them to local needs. However, their expertise is growing rapidly and will allow them to be much more innovative in the future. A very similar assessment was provided to me by another invited speaker, Prof. Stephen C.-Y. Lu Engineering Systems Research Laboratory University of Illinois at UrbanaChampaign 1206 West Green Street Urbana, IL 61801 Tel: (217) 333-6662 E-mail: lu@kbesrl.me.uiuc.edu Making these comments about Singaporean developments is in no sense to diminish their accomplishments. Lu pointed out to me that Manufacturing is not a science, at least not in a traditional definition of science. Rather, it is a true engineering. The difference here is an important one, because science is often revolutionary and engineering is always evolutionary. I agree that most manufacturing work in Singapore is not earth-shaking, but their persistent and progressive efforts in careful implementations of manufacturing technologies are, to my judgment, the right way to deal with manufacturing engineering. Further cooperation between GINTIC and Western scientists is also in the offing. Lu also told me that he is discussing the possibility of transferring one of the machine learning system developed in his Illinois laboratory to GINTIC for pre-production enhancements. A very substantial fraction of the papers presented at ICCIM'91 were written by Singaporeans, often about real problems. If the papers do not address grand challenges, then at least their proposed solutions and approaches are current and competitive. They are actively trying most of the techniques that are being discussed within the manufacturing science community today. (This is not surprising, as almost all Singaporean senior scientists have in-depth knowledge of Western approaches and in most cases have visited or studied at U.K., U.S., or other leading universities.) Furthermore, visitors to GINTIC were all extremely impressed, and if these accomplishments are judged by regional standards, the results are even more impressive. The country is trying to develop value-added solutions and perform enough focused research to make their own products more efficient, cost effective, or both. For example, we were shown a track following robot at GINTIC. The robot is built with perfectly standard technology, and if it was observed in a section of a U.S. or Japanese factory, it would have been totally unremarkable. In fact, it still had a few minor glitches. The interesting thing here was that it was entirely designed and built by a class at the university. Similarly impressive is the large amount of equipment that is available for experimentation (including a stereolithography unit), although this has all been purchased from Japanese, U.S., or European Community (EC) suppliers. (However, see the comments below in the National University of Singapore (NUS) section on potential problems.) This view of the activities in Singapore was also shared by virtually all the Western attendees at the symposium. Finally, I want to mention that Singaporean scientists are very well read and aware of what is going on elsewhere. The best example of this is their choice of invitees to ICCIM'91, but this was also clear from discussions on the conference floor. They understand the modern principles that are now guiding advanced industry, such as the ten commandments of concurrent engineering, design for assembly, design-build-teams, etc. Of course, like elsewhere, they are not always implemented as well as one would like. They are also well aware of major programs such as the Ministry of International Trade and Industry's (MITI's) IMS, micromachines, and others and participate actively in standard-making organizations. In particular, they see standards are essential to allow them to produce products that can have a large market. The key ingredients in Singaporean science, as I see it, are focus and coordination. The country has a view of where it wants to go. Essentially all government policy is thus directed toward this goal; it is not so much that general research ideas are discouraged, but instead work toward the country's goals is supported so vigorously that other activities seem to take a very minor back seat. Because of Singapore's exceptionally strong economy in recent years, there are large quantities of money available for the right kind of applied research. For example, the Government has recently announced that over the next 5 years it would spend about $S2B (about $1.3B) on the information technology part of a National Technology Plan and that by 1995 the research and development investment would reach 2% of the country's gross domestic product (GDP). The money will be focused in areas where Singapore feels it will do the most good, specifically manufacturing, computing (more generally, information technology), and biotechnology. Of these, manufacturing is seen as the key sector for the growth of Singapore and CIM is seen as the bridge between computing and manufacturing. The mission of the research community, at least with respect to manufacturing, is to invigorate the manufacturing sector with many aspects of IT. More specifically this means Provide training/consultancy/ education in use of IT in manufacturing. • Provide value-added solutions building on top of existing tools (system integration/customization). • Develop generic manufacturing applications such as manufacturing simulator, document management system, neuro-fuzzy controller, etc. • Provide general assistance in national efforts in IT. A very hard-headed look has been made into what this takes. For example, a chart was produced plotting technologies on the x-y axis of competitive (Singaporean) advantage versus benefits. Those technologies in the upper right-hand corner, knowledge systems, CIM, local/wide area networks (LAN/ WAN), Asian language, video text, and EDI value-added