Pattern-Based Methods and MBSE

Abstract

Patterns are recurring regularities, having fixed and variable parts, across engineered systems, systems of engineering, production, distribution, and sustainment, as well as the natural world. Ranging from concrete patterns of engineered product lines to abstract patterns behind architectural frameworks, reference models, ontologies, and general or domain-specific languages, patterns are implicitly involved in all MBSE practice. Methods reported in this chapter exploit the power of explicit MBSE patterns, using the leverage of acquired knowledge to speed processes, reduce rediscovery and error, and lower risk.

Although MBSE has begun exploiting them, model-based patterns have a much longer history in the foundations of the physical sciences and their support for engineering disciplines. Advanced primarily over the last 300 years, all the laws of the physical sciences are represented by model-based patterns describing and organizing observable phenomena. Indeed, that advance was largely centered on the rise of more effective representational models and modeling frameworks, mathematical and otherwise.

In the same fashion, the theoretical foundations of systems engineering are being strengthened by similar observation of system phenomena, their representation as configurable model-based patterns, and validation of credibility of such models for prediction and understanding. Properly harnessed, this creates more effective learning by groups (e.g., teams, enterprises, supply chains, industry groups), in which reusable, configurable model-based patterns are the informational proxy for learning; the techniques of pattern recognition and model verification and validation provide recognized support for formalized learning and model credibility management; and pattern-based processes are automated to enhance generation or checking of models supported by patterns.