| Summary: | Discrete event simulation is now a well established modelling and experimental technique for the analysis of manufacturing systems. Since it was first employed as a technique, much of the research and commercial developments in the field have been concerned with improving the considerable task of model specification in order to improve productivity and reduce the level of modelling and programming expertise required. The main areas of research have been the development of modelling structures to bring modularity in program development, incorporating such structures in simulation software systems which would alleviate some of the programming burden, and the use of automatic programming systems to develop interfaces that would raise the model specification to a higher level of abstraction. A more recent development in the field has been the advent of a new generation of software, often referred to as manufacturing simulators, which have incorporated extensive manufacturing system domain knowledge in the model specification interface. Many manufacturing simulators are now commercially available, but their development has not been based on any common standard. This is evident in the differences that exist between their interfaces, internal data representation methods and modelling capabilities. The lack of a standard makes it impossible to reuse any part of a model when a user finds it necessary to move from one simulator to another. In such cases, not only a new modelling language has to be learnt but also the complete model has to be developed again requiring considerable time and effort. The motivation for the research was the need for the development of a standard that is necessary to improve reusability of models and is the first step towards interchangability of such models. A standard framework for manufacturing simulators has been developed. It consists of a data model that is independent of any simulator, and a translation module for converting model specification data into the internal data representation of manufacturing simulators; the translators are application specific, but the methodology is common and illustrated for three popular simulators. The data model provides for a minimum common model data specification which is based on an extensive analysis of existing simulators. It uses dialogues for interface and the frame knowledge representation method for modular storage of data. The translation methodology uses production rules for data mapping.
|
|---|
