Improving System of Systems Progress: Dynamically Modeling, Predicting and Affecting Rate of Convergence

A systems engineer must identify effective techniques for improving the convergence of a System of Systems (SoS). The effectiveness of the identified techniques needs to be quantified so that the stakeholders may make informed cost benefit decisions. Exploring outer worlds, defending from asymmetric threats, utilizing waning natural resources, or instantly identifying an individual's esoteric ailment all require complicated and expansive cyber-physical-social systems. The magnitude and enormity of these challenges dictate that these systems be architected of other systems, constituent systems, that converge to become an SoS. An architect must predict the rate of convergence of the constituent systems needed in order to achieve the desired functionality in a cost-effective timeframe. The architect must also be able to evaluate the SoS progress and be perceptive to methods for improving their approach.;This research developed a model for quantifying and predicting the time-phased convergence of an SoS. The rate of convergence of an SoS depends on the Political, Economic, Social and Technological (PEST) environments surrounding the SoS and thereby influencing the decisions and attitudes of the stakeholders of the constituent systems. Under the influence of the PEST environment, the stakeholders of constituent systems choose to invest in connectivity and choose to belong. The rate of convergence of the SoS was modeled with a Dynamic Bayesian Network (DBN). The time-tagged factors which established the PEST environment were used as the inputted events and the time-phased convergence of the SoS was defined to be the outcome. The model could then be used to predict the future convergence based on expectations of the PEST environment. The model could also be used to identify what would need to change in the PEST environment in order to realize a desired or prescribed future rate of convergence.;A quantified case study was performed using the U.S. Smart Grid. The U.S. Smart Grid is a developing National-level SoS. The constituent systems are the electric grids of the fifty States. The PEST environment was described by collecting time-tagged PEST factors. The convergence of the U.S. Smart Grid was quantified through the grid's Advanced Metering Infrastructure (AMI). A validated DBN was used to predict the future convergence of the U.S. Smart Grid. The DBN was also used to quantify the effect that changes to the PEST factors would have on the rate of convergence and to suggest ways to improve the rate of convergence.;The functionality of an SoS emerges from the convergence of constituent systems. This research shows that the rate of convergence of an SoS can be modeled. A system engineer may use the information from such a model to inform investment decisions and to quantify the effect of proposals to influence the rate of convergence.