Abstract
Some authors suggest that transitioning requirements engineering from the traditional statements in natural language with shall clauses to model-based requirements within a Model-Based Systems Engineering (MBSE) environment could improve communication, requirements traceability, and system decomposition, among others. Requirement elements in the Systems Modeling Language (SysML) fail to fulfill this objective, as they are really a textual requirement in natural language as a model element. Current efforts to directly leverage behavioral and structural models of the system lack an overarching theoretical framework with which to assess the adequacy of how those models are used to capture requirements. This paper presents an approach to construct true model-based requirements in SysML. The presented approach leverages some of SysML’s behavioral and structural models and diagrams, with specific construction rules derived from Wymore’s mathematical framework for MBSE and taxonomies of requirements and interfaces. The central proposition of the approach is that every requirement can be modeled as an input/output transformation. Examples are used to show how attributes traditionally thought of as non-functional requirements can be captured, with higher precision, as functional transformations.
Highlights
Problem formulation, traditionally in the form of requirements, is considered by some authors to be the cornerstone of systems engineering [1]
The central proposition of the approach is that every requirement can be modeled as an input/output transformation
The central proposition of our theoretical framework is that every requirement can be modeled as an input/output transformation
Summary
Problem formulation, traditionally in the form of requirements, is considered by some authors to be the cornerstone of systems engineering [1]. Incorporating models of requirements within a central model of the system, instead of leveraging textual statements in natural language, could “facilitate requirement understanding and foster automatic analysis technique” [5]. This need demands exploring the question of what forms a good model of a requirement or a model-based requirement. SysML incorporates elements called requirement element and requirements diagram [8] These are intended to model the requirements the system is expected to fulfill. The second path is promising because it directly leverages behavioral and structural models of the system without requiring statements in natural language. The requirements and interfaces taxonomies provide context for justifying completeness of requirement diversity in the resulting models [14]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
