Reducing the cost and complexity of variant exploration with MBSE & LSP

Abstract

Variant exploration is often construed as synonymous with a significant increase in work and a decrease in clarity. Model-Based Systems Engineering (MBSE) has been heralded as a potential solution to streamline the process by supporting exploration of numerous parallel variants within a common multidimensional model. Successful decoupling of variants from their surroundings supports the parallel development of interfacing systems without delay or impediment, thereby permitting the deferment of a final decision on the preferred technology choice. Still, a number of challenges remain. Some have been addressed by previous research, though, not focusing on the effect of variant technology choices in the logical tier, nor on embedding variant management as a native part of the core MBSE model. This paper addresses those challenges via application of Liskov's Substitution Principle (LSP). It defines an MBSE based approach that explicitly documents variance points, identifying their source and scope, and integrates their management into the standard Systems Engineering (SE) workflow. Emphasis is given on providing a pragmatic approach that promotes consistency, visibility, and simplicity. The result is a model that embeds the required information artifacts without adding significant complexity. The proposed strategy does not prescribe excessive variant-specific activities, with the majority of the effort being already part of standard Requirement, Interface, and Logical SE processes. It is tolerant to an iterative approach, and proactively supports concurrent engineering through decoupling.