Development and Extension of a Deterministic System of Systems Performance Prediction Methodology for an Acknowledged System of Systems

Abstract

This paper addresses the need for predicting performance in a system of systems (SoS) during incremental development and for dealing with the inherent variability associated with predicting performance. Historically, senior decision-makers have used technical performance measures (TPM), along with modeling and simulation, to predict whether a system under development will meet performance requirements. This methodology does not appear to be directly translatable to SoS for several reasons, including the inherent complexity of the SoS and the operational flexibility the end user may have in employing the SoS. An approach for dealing with the SoS performance prediction has been presented previously. It laid out a notional approach to dealing with this issue. This approach has been generalized to address the use and integration of multiple technologies into an SoS and into the decision-maker's options in the use of these technologies that is rooted in using subject matter expert input and historical data. This methodology is used to develop a metric defined as an SoS performance measure (SPM), which serves as an equivalent in functionality to a TPM for a SoS. Similar to TPMs, an approach to developing tolerance bands is presented to be used for predicting the status of development as a function of time. The methodology is first presented as a deterministic method for predicting SoS performance during development. This method is then demonstrated using an example case to illustrate the methodology. However, many of the component variables have significant uncertainty associated with them during SoS development and integration into the SoS. The paper provides an approach for expanding the SPM concept to account for this uncertainty using a stochastic approach to address this issue.