 3 to provide some practical insights to industrial practitioners regarding how the production information can be made shareable between enterprises in the supply chain with the support of the Internet and Web technology. The development and implementation of a Web site for just carrying out the experimental simulations specifically designed in this project is relatively straightforward. Web pages are designed to follow the format of worksheets used at different stages. Input and output data are stored into and retrieved from the backend database. A typical three-tiered architecture of Web applications is adopted in an active server pages (ASP) programming environment. Java/VB (visual basic) scripts have been extensively used, in addition to SQL (structured query language) for database programming. However, significant challenges are expected if the site is intended for the second and third purposes. On the one hand, this Web site is not intended to become a Web-based general-purpose simulation system. On the other hand, some degree of customisation should be provided so that the users are able to change the settings of experimental simulations. A set of supply chain structures, typical decision levels, the production information hierarchy, the supply chain dynamic index hierarchy, supply chain dynamic models will be parametrically pre-defined and thereafter maintained in the backend database. Facilities will be provided for customizing the parameters. The building up of the model of a subject supply chain structure and the definition of simulation logics in the simulation model are some functions generally available from commercial simulation packages. Therefore, this research would avoid repeating the efforts.

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation

Figure 1 Specify supply chain structure

Integrated Manufacturing Systems 13/5 [2002] 345–358

Instead, emphasis will be placed upon the modeling itself. As a result, the simulation portal would not be able to act in a self-contained fashion. However, investigations are being carried out to devise new methods of modeling the supply chain structures and their dynamics. In this respect, it would become worthwhile to implement these new models as online Web-based systems. In addition, this simulation portal should also reflect the potential methods of exchanging production information between the supply chain players. The concept of agents will be introduced to represent players or echelons in the supply chain. The production information and supply chain dynamic performance measurements will be captured by the properties of the agents. The data exchange and communication between the echelons become the flow of data between the corresponding agents in the form of messages, either directly between them or through a commonly shared area called blackboard. The technologies (e.g. agents, blackboard, message passing, workflow management – dataflow management) used for implementing this simulation portal will provide some insights to the industrial practitioners on how they can take advantage

of the technologies to share production information with each other. Once opened for the research community to share, this portal will prove an invaluable source of resources. Experimental data can be reused in different simulations where items of production information and their value levels are changed, or where a different supply chain structure is used at another decision level. This would maintain a high degree of integrity and consistency among the findings from these simulations by different researchers, or allow the researchers to verify their findings with each other.

Case study The case study to be presented in this section has two purposes. The first purpose is to demonstrate the procedure of applying the simulation portal step by step. The second purpose is of course to investigate the effects of sharing production information on the supply chain dynamics from an inventory allocation perspective. This inventory allocation problem is based on an arborescent distribution supply chain in which all companies are independent of each other. There are two distribution channels in

[ 347 ]

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

this supply chain. Each retailer is supplied by one distributor and both distributors are supplied by a single manufacturer which operates on a make-to-order basis with limited production capacity. The details of this case study will unfold with the progress within the simulation portal from one step to the next, as shown in the following subsections.

Step 1: defining objectives of study The purpose of this study is to analyse the impacts of different levels of information sharing on operating costs of the whole supply chain and each individual company under various supply chain scenarios. Demand and inventory information are considered for sharing among companies in a supply chain. The mechanism of different levels of information sharing will be discussed in Step 4. Operating costs of the supply chain and each company will be defined in Step 5. The results can be used as a guideline for a company which is considering information sharing.

Step 2: selecting/specifying decision level There are basically three levels of decision in a supply chain: strategic, tactical and operational. Huang et al. (2001) discussed the three levels related to this project. The decisions considered in this paper are inventory allocation and replenishment, which belong to operational level.

