Analysis of Cartesian Genetic Programming's Evolutionary Mechanisms

Created by W.Langdon from gp-bibliography.bib Revision:1.4549

  author =       "Brian W. Goldman and William F. Punch",
  title =        "Analysis of Cartesian Genetic Programming's
                 Evolutionary Mechanisms",
  journal =      "IEEE Transactions on Evolutionary Computation",
  year =         "2015",
  volume =       "19",
  number =       "3",
  pages =        "359--373",
  keywords =     "genetic algorithms, genetic programming, Cartesian
                 Genetic Programming",
  ISSN =         "1089-778X",
  DOI =          "doi:10.1109/TEVC.2014.2324539",
  size =         "15 pages",
  abstract =     "Understanding how search operators interact with
                 solution representation is a critical step to improving
                 search. In Cartesian Genetic Programming (CGP), and
                 Genetic Programming (GP) in general, the complex
                 genotype to phenotype map makes achieving this
                 understanding a challenge. By examining aspects such as
                 tuned parameter values, the search quality of CGP
                 variants at different problem difficulties, node
                 behaviour, and offspring replacement properties we seek
                 to better understand the characteristics of CGP search.
                 Our focus is twofold: creating methods to prevent
                 wasted CGP evaluations (Skip, Accumulate, and Single)
                 and creating methods to overcome CGP's search
                 limitations imposed by genome ordering (Reorder and
                 DAG). Our results on Boolean problems show CGP evolves
                 genomes that are highly inactive, very redundant, and
                 full of seemingly useless constants. On some tested
                 problems we found less than 1percent of the genome was
                 actually required to encode the evolved solution.
                 Furthermore, traditional CGP ordering results in large
                 portions of the genome that are never used by any
                 ancestor of the evolved solution. Reorder and DAG allow
                 evolution to use the entire genome. More generally, our
                 results suggest that Skip-Reorder and Single-Reorder
                 are most likely to solve hard problems using the least
                 number of evaluations and the least amount of time
                 while better avoiding degenerate behaviour.",
  notes =        "also known as \cite{6815728}",

Genetic Programming entries for Brian W Goldman William F Punch