networks, are those that will get the most attention and support. Interestingly, technologies in the lower left corner are parallel processing, neural computing, fuzzy systems, and B-ISDN. (This view is not very different from that proposed for the EC.) In software more generally, the country sees that open systems, software development methodologies, object-oriented (OO) techniques, and CASE are most critical to its software industry. In a related study, Singapore feels that between 1995-2000 the key technologies that will be absorbed into manufacturing will be virtual reality, machine vision, EDI services, E-mail, high performance computing, and advanced control. Areas such as fuzzy logic, simulation, real-time systems, and electronic document management are already viewed as having been absorbed, and new technologies that are currently being absorbed include neuro-fuzzy control, image processing, and distributed processing. The trends that are seen in detailed manufacturing, at least in Singapore, are • Trend toward mass customization (i.e., everybody's product is a custom job). ⚫ Trend toward concurrent engineering (consideration of all aspects of product life cycle, from conception to disposal); also called simultaneous engineering. Table 1. Singapore Industries Trend toward flexible manufacturing (one assembly line easily changed; thus process rather than product specific). • Trend toward information-based organizations (fewer levels of midmanagement whose only job is to repackage and move information from bottom to top). • Trend toward knowledge workers and integration of information. In Singapore, industries are represented as shown in Table 1. Manufacturing, mostly discrete manufacturing, is a major part of Singapore's success. In 1989 it produced more than one-quarter of the GDP ($S63B), employed almost 400,000 people (more than one-quarter of the available manpower), and accounted for fully 65% of domestic exports. There were 3,700 firms employing more than 10 people, 580 firms employing more than 100 persons, and 125 firms employing more than 500 persons. Only the financial and business services sector contributes more to Singapore's GDP (40%). Singapore has many things going for it. • A pool of educated and computer literate workers. • Very generous government incentives for industry. • World class manufacturing technology, mostly from multinationals. Some of this is potentially available for technology transfer. • Astrong research and development (R&D) infrastructure. • Good tertiary institutions. • Asmall internal market (Singapore interact with modeling tools and engineering databases for process plans and will control machines via neural net • Lack of experienced technical peo- schemes. With the help of simulators, ple. • A reluctance to share information It is clear from the first three points that Singapore must look at almost everything they do from an "international" perspective. As the world becomes a more peaceful place and international cooperation, rather than competition, is increased, Singapore • Excellent communications network, will be a main and strong player. and a compact country. • Existing investment in CAD/CAM and automated manufacturing including more than 3,400 CAD workstations in smaller companies, 200 autonomous guided vehicles (AGVs), etc.). A group of very knowledgeable and capable leaders at the top to do planning. But the country also has some obvious weaknesses. A large gap between the multinationals and the small-to-medium sized indigenous industries (SMEs). SMES are seen as lacking both concept as well as technique, particularly in areas of measurement and quality control. Lack of downstream manufacturing design, particularly product design, which is seen to be very weak. Singapore has widespread applications in automation production and assembly systems of machine vision, industrial robots, etc., but feels that it lacks knowledge in interfacing and related software technology. As discussed in my earlier report, IT is highly developed in Singapore. In manufacturing more than 60% of manufacturers use some kind of computerized technology, although most of it is in accounting, finance, payroll, word processing, and inventory control. Naturally, computer usage is less in SMEs. For example, nearly 100% of the larger companies use computers for payroll, but only about 50% of the SMES. In the view of Singaporean scientists, future factories (year 2000) will be composed of intelligent manufacturing cells and built on the concepts of JIT (just in time), CIM, and concurrent engineering. Intelligent cells will have one to four computer-controlled processing machines capable of processing several parts simultaneously, connected by an automated material handling system and controlled by a programmable cell controller. One research project is to develop a truly adaptive manufacturing cell. It is expected to involve object-oriented modeling, expert systems, neural nets, and a simulation user interface for the operator. Neural nets will be used for pattern recognition, sensation, reaction, and control. Expert systems will provide machines with reasoning and deduction capabilities. The expert system controller will an operator can predict what will happen and study what-if situations. On top of the factory system is a collection of software, distributed manufacturing resource planning, document management system, intelligent control, and intelligent interface software. This, in turn, is built upon various standards (such as X.400, IGES, Open Systems, etc.) and basic technologies (CAD/ CAM, real time systems, etc.). In fact, standardization is an important component of Singapore's R&D programs in manufacturing. Proposed areas for R&D in manufacturing are system). Event-oriented simulation modeling is used. The simulator keeps track of the discrete events occurring