Patterns for Constructing Mutation Operators: Limiting the Search Space in a Software Engineering Application

Created by W.Langdon from gp-bibliography.bib Revision:1.3872

  author =       "Thomas Kuehne and Heiko Hamann and 
                 Svetlana Arifulina and Gregor Engels",
  title =        "Patterns for Constructing Mutation Operators: Limiting
                 the Search Space in a Software Engineering
  booktitle =    "EuroGP 2016: Proceedings of the 19th European
                 Conference on Genetic Programming",
  year =         "2016",
  month =        "30 " # mar # "--1 " # apr,
  editor =       "Malcolm I. Heywood and James McDermott and 
                 Mauro Castelli and Ernesto Costa and Kevin Sim",
  series =       "LNCS",
  volume =       "9594",
  publisher =    "Springer Verlag",
  address =      "Porto, Portugal",
  pages =        "278--293",
  organisation = "EvoStar",
  keywords =     "genetic algorithms, genetic programming",
  isbn13 =       "978-3-319-30668-1",
  DOI =          "doi:10.1007/978-3-319-30668-1_18",
  abstract =     "We apply methods of genetic programming to a general
                 problem from software engineering, namely example-based
                 generation of specifications. In particular, we focus
                 on model transformation by example. The definition and
                 implementation of model transformations is a task
                 frequently carried out by domain experts, hence, a
                 (semi-)automatic approach is desirable. This
                 application is challenging because the underlying
                 search space has rich semantics, is high-dimensional,
                 and unstructured. Hence, a computationally brute-force
                 approach would be unscalable and potentially
                 infeasible. To address that problem, we develop a
                 sophisticated approach of designing complex mutation
                 operators. We define patterns for constructing mutation
                 operators and report a successful case study.
                 Furthermore, the code of the evolved model
                 transformation is required to have high maintainability
                 and extensibility, that is, the code should be easily
                 readable by domain experts. We report an evaluation of
                 this approach in a software engineering case study.",
  notes =        "Part of \cite{Heywood:2016:GP} EuroGP'2016 held in
                 conjunction with EvoCOP2016, EvoMusArt2016 and

Genetic Programming entries for Thomas Kuehne Heiko Hamann Svetlana Arifulina Gregor Engels