An Analytic Workbench Perspective to Evolution of System of Systems Architectures

• Published in: Procedia computer science 2014, Vol.28, p.702-710
• Main Authors: Davendralingam, Navindran, DeLaurentis, Daniel, Fang, Zhemei, Guariniello, Cesare, Han, Seung Yeob, Marais, Karen, Mour, Ankur, Uday, Payuna
• Format: Article
• Language: English
• Subjects: Analytics; Architectures; Optimization; System of Systems
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2014.03.084

The development of a large group of interdependently operating systems, or ‘System of Systems (SoS)’, presents significant challenges across technical, operational and programmatic dimensions. Trades between cost, schedule, performance, and associated spectrum of risks, are essential during analysis of alternatives for both individual systems and the SoS architecture as a whole. The large number of decision variables involved, ubiquitous uncertainty and complex interactions that exist between systems creates analysis problems that go well beyond the immediate mental faculties of decision-makers. Often times, the decisions made focus on localized development at the systems level with little consideration for cascading effects on the bigger SoS picture. Hence, the process of evolving SoS architectures requires tools that provides the SoS practitioner with meaningful analytical quantifications of the SoS tradespace. In the defense arena, existing tools for such trades, have been guided by policies set forth in the Defense Acquisitions Guidebook (DAG) (5000 series) and the System Engineering Guide for System of Systems (SoS-SE), but are lacking an analytic perspective towards more informed decision-making. This paper discusses a multidisciplinary effort, funded by the DoD's Systems Engineering research Center (SERC), to establish an analytic workbench of computational tools to facilitate better-informed decision-making on SoS architectures. The work is motivated by the idea that SoS practitioners possess relevant information and archetypal questions that reflect desired outcomes at the SoS level. These archetypal, technically -driven queries are mapped to relevant methods that can provide analytical outputs to directly support SoS acquisition and architectural decisions. The applicability and respective value-added of each method in addressing various archetypal measures are presented.