Step 3: selecting/defining supply chain structure A supply chain structure must be defined for the experimental simulation. At present, only four basic SC structure templates are pre-defined in the database, as shown in Figure 1, and the analyst needs to select one of them. More sophisticated facilities should be provided in this simulation portal for defining more complicated SC structures and models that are more relevant and realistic. A supply chain with two distribution channels is selected in this case study. Each channel consists of one retailer and one distributor. Each retailer, R1 and R2, is supplied by only one distributor, D1 and D2. Both distributors are supplied by a single manufacturer M. This represents a typical distribution supply chain in the real world in which each distribution channel serves a region of the market.

Step 4: specifying production information model (PIM) This step is essential. It defines the factors that exactly will be studied in the experimental simulation, and the modes in

[ 348 ]

which these factors are shared between players. The use of the simulation portal in this step is straightforward, just a matter of filling in forms in relevant Web pages, as shown in Figure 2. The analyst must be clear about the contents being entered into the form.

Step 4.1: selecting information sharing mode Since level of information sharing is the main focus of this paper, it should be clarified first before introducing other production information used in this case study. Four modes of information sharing are identified in this study to represent four levels of sharing, as indicated in Figure 3. The dotted line indicates information flow and the solid line indicates material flow. The first mode (BC) represents the traditional information flow in the supply chain in which each company shares information based on its order (Figure 3a). An order-up-to (s, S) installation stock policy is used by retailers and distributors to replenish their inventory. This policy specifies that when the inventory position IP, which is sum of on-hand and on-order stock, falls below the reorder point s, an order with quantity Q is placed to the supplier in order to raise the inventory position to S. The (s, S) policy of each retailer and distributor is defined by equations (1)-(4). Equation (1) specifies the reorder point for average demand
per period, standard deviation of demand , supply lead time L and safety factor . The optimal value of  is given by equation (3), which is the solution for standard newsvendor problem (see Silver et al., 1998), where k is ordering cost, b is backlog cost and h is holding cost. The optimal value of Q is given by equation (2) which is the standard EOQ solution, as Axsater (1996) suggested that the standard EOQ solution is a good heuristic in a stochastic environment. It is, therefore, convenient to express S as equation (4): pffiffiffiffi s ¼
L þ  L ð1Þ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2
kðb þ hÞ Q¼ bh  ¼

1



b bþh

S ¼ s þ Q:

ð2Þ

 ð3Þ ð4Þ

In the second mode (RD) the retailer shares its demand (i.e. market demand) and inventory information with its distributor (Figure 3b). The retailer places the order according to its own policy while the distributor, by taking

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation

Figure 2 Specify parameters and sharing modes

Integrated Manufacturing Systems 13/5 [2002] 345–358

Figure 3 Four modes of information sharing

the retailer’s demand and inventory information into account, orders from the manufacturer by using a modified (s0 S0 ) policy which is specified by equations (5) and (6) (S0 is obtained as equation (4)). Note that subscript r and d indicate retailer and distributor information respectively:

s0d ¼
r ðLd þ Lr Þ þ r Qd0

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðLd þ Lr Þ

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2
r ðkd þ kr Þðbd þ hd Þ ¼ : bd hd

ð5Þ

ð6Þ

In the third mode (DM) the distributor shares its demand and inventory information with

[ 349 ]

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

the manufacturer (Figure 3c). The manufacturer determines the order size by considering the setup cost of production when the inventory position of distributor falls below its reorder point according to equation (1). The modified order size is given by equation (7) where km is production setup cost: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2
d ðkd þ km Þðbd þ hd Þ 00 Qm ¼ : ð7Þ bd hd The fourth mode (RDM) is the combination of RD and DM modes. It can be regarded as an integration of retailer, distributor and manufacturer. The retailer shares its demand and inventory information to distributor and manufacturer (Figure 3d). The manufacturer uses market demand and inventory information of retailer and distributor to determine the delivery quantity, which is given by equation (8) when the echelon inventory falls below the reorder specified by equation (5): 00 Qm

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2
r ðkd þ km þ kr Þðbd þ br þ hd þ hr Þ ¼ ð8Þ ðbd þ br Þðhd þ hr Þ

Under each sharing mode the manufacturer operates in a make-to-order basis and employs a priority rule heuristic to allocate capacity to distributor order. The distributor orders are sequenced first. Production capacity is then allocated to each order according to the sequence. In this case study order size is used to prioritize the distributor orders. Each unit of capacity is occupied for the specified lead time in order to produce one unit of product.

