A Robust Optimization Framework to Architecting System of Systems

Abstract

The architecting of a System-of-Systems (SoS) is greatly challenging due to coupling of managerial independence and highly complex operational interdependence, in achieving desired overarching capabilities. Furthermore, difficulties arise when uncertainties in the performance of individual, interconnected systems, lead to significant risks of cascading modes of failure. Mitigation of these risks through appropriate selection and design of constituent systems is important to ensure resilience of SoS architectures to operational disruptions, whilst preserving resource and cost constraints. Current guidelines and tools for architecting SoS architectures are lacking sufficient capabilities in enabling effective decision-making for SoSE practitioners. This paper presents a novel robust optimization framework to architecting a System-of-Systems (SoS). Hierarchies of operationally interdependent systems are modeled as nodes on a network that work cohesively to fulfill overarching capability objectives. Inter-nodal performance and constraints associated with connectivity under performance uncertainty are addressed. Recent advances in robust optimization methods are employed within the context of a Mixed-Integer Program (MIP) for the developed framework. A simplified case scenario using the Littoral Combat Ship (LCS) platform demonstrates application of the formulation.