The task of modeling in a manufacturing environment centers around controlling and improving the flows of material. In representing the behavior of manufacturing systems, however, modeling perspectives vary with the needs of operations improvement and information system build-up. Moreover, use of many different modeling constructs is indispensable to represent a variety of models. Accordingly, it causes redundancy, mismatching, and difficulties in reuse and change management, of modeling output. The modeling issue makes the manufacturing system integration more difficult.
One research stream of modeling is to develop a conceptual modeling framework for a unified modeling of the material flows and control information system. This dissertation presents a conceptual modeling framework called IMF-M (Integrated Modeling Framework for Manufacturing systems) which would manage a variety of models in an integrated way with physical material focus as well as control data focus. IMF-M also incorporates operations improvement and information system perspectives. To enhance readability and change management of modeling output, it is based on the advanced modeling principles including formal, declarative and executable representation scheme, top-down approach, object-oriented conceptual blackboard, dependency graph, and model repository. IMF-M is comprised of object model, activity model, and process model. In this dissertation, the modeling constructs are defined based on the set theory to present a formalism of the modeling framework. The object model captures structural aspect of manufacturing systems and has a role of conceptual blackboard which supports model sharing, reuse, and integration. Object schema and object-relationship diagram provide a formal data model to represent the object classified into material, resource, and information. The activity model describes a variety of functionalities within manufacturing system using activity schema. Mathematical models are represented using facet graph and schema. Facet graph offers flexible abstract view of mathematical models, which enhances readability and reusability of mathematical models. Process model depicts the dynamic behavior of manufacturing systems. It is comprised of process net and schemata for event, condition, and process. Specifically, a process net can depict the material path as well as informational procedure and control sequence. IMF-M provides systematic integration facility for modeler in three ways. Firstly, separation of dynamics from functionalities in the process net promotes common use of granular activity and object models in the model repository. The separation and model repository concept enforces consistency of models and prevents the modeling output from mismatching. Physical flow and control information system can be represented in an integrated way by linking two process modeling output via event. Secondly, IMF-M supports model integration such as deep integration and functional integration. A new model can be represented using predefined model. IMF-M also permits coupling of heterogeneous models. The integration facility of IMF-M contributes to integrated management of decision models existing in manufacturing systems. Thirdly, definitional dependencies among modeling constructs of IMF-M can show an influence closure of a change, which propagates model to model change. Executable representation also facilitates model to implementation change by the model evaluation facility. IMF-M supports identification of change influence and model evaluation.
To illustrate the practical usability of IMF-M, the modeling approach and representation of modeling constructs are applied to the case of manufacturing system of a tire manufacturer in Korea. And formalism is explained by a formal definition of structure and modeling constructs of IMF-M based on the set theory.
