Using knowledge-based systems to manage quality attributes in software product lines

Abstract

Product configuration in a feature model in software product line engineering is a process, in which the desired features are selected based on the customers' functional requirements and non-functional requirements. The functional requirements of the target product can be satisfied by including the proper functional features. However, there is no such a straightforward way to realize the non-functional requirements and quality attributes of the target product. In our early work, we have developed a quantitative based method to assess the quality attributes for a configured product. However, this approach cannot adequately represent the inter-relationships among quality attributes which play an important role in product configuration process. We supplement our previous work by introducing a quality attribute knowledge base (QA_KB) to represent the inter-relationships among different quality attributes in a SPL. Furthermore, we develop algorithms for configuring a product based on customers' quality requirements. We also use a case study to illustrate our approach.