Evaluating System Architecture Quality and Architecting Team Performance Using Information Quality Theory

As engineering projects grow in complexity, estimating the required engineering effort during the development phase of a project has become more art than science. Predictions of required engineering effort are based on empirical models fitted to historical data with additional subjective factors, such as “team cohesion,” applied based on the judgment of the user of the model. Recent work has shown that information quality theory can be used to estimate the engineering effort necessary to develop system requirements by evaluating the change in requirements uncertainty during the requirements engineering process. This paper builds upon that foundation by generalizing the information quality theory to apply to the variety of system engineering artifacts generated during the development phase of a program. The generalized form of information quality theory is implemented in a model for application to the system architecture definition process. Engineering effort required for system architecture definition on four programs is used to evaluate both the performance of the model and the performance of the system architecture definition teams. The results of the model and program evaluations show clear benefits in reducing the engineering effort required to define a system's architecture when there is reuse from previous programs. The results of the evaluation also show the architecture definition team benefits from “momentum” and performs more efficiently if the team has completed a system's architecture definition on a recent program.