2020 · 27 citations · 29 references
ComputingComputer ScienceSystems EngineeringDesign AutomationFeature EngineeringSystems AnalysisConfiguration ManagementSatisfiability Modulo TheoriesAutomatic ConfigurationBoolean SatisfiabilityBoolean Satisfiability Modulo TheoriesConfiguration Management ProcessEmpirical EvaluationFeature ModelsSatisfiabilityVariability Modeling
Handling configurable systems with thousands of configuration options is a challenging problem in research and industry. One of the most common approaches to manage the configuration options of large systems is variability modelling. The verification and configuration process of large variability models is manually infeasible. Hence, they are usually assisted by automated analyses based on solving satisfiability problems (SAT). Recent advances in satisfiability modulo theories (SMT) could prove SMT solvers as a viable alternative to SAT solvers. However, SMT solvers are typically not utilized for variability analyses. A comparison for SAT and SMT could help to estimate SMT solvers potential for the automated analysis. We integrated two SMT solvers into FeatureIDE and compared them against a SAT solver on analyses for feature models, configurations, and realization artifacts. We give an overview of all variability analyses in FeatureIDE and present the results of our empirical evaluation for over 122 systems. We observed that SMT solvers are generally faster in generating explanations of unsatisfiable requests. However, the evaluated SAT solver outperformed SMT solvers for other analyses.
29
Satisfiability modulo theories
Leonardo de Moura, Nikolaj Bjørner · Communications of the ACM · 2011 · 629 citations
Feature-oriented product line engineering
K.C. Kang, Jaejoon Lee, Patrick Donohoe · IEEE Software · 2002 · 507 citations
A survey of variability modeling in industrial practice
Thorsten Berger, Ralf Rublack, D. Nair et al. · 2013 · 395 citations