An Analytic Portfolio Approach to System of Systems Evolutions

Abstract

The development of large collections of systems or a ‘System-of-Systems (SoS)’ is challenging due to the large number of systems involved, complex dynamics attributed to interdependencies between systems, and inherent technical and programmatic uncertainties. The sheer number of decision variables involved in SoS development exceeds the mental faculties of the SoS practitioner, prompting the need for effective analytical support frameworks. Current frameworks and guidelines in addressing SoSE challenges lack analytical means of objective SoS level decision-making. Research in this paper adopts computational decision support methods from financial engineering that allows SoS practitioners the means to identify optimal ‘portfolios’ of systems based on dimensions of capability, cost and operational risk. The SoS architecture is represented as a hierarchical collection generic nodes that interact to provide the overarching SoS level set of capabilities based on an archetypal set of inter- nodal behaviors. Our research leverages a Conditional Value-at-Risk (CVaR) perspective to managing risks that can incorporate agent based simulation data in the decision-making process. We demonstrate the method using a LCS inspired Naval Warfare Scenario (NWS) as an illustrative case study.