Step 4.2: selecting production information for impact analysis Five types of information are selected in PIM for further study. They are demand variance, holding cost, backlog cost, ordering cost, production setup cost and production capacity. Demand variance indicates the fluctuation of consumer demand at retailer level. It is well known that demand variance is one of the sources of bullwhip effect. Investigating different levels of demand variance is important to enhance our understanding of the values of information sharing. Holding cost is incurred per unit stock in on-hand inventory and on-order inventory. Backlog cost is incurred for each unit short of inventory per period. This cost only applies for retailer and distributor. Ordering cost is incurred when a company places an order. This cost may include administrative cost of ordering and transportation cost. Studying impacts of cost parameters is important because the performance index in this study is operating cost. Moreover, the cost

[ 350 ]

parameters affect the local optimal policy of each company, as presented in equations (1)-(7). Production setup cost is incurred for each order received by the manufacturer when production starts. Production capacity, CP, means the number of products that can be produced concurrently during the production lead time. It is also an influential factor of supply chain performance. Evaluation of performance of capacitated supply chain by using analytical approach (e.g. Gavirneni et al., 1996; Cachon and Fisher, 2000) is, however, more difficult than simulation approach. Therefore it is appealing to include this factor in this study. Figure 5 shows a summary of PIMs that are being studied by using the portal. The analyst can select other information for in-depth study by clicking the check box beside the information node. After selecting information, the analyst needs to input levels of value for each selected information in Step 7.

Step 4.3: specifying values for other parameters Other parameters (e.g. market demand, transport and production lead time) of each firm can be specified consequently. In addition to mean values, variances of some parameters can also be specified in the portal. The values are summarised in Table I.

Step 5: selecting dynamic performance index (DIM) Different performance indicators or indexes must be specified for each firm as well as the entire supply chain in order to assess the impacts of information sharing. Cost is the main performance index considered in this study. The average operating cost of retailers and distributors includes fixed ordering, inventory holding cost and backlog cost. It is defined by: ! T X ðhIPt þ bBLt þ kmaxðQt ; 0ÞÞ =T: ð9Þ C¼ t¼Ts

IPt is inventory position and BLt is backlog at the end of period t. T is the length of simulation and Ts is the warm-up period. Note that the subscript is omitted as equation (9) applies for both retailers and distributors. The average operating cost of manufacturer consists of production setup cost and process cost. It is defined by: ! T X X Cm ¼ ðkm maxðQdt ; 0Þ þ gQdt Þ =T: ð10Þ t¼Ts

d

Qdt is production size for distributor d at period t and g is process cost per unit. The

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

Table I Mean values and variances used in simulation Information Demand (units) Production lead time (periods) Transportation lead time between manufacturer and distributor (periods) Transportation lead time between distributor and retailer (periods) total supply chain cost is the sum of operating cost of each firm. The analyst can select total cost of each firm and cost component (e.g. backlog cost) for detail analysis (Figure 4).

Standard deviation

10 3 5 2

Specified in Step 7 0 1 1

The following are assumed in this simulation model: . each retailer faces the same demand .

Step 6: establishing simulation model Conceptually, a mathematical model, which represents the operations in a supply chain, is needed to map PIM to DIM. A multi-agent approach is employed to model the supply chain. This multi-agent model provides a practical foundation of developing a distributed decision support system for the whole supply chain. Discussion of multi-agent modeling is out of the scope of this paper. The reader may refer to Swaminathan et al. (1997) and Parunak et al. (1999) for successful application of multiagent modeling in addressing supply chain problems.

