A System-of-Systems Approach to Enterprise Analytics Design: Acquisition Support in the Age of Machine Learning and Artificial Intelligence

Abstract

System-of-Systems (SoS) capability emerges from the collaboration of multiple systems, which are acquired from independent organizations. Even though the systems contribute to and benefit from the larger SoS, the data analytics and decision-making about the independent system is rarely shared across the SoS stakeholders. The objective of the research presented in this paper is to identify how the sharing of datasets and the corresponding analytics among SoS stakeholders can lead to an improved SoS capability. Our objective is to characterize how appropriate use of data sets may lead to deployment of different predictive (predicting an outcome from data) and prescriptive (determining a preferred strategy) analytics and lead to better decision outcomes at the SoS level. We build and demonstrate a framework for this objective, based on extensive literature review, which utilizes appropriate predictive and prescriptive methodologies for SoS analysis. Additionally, we propose to utilize machine learning techniques to predict the achievable SoS capability and identify sources of uncertainty derived by sharing partial datasets. A case study demonstrates the use of the framework and prospects for future improvements.