System architectures assessment based on network metrics

Characterising system architectures and describing a system’s properties using quantitative metrics is required for the comparison and evaluation of engineering systems. The aim of this article is to discuss the assessment of system architectures based on network metrics derived from literature. In this paper, network metrics such as interaction density, Newman modularity index (Q), centrality measures, cyclomatic number, and graph energy measure are included into a methodological approach for evaluating five hypothetical system architecture patterns: bus-modular, sequential, hierarchical, core-periphery and integral. Network and graph theory offer a conceptual approach to model and analyse engineering system architectures. The contribution of the article is a network metrics assessment founded on describing emergent system architecture properties as integrality, modularity, centrality, cyclicality and complexity for holistically assessing system architectures. The network metrics offer quantitative tools to gain valuable insights and can function as decision aid tools during the redesign and early development of engineering systems.