Automatic detection of incomplete requirements via symbolic analysis

Abstract

The usefulness of a system specification depends in part on the completeness of the requirements. However, enumerating all necessary requirements is diffficult, especially when requirements interact with an unpredictable environment. A specification built with an idealized environmental view is incomplete if it does not include requirements to han- dle non-idealized behavior. Often incomplete requirements are not detected until implementation, testing, or worse, after deployment. Even when performed during requirements analysis, detecting incomplete requirements is typically an error prone, tedious, and manual task. This paper intro- duces Ares, a design-time approach for detecting incomplete requirements decomposition using symbolic analysis of hierarchical requirements models. We illustrate our approach by applying Ares to a requirements model of an industry- based automotive adaptive cruise control system. Ares is able to automatically detect specific instances of incomplete requirements decompositions at design-time, many of which are subtle and would be difficult to detect, either manually or with testing.