Detecting of requirements inconsistency: an ontology-based approach

Detecting of requirements inconsistency is key to the development of heterogeneous software system. But at present, although there is a lot of work on the detecting of requirements inconsistency, most of them is limited in detecting of requirements inconsistency at syntax level, we still lack of a promising method for the detection of requirements inconsistency at semantic level. Based on an abstract requirement refinement process model, this paper takes domain ontology as infrastructure for the refinement of software requirements, software requirements can then be seen as interactions between software agents and the ambiences they exist. We assume that the real effect of these interactions is to make the state of ambience changed, so we explicitly represent requirements of software agents as a set of state transition diagrams, each is for one entity in the ambience. And we realized semantic level inconsistency detecting of software requirements based on state transition of ambience.