Mean

.

distribution; production and transportation setup time do not exist; a setup cost which is independent of production quantity is incurred when a

. .

production is started; production lead time is deterministic; market demand and transportation lead

.

time follow normal distribution; unfulfilled order is backlogged in retailers

.

and distributors; supply of material for production is

.

unlimited; holding cost of material in manufacturing

.

site is negligible; production capacity level is constant.

Figure 4 Select indexes for performance measurement DIM

[ 351 ]

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

Step 7: designing experiments A total of 15 items of production information have been specified in this case study as the control factors (parameters). One factor is assigned to demand variance; 12 factors are assigned to holding, backlog and ordering cost of each retailer and distributor. The remaining two factors are assigned to production setup cost and capacity of manufacturer. All the 15 control factors have two levels of value. For all sharing modes, an orthogonal array L16(215) of the 15 control factors is constructed to find out the impact of these factors under each sharing mode. Advantages of using orthogonal array over full factorial combination to generate experiment set are discussed in Roy (1990). Figure 5 shows a summary of selected PIM and allows the analyst select the type of array. The analyst needs to specify the value of each level. A column number is assigned to control factors according to the sequence of selection. The analyst can change column assignment by moving the selected control factor up and down. The complete orthogonal array is generated as shown in Figure 6.

Step 8: running simulation After confirming the PIM, DIM and experiments, the investigator can run the simulation. Length of each simulation run is

Figure 5 Summary of selected PIM

[ 352 ]

600 periods. The results of first 100 periods are discarded in order to eliminate the effect of transient response. The number of replications of each experiment is 20. The analyst can also specify the simulation parameter of each experiment defined in Step 7, as shown in Figure 7.

Step 9: analysing dynamic performance measurements Once the results from the simulation experiments are obtained, the analyst can study the dynamic behaviour of each firm (as represented by an agent in the agent-based simulation model) by selecting different dynamic characteristics like inventory, demand, order size and backlog. The analyst may gain additional insights (e.g. fluctuation of inventory and order size) which cannot be obtained by using average performance measures (Towill et al., 1992; Parunak et al., 1998). Figure 7 shows the dynamics of inventory of one distributor.

Step 10: analysing impacts of PIM and sharing modes The effects of the sharing mode and the 15 control factors (items of production information) are calculated for each of their level, and presented in Figures 8 and 9

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation

Figure 6 DOE orthogonal array

Integrated Manufacturing Systems 13/5 [2002] 345–358

Figure 7 Dynamics of simulation results of the chosen DIM (costs of the total supply chain and individual agents)

[ 353 ]

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

respectively. Based on these charts, the analyst is able to draw some implications.

Step 10.1: analysing impacts of sharing modes Let us examine Figure 8 first. This chart is obtained by averaging the cost measures (i.e. DIM) of each experiment. It shows the varying impacts of different levels of information sharing mode on the total supply chain cost and the breakdown of cost performance of each agent (firm). It becomes apparent from the chart that the RD mode results in the highest total supply chain cost while the RDM mode results in the lowest. This is perhaps because the distributor orders less from the modified policy, leading to higher backlog in both retailer and distributor. The cost for the distributor is reduced due to reduction in inventory while the cost for retailers is increased due to increase in backlog. The smaller size of distributor orders leads to higher production cost for the manufacturer. In the DM mode, the inventory holding cost for distributors is higher due to a larger order size which is determined by the manufacturer. The manufacturer benefits because of the optimal production size. The cost for retailers in the DM mode is very close to that in the BC mode because sufficient supply from the distributor is guaranteed. But the overall cost is still higher than that in the BC mode. In the RDM mode extra backlog of distributors and retailers introduced by the RD mode is compensated by extra inventory due to larger order size in the DM mode. The overall result is that the manufacturer can produce in optimal batch size, the distributors reduce the inventory level and the retailers can maintain the service level. Hence the cost performance is superior to other modes.

