Automated Analysis of Feature Models Using Atomic Sets.
Sergio Segura · 2008
Scalability is recognized as a key challenge in the automated analysis of Feature Models (FMs). Current solutions in this context mainly propose using different logic paradigms as a way to improve the performance at the solution level while the problem remains the same. Atomic Sets (ASs) were proposed as a promising solution for the simplification of FMs (i.e. reduction of the number of variables) in the context of automated analysis. However, years after their introduction, the lack of specific algorithms and performance results still hinder its integration into current proposals and tools. In this paper, we set the basis for the usage of ASs as a generic technique for the automated analysis of FMs. In particular, we first propose a specific algorithm to construct the ASs of an FM. Then, we present a performance test measuring the degree of improvement (in time and memory) when implementing ASs into CSP, BDD and SAT-based solutions.