Improving OPM Conceptual Models by Incorporating Design Structure Matrix

Abstract

We combine Object-Process Methodology (OPM) with Design Structure Matrix (DSM) and its Multi Domain Matrix (MDM) extension to capitalize on the potential synergies of these two approaches to model-based systems engineering. OPM can provide a very detailed system model, but it is limited in examining the system as a whole, especially when the system is large and complex. DSM, which provides a matrix representation of the system, lists all the elements and shows relations between model elements. This matrix is amenable to matrix manipulation like clustering and partitioning, enabling the evaluation of relations among model elements. We follow a methodical five-step approach for applying DSM to analyze and improve complex OPM system models. Dedicated preprocessing algorithms have been developed and applied to eliminate states from objects and flatten the hierarchical OPM model while maintaining its semantics. The resulting DSMs and DSM variations of an OPM model are then analyzed, this can be achieved by domain experts who can get ideas or suggestions to improve the model based on quantitative metrics. After examining outcomes of clustering or partitioning algorithms on the DSM, the modularity of the model can be improved, yielding possible gains for the system under development.