A genetic algorithm for optimized feature selection with resource constraints in software product lines

• Published in: The Journal of systems and software Vol. 84; no. 12; pp. 2208 - 2221
• Main Authors: Guo, Jianmei, White, Jules, Wang, Guangxin, Li, Jian, Wang, Yinglin
• Format: Journal Article
• Language: English
• Published: New York Elsevier Inc 01.12.2011; Elsevier Sequoia S.A
• Subjects: Algorithms; Automation; Commonality; Computer programs; Configurable; Configuration; Cost engineering; Feature models; Feature selection; Genetic algorithm; Genetic algorithms; Heuristic; Mathematical models; Optimization; Optimization algorithms; Product derivation; Software; Software engineering; Software product lines; Studies; Product derivation; Software product lines; Configuration; Genetic algorithm; Optimization; Feature models
• ISSN: 0164-1212; 1873-1228
• DOI: 10.1016/j.jss.2011.06.026

► We propose GAFES for optimized feature selection in software product lines (SPLs). ► GAFES is a GA-based heuristic for SPL feature selection with resource constraints. ► A repair operator transforms an arbitrary feature selection into a valid one. ► GAFES produces solutions with 86–97% optimal of prior feature selection techniques. ► GAFES achieves 45–99% faster than prior feature selection techniques. Software product line (SPL) engineering is a software engineering approach to building configurable software systems. SPLs commonly use a feature model to capture and document the commonalities and variabilities of the underlying software system. A key challenge when using a feature model to derive a new SPL configuration is determining how to find an optimized feature selection that minimizes or maximizes an objective function, such as total cost, subject to resource constraints. To help address the challenges of optimizing feature selection in the face of resource constraints, this paper presents an approach that uses G enetic A lgorithms for optimized FE ature S election (GAFES) in SPLs. Our empirical results show that GAFES can produce solutions with 86–97% of the optimality of other automated feature selection algorithms and in 45–99% less time than existing exact and heuristic feature selection techniques.