Figure 8 Effects of sharing mode on the total supply chain and individual costs

[ 354 ]

Although the ordering policies are not global optimal as the reorder point and order size are determined ‘‘optimally’’ from each local agent perspective, they can be easily implemented in a real supply chain. The investigator can test different feasible policies, instead of the local optimum one, in the simulation model as control factors to seek ‘‘optimal’’ policy in a realistic sense. Since not all sharing modes (except BC) are feasible in a real situation (e.g. retailer is not willing to share information to manufacturer due to sensitivity of the information or lack of information technology), this system can support investigation of partial information sharing in which only one distribution channel is engaged in information sharing while the other still operates in traditional mode (i.e. BC). Furthermore, it is clear that from the discussion of this case study, not all firms in the supply chain benefit by information sharing. The investigator can introduce some financial incentives in the simulation model in order to balance the benefits among the companies. For example, different levels of price discount together with different levels of information sharing (i.e. sharing modes) may be tested in the model to find out which combination is ‘‘optimal’’ from both a local and global perspective in the supply chain.

Step 10.2: analysing impacts of PIM control factors Next, let us move to discuss the interaction between the 15 PIM control factors and sharing modes. Table II shows the F-value which indicates the significance of the effect of each PIM control factor on the supply chain and individual operating costs. For analysing cost of individual firm, R1, D1 and M are selected. As distribution channels formed by R1, D1 and R2, D2 are symmetric to each other, the results of one channel is very similar to the other.

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation

Figure 9 Pareto charts: Interaction effects of PIM control factors on total cost of supply chain, Retailer 1, Distributor 1 and Manufacturer

Integrated Manufacturing Systems 13/5 [2002] 345–358

Table II ANOVA results (F-value) for the interaction effects of 15 control factors on total cost of supply chain, Retailer 1 (R1 cost), Distributor 1 (D1 cost) and Manufacturer (M cost) under different sharing modes PIM A B C D E F G H I J K L M N O

Supply chain cost Demand variance R1 Backlog cost R1 Holding cost R2 Backlog cost R2 Holding cost R1 Ordering cost D2 Holding cost D1 Backlog cost R2 Ordering cost Setup cost D2 Ordering cost Capacity D1 Holding cost D1 Ordering cost D2 Backlog cost

a

499.38 0.37 0.23 0.54 1.24 1.85b 174.34a 0.01 2.33b 275.63a 1.89b 65.80a 246.97a 1.95b 0.97

R1 cost a

6.39 2.20b 0.45 1.28 1.04 1.53 0.28 0.054 1.23 0.13 0.34 9.25a 6.18a 4.07a 0.01

D1 cost a

35.80 14.32a 109.21a 0.40 0.22 3.55a 1.14 1.05 0.84 53.38a 0.28 16.30a 37.90a 14.21a 0.80

M cost 9.75a 1.53 0.44 0.99 0.64 0.25 19.61a 0.09 1.50 1,132.47a 20.15a 42.88a 20.54a 25.75a 0.42

Notes: a PIM control factors with significance greater than 95 per cent b PIM control factors with significance ranging from 80 per cent to 95 per cent [ 355 ]

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

[ 356 ]

For each cost performance measure three sets of PIM are classified based on their F-value. Class I represents a set of PIM control factors (indicated by superscript ‘‘a’’ in Table II) of which the significance is greater than 95 per cent. PIM control factors (indicated by superscript ‘‘b’’ in Table II) with significance ranging from 80 per cent to 95 per cent are classified in class II. The remaining insignificant factors are classified in class III. The PIM control factors are ranked by their F-values, from the largest to the smallest, in each operating cost measure in Figure 9. The class labels of PIM control factors are also shown below the horizontal axis. As can be seen from Figure 9, only a few factors, which significantly impact on supply chain performance, need to be addressed carefully. It is obvious that demand variance, holding costs, production setup cost and capacity, which belong to class I, are influential on operating costs. These factors are discussed in this section one by one. Class I. Demand variance significantly affects the benefits of sharing information. The stock-out risk and average backlog is higher under higher demand variance. This outweighs the cost saving due to information sharing in RD mode. Under DM and RDM mode, the operating cost of both retailers and distributors is lower that in RD mode. This is because average order size and on-hand inventory are increased under these modes. On the other hand, the manufacturer gains cost saving under RD, DM and RDM mode as the average order size from distributors is increased. Capacity is another influential factor on the benefits of information sharing. The manufacturer’s holding cost is higher under low capacity. As the production lead time of each order is longer, finished goods inventory of each order is higher. The longer order lead time increases the backlog cost and on-order cost of distributors and retailers. This actually reduces and even cancels out the cost saving caused by sharing information. It is obvious that DM mode is most beneficial to the manufacturer under high setup cost. However it is not the case for the distributor as it needs to hold more inventory. This explains the high interaction between setup cost and sharing modes with respect to distributor and manufacturer operating costs. Class II. Cost parameters of the retailer seem to have little interaction with sharing mode with respect to the retailer’s operating cost. The reason is that the local optimal policy, which adjusts the reorder point and

