טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
Ph.D Thesis
Ph.D StudentLevi Natali
SubjectNarrowing the Gap between Systems and Software
Engineering by Integrating Computations into
Object-Process Methodology
DepartmentDepartment of Industrial Engineering and Management
Supervisors PROF. Dov Dori
DR. Ahmad Jbara


Abstract

Lack of a modeling framework that integrates systems with software engineering is a major cause of product development problems. The current model-based systems engineering approach applies a variety of model kinds, each with its own fidelity level, with disparate or loosely integrated models. The gap between the system model and the software levels leads to a great fidelity gap, as the software is often not aligned with the model of the system it is supposed to control. This doctoral thesis aims to overcome this system-software modeling gap by integrating computational, software-related, and model execution capabilities into Object Process Methodology - OPM ISO 19450-based conceptual modeling, resulting in a holistic unified executable qualitative-quantitative modeling framework. The gap is bridged by extending OPM with a Methodical Approach to Executable Integrative Modeling (MAXIM). We exemplify MAXIM usage by a model of a civil aircraft landing gear braking system, showing that engineers from different domains can collaborate during early system development phases to jointly construct a holistic model, combining qualitative and quantitative aspects. During the research, we came across the model fidelity hierarchy, which is presented via a case study of modeling an aircraft landing gear. OPM with MAXIM enables continuous, seamless modeling approach with increasing accuracy. As errors are revealed during early system lifecycle stage, they are exponentially less costly to correct than those revealed downstream. The principles of MAXIM are presented and demonstrated within OPCloud?a web-based collaborative conceptual OPM modeling framework.

Models have traditionally been mostly either prescriptive, expressing the function and structure of a system-to-be, or descriptive, specifying a system so it can be understood and analyzed. MAXIM provides a new methodology for developing and engaging with a new, third family of models?diagnostic models.  As a case in point, we have built a model for assessing potential pediatric failure to thrive (FTT). As part of this Thesis, we present FTTell?an executable model-based medical knowledge aggregation and diagnosis tool, in which the qualitative considerations and quantitative parameters of the problem are modeled using MAXIM. The efficacy of the tool is demonstrated on data collected from 100 children, providing 87% correct diagnosis. Pediatricians can use this model-based standardized approach to improve their FTT diagnosis for appropriate timely intervention.

The development of MAXIM as a major extension of OPM, which promotes it from a conceptual modeling language and methodology to a combined conceptual-computational modeling approach, lays out a solid foundation for extending it further to support full-fledged simulation and execution of systems of all kinds. Moreover, the ability to inject code into model processes will potentially allow smoother reverse engineering. In case we do not know how to make the translation between code and graphical representation we can retain the code as is and simply inject it into the process with slight adaptations. Another direction is teaching programming for novices by starting with visual programming and progressively moving to traditional coding. MAXIM provides us with an approach that combines traditional coding with visual programming, benefiting from both modalities.