Final Report of IMS Test Case 7

GNOSIS: Knowledge Systematization: Configuration Systems for Design and Manufacturing


GNOSIS partners have been aware that the test case is an experiment in the value of international collaboration in IMS precompetitive research, and the technical work package activities have focused on satisfying the GNOSIS objectives through various forms of collaboration. To achieve any meaningful result, it was necessary to choose a small number of sub-projects, each exemplifying some form of collaborative activity in knowledge systematization. Thus, it was agreed to exchange only a few tools and a few problems of the many available. Individual partners agreed to participate in specific roles in different collaborative activities, even though their interests in GNOSIS were far broader. The fact that it was possible to achieve this agreement is itself a significant result, indicating a spirit of trust and willingness to bring individual partner objectives into line with those of the project as a whole, an attitude essential to significant international collaboration.

3.1 Project Deliverables

The GNOSIS test case deliverables encompass a wide range of outcomes of value to individual partners, to industry in their countries and regions, and to the international manufacturing community. The main deliverables identified are:-

3.2 Work Package Tasks

Following the workplan's guidelines (Section 2.3) each work package developed its own program, and the overlapping membership provided coordination. TW1 and TW5, in particular, had specific objectives of integrating the other packages, in terms of knowledge systematization activities, and post mass production paradigm issues, respectively. TW1 coordinated a consortium-wide joint demonstration with work packages TW2, TW3 and TW4. Each work package generated subtasks that were themselves assigned to collaborative groups of partners, typically across several regions. Five major types of subtask are apparent: Much more has been achieved than seemed realistic when GNOSIS commenced, and it is not possible to do more than highlight some significant results from each form of activity. The results to date have been documented in a wide range of technical reports that are available to interested parties.

3.2.1 State of the Art Surveys

Each of the Technical Work Packages has undertaken surveys of the state of the art in one or more significant aspects of their areas of interest. These surveys are currently in final form and will constitute major deliverables from GNOSIS to the IMS community.

3.2.2 Systematization of Knowledge and Views

TW5 used a group elicitation methodology to develop a framework encompassing all the issues relating to a post mass production paradigm. The six elements were: boundary conditions; initial conditions; critical drivers; implementation; expected results; obstacles. Five partners representing a range of viewpoints and regions then developed their own perspectives on issues within the framework, and these were consolidated into categories through an iterative process. The result is a systematization of the major issues of a post mass production paradigm.

3.2.3 Shared Tools

Many tools were shared between individual partners. Two major experiments on widespread sharing were carried out. One involved the SYSFUND functional design tools from the University of Tokyo, an example of advanced university research, and the other the QUEST factory simulation tools from Deneb Robotics, an example of a leading edge industrial project. Thirteen partners were supplied with SYSFUND as IMS background technology. Manuals, training (in Tokyo and on-site), and support through email were provided. Six partners were supplied with QUEST on a variety of different bases. The timescale of GNOSIS has made the full investigation and use of these complex tools difficult, but the preliminary results are very encouraging and results of the use of the tools, and particularly their integration with other tools, figure prominently in the individual partner and TW reports. One lesson learned is that the resources required to effectively use such tools is very high, and it is generally appropriate to focus on the exchange of a few specific tools.

XDSP, a designer's spread sheet from Kyushu Institute of Technology, was applied to the elevator configuration problem, which is a classic benchmark for configuration, and was also applied to real construction problems (temporary platform design and structural design of long-life foundation) by Shimizu.

3.2.4 Shared Data Sets

The first data set was provided by the French company, Télémécanique, and was the manufacturing information for a family of contactors. It was initially used by TW2 for research on configuration management, and it was also adopted in other work packages. The second data set complemented the first by providing a STEP data set for a mechanical part from the University of Calgary DME. It was initially developed in TW3 for research on configurable manufacturing, and was also adopted in TW1.

The use of these data sets went across work packages TW1-4, and was summarized in the report "Main Joint Demonstration of GNOSIS Technical Work Package 1" and the "TW3 Technical Work Package Report" which are available. There were additional data sets that were used by individual partners. Activities and demonstrations based on shared data sets include:

3.2.5 Shared Models

Different models and associated software were exchanged in the area of product configuration management. They all concern the increasing complexity of products, the necessity to reduce the time to market and the costs as well as to improve overall quality.

IPA provided the fractal company concept which concern enterprise organization, VTT the virtual factory model, and LLP-CESALP, the OLYMPIOS model which concerns information and decision systems following the "consumer-supplier" relationship concept and the GENERIS model which provides links between equipment and organization.

3.3 Technology Transfer and Dissemination Activities

An open-day on GNOSIS activities was held at the inter-regional meeting in Kyoto on the 12th of November and attended by over 30 industrial representatives. Partners have similar presentations planned for industry in their own countries.

Partners who have industry training activities such as VTT, IPA, ADEPA and ARC will use these activities to disseminate GNOSIS material and results.

The exchange of researchers has also been an important basis for transfer and dissemination, but due to the short time scale only a few arrangements could be made. The main exchange was for short periods of SYSFUND training. One researcher from VTT stayed at Shimizu Corporation for two months to study the application of the virtual factory in construction. A researcher from Shimizu Corporation is at VTT for one year in 1994.

It is important to note that we focused on a few activities so as to achieve some in-depth studies of significance for longer-term planning. For example, the two widely exchanged tools are representatives of some 30 tools identified as potentially exchangeable between partners, some of which were exchanged between individual partners. Similarly, the two data sets exchanged are representatives of large numbers of industrial data sets that could have been made available to partners, some of which were exchanged between individual partners. It was felt in planning GNOSIS tasks that the resources necessary to effectively accept and use such material were very high, as was the burden of support upon the supplying partner. This was confirmed in the actual exchanges that took place. Thus, in planning future projects it is important to take into account that the dominant factor is not the major potential for technology transfer and dissemination, but rather the costs of doing this for both donors and recipients.

The final versions of the overall GNOSIS report, the five technical work package reports, and a number of specific technical reports on GNOSIS projects are being issued at the end of the project in February 1994 and distributed to the IMS community. In addition a number of journal papers, technical magazine articles are planned. Press releases on the GNOSIS consortium have been issued.

Contents, Previous Section, Next Section, IMS Page, KSI Page 1-Sep-94