order size based on different cost parameters, employed by the retailer maintains a consistent performance between different sharing modes. The effect of interaction between holding cost of retailer and sharing modes on distributor operating cost is significant. Since higher retailer’s holding cost leads to smaller retailer order size, the average backlog of the distributor is reduced. The cost saving, when the retailer’s holding cost is high, in RD mode is more than that when the retailer’s holding cost is low. In DM mode, however, high on-hand inventory plus reduced shipment, which is due to high retailer’s holding cost, leads to higher distributor’s cost. This situation still occurs in RDM mode as the distributor needs to hold more on-hand inventory. Like the retailer’s holding cost, its backlog cost also shows significant interaction with sharing mode with respect to the distributor’s operating cost. Higher backlog cost increases the retailer’s order size. The distributor then holds fewer inventory in BC and RD mode. The distributor’s cost in DM mode is smaller when the retailer’s backlog cost is higher. This is because larger demand from the retailer compensates for some of the increased distributor’s inventory. The distributor’s cost performances under high and low retailer’s backlog cost in RDM mode are similar, as holding cost of distributor is reduced in this mode. Holding cost of distributor has high interaction with sharing modes with respect to distributor’s operating cost. As higher holding cost leads to smaller average order size, this increases the risk of backlog in RD and RDM mode and hence, the operating cost is higher. The increase in holding cost actually outweighs the cost saving of sharing information under these two modes. The interaction between distributor’s backlog cost and sharing modes is very small as the saving of reduced backlog, which is caused by higher unit backlog cost, cannot cover the increased holding cost (both on-hand and on-order) due to larger average order size. Hence the operating cost is higher when the backlog cost is higher under all sharing modes. Ordering cost of the distributor, like ordering cost of the retailer, does not have high interaction with sharing modes. The reason is that the high ordering cost compensates for the cost saving in RD and RDM mode. The interaction between distributor’s cost parameters and sharing modes with respect to retailer’s operating cost is similar to that of distributor’s operating cost. This is

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

because the retailer performance is directly affected by the distributor. Class III. There is little interaction effect of cost parameters of one distributor (R1) on another distributor (R2) under different sharing modes. This is because the capacity of the manufacturer can absorb change in average order size of the distributor due to change in cost parameters. The same argument can be used to explain the little interaction on the other retailer (i.e. R2) performance. The effect of sharing modes on distributor’s operating cost is less significant when retailer’s ordering cost is high. The reason is that high ordering cost outweighs the cost saving gained from information sharing in RD and RDM modes (i.e. from equations (6) and (8) respectively). Moreover, the higher order size from the retailer increases the holding cost of the distributor as indicated in BC and DM mode.

