Semantic-based requirements analysis and verification

Requirements verification for complex software system has become more and more important in requirements engineering and it is one of the most helpful strategies for improving the quality of software system. Related work shows that requirement elicitation and analysis can be facilitated by semantic technologies and domain ontology. A new methodology for verifying software requirements with structural and formal semantics based on domain ontology is proposed in this paper. Requirement specification processed with natural language can be decomposed into a set of atomic requirement items which are represented with triplet of semantic elements in their domain ontology. Mapping semantic requirement items to concepts and their relationships in the domain ontology leads to a set of inference rules that are represented by description logics to verify the completeness, consistency and correctness of requirement specifications. The process of requirements verification is defined in order to evaluate the quality of requirement specifications, and its effectiveness is analyzed and evaluated with an experimental case study.