Feature Model Synthesis with Genetic Programming (bibtex)
by Lukas Linsbauer, Roberto E. Lopez-Herrejon, Alexander Egyed
Abstract:
Search-Based Software Engineering (SBSE) has proven successful on several stages of the software development life cycle. It has also been applied to different challenges in the context of Software Product Lines (SPLs) like generating minimal test suites. When reverse engineering SPLs from legacy software an important challenge is the reverse engineering of variability, often expressed in the form of Feature Models (FMs). The synthesis of FMs has been studied with techniques such as Genetic Algorithms. In this paper we explore the use of Genetic Programming for this task. We sketch our general workflow, the GP pipeline employed, and its evolutionary operators. We report our experience in synthesizing feature models from sets of feature combinations for 17 representative feature models, and analyze the results using standard information retrieval metrics.
Reference:
Feature Model Synthesis with Genetic Programming (Lukas Linsbauer, Roberto E. Lopez-Herrejon, Alexander Egyed), In Proceedings of the Search-Based Software Engineering - 6th International Symposium, (SSBSE 2014), Fortaleza, Brazil (Claire Le Goues, Shin Yoo, eds.), Springer, volume 8636, 2014.
Bibtex Entry:
@Conference{Linsbauer2014,
  author    = {Lukas Linsbauer and Roberto E. Lopez-Herrejon and Alexander Egyed},
  title     = {Feature Model Synthesis with Genetic Programming},
  booktitle = {Proceedings of the Search-Based Software Engineering - 6th International Symposium, (SSBSE 2014), Fortaleza, Brazil},
  year      = {2014},
  editor    = {Claire Le Goues and Shin Yoo},
  volume    = {8636},
  pages     = {153--167},
  publisher = {Springer},
  abstract  = {Search-Based Software Engineering (SBSE) has proven successful on
	several stages of the software development life cycle. It has also
	been applied to different challenges in the context of Software Product
	Lines (SPLs) like generating minimal test suites. When reverse engineering
	SPLs from legacy software an important challenge is the reverse engineering
	of variability, often expressed in the form of Feature Models (FMs).
	The synthesis of FMs has been studied with techniques such as Genetic
	Algorithms. In this paper we explore the use of Genetic Programming
	for this task. We sketch our general workflow, the GP pipeline employed,
	and its evolutionary operators. We report our experience in synthesizing
	feature models from sets of feature combinations for 17 representative
	feature models, and analyze the results using standard information
	retrieval metrics.},
  doi       = {10.1007/978-3-319-09940-8_11},
  file      = {:Conferences\\SSBSE 2014 - Feature Model Synthesis with Genetic Programming\\Feature Model Synthesis with Genetic Programming-preprint.pdf:PDF},
  keywords  = {FWF P25289, FWF M1421},
}
Powered by bibtexbrowser