Conclusions In this paper we have analysed the impacts of information sharing on supply chain dynamics by using the proposed framework. A Web-based system is developed to facilitate the analysis. A case study on a distribution supply chain is used to illustrate this framework. The results indicate that sharing information may not be beneficial to all supply chain members. The effects of several factors (PIM) on operating cost are also identified in the case study. The following directions are identified for future research: . More complicated supply chains (e.g. more distribution channels, more intermediate tiers and assembly structure) can be investigated by using the facilities provided by this portal. . As there is imbalance of benefits due to information sharing, investigation of an incentive system for motivating information sharing in the supply chain is necessary. . Different inventory control policies (e.g. periodic review) and capacity allocation policies can be analysed based on this framework. . Optimal configuration (e.g. sharing mode, control parameters) of information sharing in a supply chain can be searched by using computational algorithm (e.g. genetic algorithm, simulated annealing). Further research is needed in developing such optimization algorithm based on the simulation model.

Based on this Web-based system, a handbook for supply chain design/redesign will be established. Guidelines have played significant roles in design decision-making process. In the field of product design, axiomatic design theory proposed by Suh at MIT has become widely accepted as a generic design framework (where guidelines are called axioms, theorems and corollaries). In the field of product design for manufacture and assembly, guidelines have also played essential roles. We will investigate the possibility of introducing the Web-based guideline system (Huang et al., 2001) for representing supply chain design guidelines.

References Axsater, S. (1996), ‘‘Using the deterministic EOQ formula in stochastic inventory control’’, Management Science, Vol. 42 No. 6, pp. 830-4. Bhatt, G.D. (2001), ‘‘Business process improvement through electronic data interchange (EDI) systems: an empirical study’’, Supply Chain Management: An International Journal, Vol. 6 No. 2, pp. 60-73. Cachon, G.P. and Fisher, M. (2000), ‘‘Supply chain inventory management and the value of shared information’’, Management Science, Vol. 46 No. 8, pp. 1032-48. Cohen, S.L. (2000), ‘‘Asymmetric information in vendor managed inventory systems’’, PhD thesis, Stanford University. Conway, D.G. (2000), ‘‘Supplier affiliated extended supply chain backbones’’, Information Systems Frontiers, Vol. 2 No. 1, pp. 57-64. Croom, S. (2001), ‘‘Restructuring supply chains through information channel innovation’’, International Journal of Operations and Production Management, Vol. 21 No. 4, pp. 504-15. Davis, M. and O’Sullivan, D. (1998), ‘‘Communications technologies for the extended enterprise’’, Production Planning and Control, Vol. 9 No. 8, pp. 742-53. Dong, Y. and Xu, K. (2001), ‘‘A supply chain model of vendor managed inventory’’, Transporatation Research Part E. Forrester, J.W. (1961), Industrial Dynamics, MIT Press, Cambridge, MA. Furst, K. and Schmidt, T. (2001), ‘‘Turbulent markets need flexible supply chain communication’’, Production Planning and Control, Vol. 12, No. 5, pp. 525-33. Gavirneni, S., Kapuscinski, R. and Tayur, S. (1996), Value of Information in Capacitated Supply Chains, Graduate School of Industrial Administration, Carnegie Mellon University, Pittsburgh, PA. Graham, G. and Hardaker, G. (2000), ‘‘Supplychain management across the Internet’’, International Journal of Physical Distribution and Logistics Management, Vol. 30 No. 3/4, pp. 286-95.

[ 357 ]

Jason S.K. Lau, George Q. Huang, and K.L. Mak Web-based simulation portal for investigating impacts of sharing production information on supply chain dynamics from the perspective of inventory allocation Integrated Manufacturing Systems 13/5 [2002] 345–358

The authors wish to thank the Hong Kong Research Grant Council and the Committe on Research and Conference Grants of the Hong Kong University for the financial support.

