Integrated information management architecture: an object-oriented approach for problem representation, model formulation, and model execution
Kang‐Woo Lee · 1992
The research emphasis in traditional modeling has been on the development of efficient algorithms and methods for solving models. Indeed, a wealth of solution algorithms exist today for solving a wide range of models. However, models have been less than successful in achieving their full potential and in attracting wide organizational acceptance due, mainly, to the following problems: low modeling productivity, the difficulty of creating and using models, the lack of support for multiple modeling domains as well as for the entire modeling life-cycle, and the lack of data-model integration. A possible explanation for the above observation is the lack of a formal modeling methodology that addresses the problems mentioned above. This has prompted several researchers to focus their work on the modeling process itself in an attempt to develop a theory of modeling. One line of research in this emerging discipline is the Extended Entity-Relationship (E2R) approach to conceptual modeling and its associated object-oriented implementation. This dissertation develops an object-oriented modeling framework based on the E2R approach and a computer-based modeling support system that addresses the above mentioned modeling issues. The modeling framework developed in this study integrates the entity-relationship approach in database design, knowledge representation and inference techniques in artificial intelligence, and object-oriented programming. Some characteristics of the proposed modeling system are as follows. Models are built by aggregating basic units of knowledge (objects), a feature that supports model decomposition and coupling. Model abstraction makes a clear distinction between the schematic definition of a model and model instances, which promotes model reusability. The proposed system supports knowledge-based modeling, thus allowing non-technical users to build and run models without the intervention of intermediaries. Incremental modeling enables model refinement and specialization. Also, the proposed system provides a unified environment for both data and model management.