Fitness Causes Bloat

Created by W.Langdon from gp-bibliography.bib Revision:1.4504

  author =       "W. B. Langdon and R. Poli",
  title =        "Fitness Causes Bloat",
  institution =  "University of Birmingham, School of Computer Science",
  address =      "Birmingham, B15 2TT, UK",
  number =       "CSRP-97-09",
  month =        "24 " # feb,
  year =         "1997",
  keywords =     "genetic algorithms, genetic programming",
  file =         "/1997/",
  URL =          "",
  abstract =     "The problem of evolving an artificial ant to follow
                 the Santa Fe trail is used to demonstrate the well
                 known genetic programming feature of growth in solution
                 length. Known variously as ``bloat'', ``redundancy'',
                 ``introns'', ``fluff'', ``Structural Complexity'' with
                 antonyms ``parsimony'', ``Minimum Description Length''
                 (MDL) and ``Occam's razor''. Comparison with runs with
                 and without fitness selection pressure shows the
                 tendency for solutions to grow in size is caused by
                 fitness based selection. We argue that such growth is
                 inherent in using a fixed evaluation function with a
                 discrete but variable length representation. Since with
                 simple static evaluation search converges to mainly
                 finding trial solutions with the same fitness as
                 existing trial solutions. In general variable length
                 allows many more long representations of a given
                 solution than short ones of the same solution. Thus
                 with an unbiased random search we expect longer
                 representations to occur more often and so
                 representation length tends to increase. I.e. fitness
                 based selection leads to bloat.",
  size =         "16 pages",

Genetic Programming entries for William B Langdon Riccardo Poli