A Model-Based Systems Architecting and Integration Approach Using Interlevel and Intralevel Dependency Matrix

The way system architects typically decompose a high-level system architecture has an outsized impact on interface complexity of systems and system-of-systems (SoS). In existing practice, systems are decomposed based on empirical knowledge of system architects and the prevailing architecture of organizations responsible for developing systems. The decomposition of systems (and SoS) involves aspects of engineering, management, and social sciences. How the decomposition approach breaks down a complex system determines the effectiveness of the integration strategy. In this article, we recommend that a complex system (and more specifically, a complex sociotechnical system) should be decomposed and aligned along five perspectives or layers: operation, function, physical, implementation, and organization. At each layer, the system architecture exhibits a certain level of complexity. Aligning modules at each layer with modules at the level above can help reduce the complexity of the overall system making it more manageable. This article presents a system architecting methodology that employs interlevel and intralevel dependency matrix (I2DM) to enable this approach. The I2DM methodology employs stochastic optimization in combination with heuristics to align modules in the different layers and then identifies problematic interactions that could pose problems in system integration. The article also describes how techniques from pattern recognition, data analytics, and clustering can be applied to system architecture models to formulate and assess the efficacy of a system integration strategy. The ontology underlying the systems architecting and integration strategy is also presented.