Examination of modeling languages to allow quantitative analysis for model-based systems engineering

Abstract

Model-based systems engineering (MBSE) needs a formal language, one defined with explicit rules between its elements, in order to support the use of formal modeling in systems engineering. This thesis examines desirable features in the context of quantitative modeling for systems engineering modeling language. Object Management Group’s UML and SysML and Vitech’s System Definition Language are then analyzed in terms of these features. The first important feature is the capability for interoperability between different MBSE tools combined with the ability to integrate the use of specialty tools to interact with and manipulate the system model. Flexibility is necessary in describing and defining entities in the system modeling language. This allows supporting project specific concerns in the system semantics, making MBSE tool support simpler. Finally, support for non-fixed value properties for entities, particularly random variables, is essential to representing system behavior. Existing system modeling languages have shortcomings that should be addressed to improve the conduct of MBSE. Random variables are inconsistently supported. Behavior modeling allows intermingling event timelines for different entities, preventing automated analysis of possible event sequences. Finally, support of parametric modeling is not universal, and the semantics for the use of black box entities to represent external analysis is ad-hoc.