Assessing Performance Uncertainty in Complex Hybrid Systems

Abstract

Complex systems are often a derivative of the current trend that dictates increased functionality and performance. Mobile phones that provide internet access and music player, fighter aircraft with fly-by-wire for flight control, and energy grid that also monitors usage and recommends corrective actions are few of the examples of emerging complex systems. There are a large number of causes that could lead to increased complexity of a system, such as (1) increase in number of parts, (2) increase in number of interactions, and (3) integration of multiple technologies. Increase in complexity makes these systems significantly challenging in terms of design and optimization, resulting in large time and resource investment. Therefore, it is typical for complex systems to use non-mature technologies that are expected to mature along the way. However, such infusion introduces significant levels of uncertainty, which historically have led to explosion in development cost. In this paper, we assess the impact of uncertainty on a complex system. Through numerical simulations, we demonstrate that increasing complexity does not necessarily result in vulnerability of a system to uncertainties. These results challenge the conviction held by complex system design community, which assumes complexity and uncertainty have a direct correlation. Our results indicate that selection of system architecture is critical for system robustness. This should provide motivation for further research in the complex system arena.