Knowledge Based Method to Validate Feature Models

Feature model has been used to support requirements analysis and domain engineering in Software Product Line by representing variability. This paper proposes a knowledge base method to validate feature model. Validation of feature model has been split into two main processes, automated consistency check (by defining rules control the variation selections considering cross-tree constraint dependencies) and the second process is automated error detection(two types of errors, dead feature and inconsistency, was defined and validated). Variability was described among feature model, and then variability was represented as a knowledge base containing predicates and rules. In addition to validation, the proposed method can be used to identified and provide auto support for some operations on the automated analysis of feature models (propagation, cardinality validation, explanation, and optimization).