A Role-based feature model componentization framework and related algorithms

Abstract

As a kind of requirement model, feature model represents functions of a family of products in a uniform form. To solve the chaos and entanglement problem in the process of feature model description and configuration, this paper designs a feature model componentization framework. With the help of the concept of Role and Reference Role, the framework proposes a process to analyze and decompose complex requirement specification to simple and coherent roles, and then implements Feature -Role-Component algorithms which map all features generated by requirement elicitation and analysis to different model components. In the framework and algorithms, role and reference role plays as an intermediary who decouple the feature and component, which makes convenient feature variants selection and composition and enhances the componentization level of the system. By decomposing the system functions in the early phase of software lifecycle, the framework improves the flexibility and adaptability of software artifact, which makes a stable foundation for higher quality product.