[ 358 ]

Helo, P.T. (2000), ‘‘Dynamic modelling of surge effect and capacity limitation in supply chains’’, International Journal of Production Research, Vol. 38 No. 17, pp. 4521-33. Huang, G.Q., Lau, J.S.K. and Mak, K.L. (2001), A Framework of Investigating the Impacts of Sharing Production Information on Supply Chain Dynamics’’, Technical Report, Department of Industrial and Manufacturing\Systems Engineering, University of Hong Kong. Kehoe, D.F. and Boughton, N.J. (2001), ‘‘New paradigms in planning and control across manufacturing supply chains – the utilisation of Internet technologies’’, International Journal of Operations & Production Management, Vol. 21 No. 5/6, pp. 582-93. Lee, H., Padmanabhan, V. and Whang, S. (1997), ‘‘Information distortion in a supply chain: the bullwhip effect’’, Management Science, Vol. 43 No. 4, pp. 546-58. Lee, H.L., So, K.C. and Tang, C.S. (2000), ‘‘The value of information sharing in a two-level supply chain’’, INFORMS, Vol. 46 No. 5, pp. 626-43. Metters, R. (1997), ‘‘Quantifying the bullwhip effect in supply chain’’, Journal of Operations Management, Vol. 15, pp. 89-100. National Research Council (2000), Surviving Supply Chain Integration: Strategies for Small Manufacturers, National Academy Press, Washington, DC. Parunak, H.V.D., Savit, R. and Riolo, R.L. (1998), ‘‘Agent-based modeling vs equation-based modeling: a case study and users’ guide’’, Proceeding of Workshop on Multi-agent Systems and Agent-based Simulation, Springer, Berlin. Parunak, H.V.D., Savit, R., Riolo, R.L. and Clark, S.J. (1999), DASCh – Dynamic Analysis of Supply Chains, Final Report, Center for Electronic Commerce, ERIM, Ann Arbor, MI. Roy, R. (1990), A Primer on the Taguchi Method, Van Nostrand Reinhold, New York, NY. Silver, E.A., Pyke, D.F. and Peterson, R. (1998), Inventory Management and Production

Planning and Control, 3rd ed., John Wiley & Sons, New York, NY. Simchi-Levi, D., Simchi-Levi, E. and Kaminsky, P. (2000), Designing and Managing the Supply Chain, Irwin/McGraw-Hill, Boston, MA. Singer, T. (1999), ‘‘Sharer beware’’, Inc. Magazine, Vol. 21 No. 4. Sterman, J.D. (1989), ‘‘Modeling managerial behavior: misperceptions of feedback in a dynamic decision making experiment’’, Management Science, Vol. 35 No. 3, pp. 321-39. Strader, T.J., Lin, F. and Shaw, M.J. (1998), ‘‘Information infrastructure for electronic virtual organization management’’, Decision Support Systems, Vol. 23, pp. 75-94. Swaminathan, J.M., Sadeh, N.M. and Smith, S.F. (1997), Effect of Sharing Supplier Capacity Information, Haas School of Business, University of California, Berkeley, CA. Tan, G.W. (1999), ‘‘The impact of demand information sharing on supply chain network’’, PhD thesis, University of Illinois at Urbana-Champaign, IL. Towill, D.R. (1991), ‘‘Supply chain dynamics’’, International Journal of Computer Integrated Manufacturing, Vol. 4 No. 4, pp. 197-208. Towill, D.R., Naim, M.M. and Wikner, J. (1992), ‘‘Industrial dynamics simulation models in the design of supply chains’’, International Journal of Physical Distribution & Logistics Management, Vol. 22 No. 5, pp. 3-13. Warkentin, M., Sugumaran, V. and Bapna, R. (2001), ‘‘E-knowledge networks for interorganizational collaborative e-business’’, Logistics Information Management, Vol. 14 No. 1/2, pp. 149-62. Wikner, J., Towill, D.R. and Naim, M. (1991), ‘‘Smoothing supply chain dynamics’’, International Journal of Production Economics, Vol. 22 No. 3, pp. 231-48. Wu, S.D. and Meixell, M.J. (1998), Relating Demand Behavior and Production Policies in the Manufacturing Supply Chain, Lehigh University, Bethlehem, PA. Zhao, X. and Xie, J. (2002), ‘‘Forecasting errors and the value of information sharing in a supply chain’’, International Journal of Production Research, Vol. 40 No. 2, pp. 311-35.