Anomaly detection in Context-aware Feature Models
This addresses a domain-specific problem for software engineering by improving anomaly detection in feature models, though it is incremental as it builds on existing methods like HyVarRec.
The paper tackles the problem of detecting anomalies in Context-aware Feature Models, which is challenging due to contextual influences, and shows that using Quantified Boolean Formula (QBF) solvers outperforms common techniques like iterative SAT solver calls.
Feature Models are a mechanism to organize the configuration space and facilitate the construction of software variants by describing configuration options using features, i.e., a name representing a functionality. The development of Feature Models is an error prone activity and detecting their anomalies is a challenging and important task needed to promote their usage. Recently, Feature Models have been extended with context to capture the correlation of configuration options with contextual influences and user customizations. Unfortunately, this extension makes the task of detecting anomalies harder. In this paper, we formalize the anomaly analysis in Context-aware Feature Models and we show how Quantified Boolean Formula (QBF) solvers can be used to detect anomalies without relying on iterative calls to a SAT solver. By extending the reconfigurator engine HyVarRec, we present findings evidencing that QBF solvers can outperform the common techniques for anomaly analysis.