|M.Sc Student||Renick Aharon|
|Subject||Incorporating Quantitative Aspects into OPM-Based|
Conceptual Models with MATLAB Computational
|Department||Department of Industrial Engineering and Management||Supervisor||Professor Dov Dori|
|Full Thesis text|
Modeling is an important part of the lifecycle of systems, starting from the early design stages. Modeling is also very useful in the process of studying an unfamiliar, existing system. Conceptual modeling methodologies disregard certain aspects of the system, making modeling or understanding a model a simpler task, as they convey the important aspects of a system in an effective way. One of the shortcomings of conceptual modeling methodologies is the simplification of the system being modeled at the expense of suppressing computational aspects. This research presents two approaches for solving this computational simplification problem for conceptual models that use Object Process Methodology (OPM), an emerging ISO 19450 standard modeling methodology. OPM offers a holistic approach for modeling systems that combines the structure and behavior of the system in a single diagram type. We expand the quantitative aspects of an OPM model by representing complex quantitative behavior using alternative approaches that employ MATLAB or Simulink without compromising the holism and simplicity of the OPM conceptual model. The first approach, AUTOMATLAB, expands the OPM model to a full-fledged MATLAB-based simulation. The second, OPM Computational Subcontractor approach, replaces low-level processes of the OPM model with computation-enhanced MATLAB functions or Simulink models. We demonstrated the two approaches with MATLAB and Simulink enhanced OPM models of a biological system and a radar system, respectively. Evaluation of the AUTOMATLAB approach, which compared information systems engineering students' performance in system modeling and analysis with and without the AUTOMATLAB layer, has resulted in significant differences in certain modeling aspects, indicating several benefits of the additional AUTOMATLAB layer compared to a non-enhanced OPM model.