A manufacturing model to enable knowledge maintenance in decision support systems
The product development process, within a typical manufacturing company, utilises huge amounts of knowledge related to manufacturing and design activities. Knowledge based systems are increasingly being used to support manufacturing and design decisions. These systems are important tools for obtaining a competitive advantage and leverage using company "know-how". However, it is important to define suitable knowledge structures in the creation of these decision support systems. Due to the significant volume of knowledge generated in the manufacturing and design stage, there is a need to create structures and methods that readily manage and maintain the knowledge in order to a) assure the long-term use of these systems b) improve the company's competitiveness. The research reported in this thesis explores and defines a Manufacturing Facility Information and Knowledge Model (MFIKM) allowing a) the ability to store and manage various types of knowledge, b) the capturing of valuable new knowledge using a knowledge maintenance method. The understanding of an information and knowledge infrastructure using different types of knowledge categorisation has been explored. The major emphasis has been placed on understanding the facility knowledge structure related to processes and resources supporting process planning decisions. Using a knowledge maintenance life cycle as a method to maintain knowledge, it was possible to capture new and valuable machining knowledge using different types of representations. Knowledge models and methods are essential in the definition of structures to support manufacturing decisions allowing knowledge management and maintenance. It has been shown that the knowledge structures defined for the new model can serve as a source and repository for different types of knowledge allowing the support of manufacturing decisions with up-to-date knowledge. The framework defined enables the structuring of facility knowledge, processes, and resources, as super classes; improving the understanding of the relationships and dependencies among them, and allowing accessibility depending on the characteristics of each. A UML tool helped in the creation of new structures detailing attributes for the classes defined. An experimental system has been implemented using the object-oriented database ObjectStore© and the Visual C++ programming environment. The MFIKM has been explored using scenarios from machining knowledge to successfully demonstrate the feasibility of knowledge maintenance supporting process planning decisions using the knowledge structures defined.