Evolutionary Consequences of Coevolving Targets

Created by W.Langdon from gp-bibliography.bib Revision:1.4221

  author =       "Ludo Pagie and Paulien Hogeweg",
  title =        "Evolutionary Consequences of Coevolving Targets",
  journal =      "Evolutionary Computation",
  volume =       "5",
  number =       "4",
  pages =        "401--418",
  year =         "1997",
  month =        "Winter",
  keywords =     "genetic algorithms, genetic programming, coevolution,
                 co-evolution, genotype-phenotype mapping, information
                 integration, generalisability, mutational stability,
  URL =          "http://www.santafe.edu/~ludo/articles/ecct.ps.gz",
  URL =          "http://citeseer.ist.psu.edu/pagie98evolutionary.html",
  DOI =          "doi:10.1162/evco.1997.5.4.401",
  size =         "28 pages",
  abstract =     "Most evolutionary optimisation models incorporate a
                 fitness evaluation that is based on a predefined static
                 set of test cases or problems. In the natural
                 evolutionary process, selection is of course not based
                 on a static fitness evaluation. Organisms do not have
                 to combat every existing disease during their lifespan;
                 organisms of one species may live in different or
                 changing environments; different species coevolve. This
                 leads to the question of how information is integrated
                 over many generations.

                 This study focuses on the effects of different fitness
                 evaluation schemes on the types of genotypes and
                 phenotypes that evolve. The evolutionary target is a
                 simple numerical function. The genetic representation
                 is in the form of a program (i.e., a functional
                 representation, as in genetic programming). Many
                 different programs can code for the same numerical
                 function. In other words, there is a many-to-one
                 mapping between genotype (the programs) and phenotypes.
                 We compare fitness evaluation based on a large static
                 set of problems and fitness evaluation based on small
                 coevolving sets of problems. In the latter model very
                 little information is presented to the evolving
                 programs regarding the evolutionary target per
                 evolutionary time step. In other words, the fitness
                 evaluation is very sparse. Nevertheless the model
                 produces correct solutions to the complete evolutionary
                 target in about half of the simulations. The complete
                 evaluation model, on the other hand, does not find
                 correct solutions to the target in any of the
                 simulations. More important, we find that sparse
                 evaluated programs are better generalisable compared to
                 the complete evaluated programs when they are evaluated
                 on a much denser set of problems. In addition, the two
                 evaluation schemes lead to programs that differ with
                 respect to mutational stability; sparse evaluated
                 programs are less stable than complete evaluated
  notes =        "p6 'We embedded the populations of solutions and
                 parasites in space. We used a 2-D toroidal square
                 lattice...' p10 Solutions evolved without coevolution
                 are more difficult to find and of a different form. p11
                 'The complete sampling results in severe selection
                 against the occurrence of errors.' p11 'increased
                 complexity reduces the generalisability of the
                 solutions' p11 Mutational stability p14 'statically
                 evaluated programs are more stable than the' coevolved
                 programs. p15 adds 'much less generalizable'.

                 p20 'the coevolutionary evaluation scheme works best if
                 the population size is of the same order as the size of
                 the complete set of problems.'

                 p20 'coevolving evaluation scheme needs much larger
                 populations... computational cost nevertheless favours
                 the coevolutionary...' p20 'random evaluation scheme'
                 p20 'In the 2-D model the coevolving evaluation scheme
                 is not much more efficient than the random evaluation
                 scheme.' but see following text. p21 Figure 7. p23
                 'Thus the sparseness of the evaluation helps (rather
                 than hinders) the search.'",

Genetic Programming entries for Ludo Pagie Paulien Hogeweg