MEASURING FUNCTIONAL ROBUSTNESS WITH NETWORK TOPOLOGICAL ROBUSTNESS METRICS

This paper describes a study on the use of network topological robustness metrics to measure the functional robustness of complex engineered systems. The goal of the research is to identify network metrics capable of accounting for the functional robustness of a system without the need for a behavioral simulation. Three network topology metrics, average shortest path length, network diameter, and a robustness coefficient, were evaluated given their prominence in characterizing the topological robustness of networks but lack of evidence of their correlation to functional robustness in engineered systems. A bipartite 'behavioral' network of a drivetrain was modeled and simulated for failure using a model-based simulation package, Modelica, and a network attack approach. Average shortest path length and the robustness coefficient showed consistent topological disintegration, revealing the effect of a failure on system performance. Network diameter does not show topology changes when the failure is located outside of the cluster containing the failure. The research demonstrates the plausibility of certain network metrics to characterize the functional robustness of systems.