This research details the development of a design process and decision-making environment to support the acquisition of an air-based, interoperable, autonomous system-of-systems to support suppression of enemy air defenses missions. 12 The design is done using a three part process, which includes the concurrent development of the DoDAF products, modeling and simulation environment, and decision support environment. In particular, the research focuses on understanding how the interactions, interoperability, and autonomy level of the unmanned aircraft affect the overall mission performance.

Systems can be arranged in useful ways to perform tasks which, executed in an order, unlock overall capabilities. The goal of a system of systems engineer is to find the system arrangements and task execution order that allow the system of systems successfully achieving its end goal efficiently. The difficulties in this work are the sheer number of possible arrangements and work flows, uncertain and missing information about systems, and greater flexibility required in evaluation compared to system evaluation. This paper describes a modeling and simulation solution that deals with all the three difficulties enumerated as part of a research work sponsored by the Office of Naval Research. The modeling and simulation solution implemented here is in the form of a discrete event simulation. Because system of systems design involves itself with many missions, the simulation is developed to handle various missions and systems. However, in this paper the use case is a suppression of enemy air defenses mission carried out by a variety of naval military systems.

No abstract

No abstract

J ust as we put components together to build systems for various purposes, we are now interfacing systems with each other so that they can collectively achieve goals that none of them could achieve alone. This collective is what is referred to as a system of systems. Systems engineering, given its focus on the holistic design of a system, was the logical field to turn to for dealing with systemof-systems problems. Systems engineers used models to describe, define, document, and also potentially simulate systems. The same approach is applicable to system-of-systems problems; however, aspects such as distributed authority, independent management, the high degree of complexity, humans in the loop, and the large numbers of interactions can create emergent behaviors for the system of systems, which are very difficult to model, detect, and output for further analysis (Sage and Cuppan 2001: 325-345; US Department of Defense 2008; Haskins 2011). Additionally, the outcomes of actions taken by systems within the simulation are not certain (e.g., a bomb can hit or miss a target). The combination of these aspects is what causes simulations of systems of systems to be stochastic.Initially a description problem, the study of systems of systems has adapted to the system engineering standards such as Systems Modeling Language (SysML), the US Department of Defense Architecture Framework (DoDAF), and NATO Architecture Framework diagrams. These formats practically define system-of-systems architectures in varying levels of detail. Each of these description languages has its own strengths and shortcomings. While the community is continuing to develop such descriptive languages, there is also room to explore our options in evaluative models that can be "executed." Whereas description can find the characteristics of the system of systems under study, execution provides measures of performances for the system of systems in its virtual operating environment. Such performances could include mission effectiveness, cost, duration, and attrition rate. The ability to analyze the performance of a system of systems enables better design of the system.Systems engineers must consider all the mentioned characteristics of the system of systems when they choose an executable modeling technique. Unsurprisingly, modeling techniques that are used in the field of large-scale operations research are appropriate for system-of-systems simulation. The engineer must also remember that all models are wrong, and that the models' adequacy is determined by the scope of the problem. Having worked with many different executable models for system-of-systems simulations, we have found that the most useful tools for evaluating systems of varying size and purpose are graph models, Markov chains, Petri nets, system dynamics, discrete-event models, and agent-based models. The selection depends on the problem. If there are many design alternatives to compare, we desire models that run quickly such as mathematical models. For combinatorial problems we seek flexibility in...