in the system and updates entities like work center parts and products at the completion of each event. Rules and knowledge of production planners will be stored in the knowledge-based analyzer module. I met with Nara and listened to him explain the main ideas of his simulator. This is a nice project, but it is just getting started. I also met with one of the team members, a recent B.S. graduate who had very little experience in this area and was learning while programming in C++. Nara has many years of actual shop floor production experience. The second project is by Dr. Beng Siong Lim CIMIDES is a computer-integrated manufacturing information and data exchange system, which is in reality a network of mini expert systems and a director of information. The director coordinates the data exchange between the individual systems, both the raw data as well as deduced information, maintains consistency, and schedules and prioritizes contributions by use of a queue. Each mini expert system controls a specific area and has its own particular rules. These systems include the following. • Intelligent Product Configuration System • Knowledge-Based Product Designer • Intelligent Component Designer • Intelligent Knowledge-Based Object-Oriented Process Planner • Modular Fixture Design Expert It has the usual programs and some System unusual ones such as graduate programs in International Construction • Intelligent Quality Assurance Management and Hotel AdministraPlanner tion. Special laboratories and centers include the Institute for Manufactur • Intelligent Material and Equipment ing Technology, Center for Advanced Planner Autonomous Guided Vehicle Planner Construction Studies, Microelectronics Center, Computer Graphics Center, and Center for Transportation Studies. NTU emphasizes work-study programs; in the United States they are • Expert Production Scheduling called co-ops, but Singapore calls them System "business attachments." There are also a number of collaborative programs • Flexible Manufacturing System with Western universities, including Controller Each mini expert system is developed using an object-oriented representation and the corresponding knowledge bases are organized into a ring structure. The main difference between this approach and other CIM systems is that it employs a common knowledge base rather than a common database. GINTIC INSTITUTE OF CIM From my perspective, the most interesting aspect of ICCIM'91 was the opportunity to see firsthand the work being done at GINTIC. This facility is located on the campus of Nanyang Technological Institute (NTI) about a 40-minute bus ride from the center of Singapore City, and essentially at the far west side of the island. Although NTI has been in existence since 1972, on 1 August 1991 it changed its name from an institute to a university (NTU); literature refers to both although they are the same place. Nanyang is the "other" large tertiary institution in Singapore, Singapore's National University being the first. NTU caters to about 9,000 mostly residential undergraduate students and about 400 graduate students, although some of those are part-timers. Its focus is on science, engineering, business, and management. the Sloan School at the Massachusetts Institute of Technology (MIT), the University of Warwick, and Loughborough University of Technology. The NTU campus, which rises somewhat Phoenix-like out of the tropical vegetation, is a study in modern cast concrete. The buildings, including many dormitories, appear to be no more than a few years old, although doubtless some are from the 1970s. In my earlier report I described GINTIC briefly, but had no opportunity to visit at that time. This was originally established by NTI in partnership with Grumman in 1985 under the name “The Grumman International/ NTI CAD/CAM Center" as a 5-year program to develop local CAD/CAM expertise. When the collaborative arrangement with Grumman ended in 1989, the organization was renamed GINTIC Institute of CIM; its 5-year period 1989-1994 is thought to be the second phase of its growth and is funded by the National Science and Technology Board at $S50.3M. The GINTIC mission is to develop local expertise in CIM applications by working with local partners on collaborative research, technology transfer, and training and education. GINTIC's role is perfectly consistent with the comments that I made earlier about manufacturing scimade earlier about manufacturing science research in Singapore. It should be noted that the current GINTIC efforts are mainly focused on the software aspects of CIM. In fact, they have decided to establish a sister institute to GINTIC, called the Institute for Manufacturing Technology (IMT), with its own separate building next to GINTIC, that will focus on the hardware developments for CIM. Currently, GINTIC is housed in several buildings near the Mechanical Engineering Department on the NTU campus, but a new building is to be built this year to bring the groups together and give them additional space. Currently there is a staff of 84 divided into four groups, • Research • Applications to provide services to industry • Business to promote GINTIC and seek out commercialization opportunities • Systems to provide support for the other groups There is a management board and an international advisory panel, the latter including senior university scientists from Germany, the United States, Japan, and the United Kingdom. Research programs at GINTIC reflect the comments and views that have been made earlier. • Management aspects of CIM ⚫ System design ⚫ System modeling and simulation • Factory automation • Product design The emphasis on industrial technology transfer is apparent. For example, last year a CAD/CAM project for the jewellers industry was begun, as was a knee joint prosthesis project for Singapore Aerospace and the National |