NAME

Algorithm::Evolutionary - Perl module for performing paradigm-free evolutionary algorithms.

SYNOPSIS

  #Short way of loading a lot of modules, POE-style
  use Algorithm::Evolutionary qw( Op::This_Operator
                                  Individual::That_Individual
                                  Fitness::Some_Fitness); 
  

DESCRIPTION

Algorithm::Evolutionary is a set of classes for doing object-oriented evolutionary computation in Perl. Why would anyone want to do that escapes my knowledge, but, in fact, we have found it quite useful for our own purposes. Same as Perl itself.

The design principle of Algorithm::Evolutionary is flexibility: it should be very easy to extend using this library, and it should be also quite easy to program what's already there in the evolutionary computation community. Besides, the library classes should have persistence provided by XML modules, and, in some cases, YAML.

The algorithm allows to create simple evolutionary algorithms, as well as more complex ones, that interface with databases or with the web.

The project is hosted at Sourceforge . Latest aditions, and nightly updates, can be downloaded from there before they are uploaded to CPAN. That page also hosts the mailing list, as well as bug reports, news, updates, whatever.

In case the examples are hidden somewhere in the .cpan directory, you can also download them from the CVS repository, and the -examples tarballs in the file download area of that repository

You can also get help from the project forum (where I usually hang around or from the mailing list (information at its web site

I have used this continously for my research all these year, and any search will return a number of papers; a journal article is already submitted, but meanwhile if you use it for any of your research, I would be very grateful if you quoted papers such as this one:

 @InProceedings{jj:2008:PPSN,
   author =     "Juan J. Merelo and  Antonio M. Mora and Pedro A. Castillo
                 and Juan L. J. Laredo and Lourdes Araujo and Ken C. Sharman
                 and Anna I. Esparcia-Alcázar and Eva Alfaro-Cid
                 and Carlos Cotta",
   title =      "Testing the Intermediate Disturbance Hypothesis: Effect of
                 Asynchronous Population Incorporation on Multi-Deme
                 Evolutionary Algorithms",
   booktitle =  "Parallel Problem Solving from Nature - PPSN X",
   year =       "2008",
   editor =     "Gunter Rudolph and Thomas Jansen and Simon Lucas and
                  Carlo Poloni and Nicola Beume",
   volume =     "5199",
   series =     "LNCS",
   pages =      "266-275",
   address =    "Dortmund",
   month =      "13-17 " # sep,
   publisher =  "Springer",
   keywords =   "genetic algorithms, genetic programming, p2p computing",
   ISBN =       "3-540-87699-5",
   doi =        "10.1007/978-3-540-87700-4_27",
   size =       "pages",
   notes =      "PPSN X",
 }

Some information on this paper and instructions for downloading the code used in it can be found in our group blog

DISCUSSION, FEATURE REQUESTS

Head to the CPAN forum for this module: http://www.cpanforum.com/dist/Algorithm-Evolutionary

BUGS?

Have you found any bugs? Use the CPAN tracker to inform about them (http://rt.cpan.org/Public/Dist/Display.html?Name=Algorithm-Evolutionary) or email the author (below).

AUTHOR

Main author and developer is J. J. Merelo, jmerelo (at) geneura.ugr.es who blogs at BloJJ and twitters at twitter.com/jmerelo. There have also been some contributions from Javi García, fjgc (at) decsai.ugr.es and Pedro Castillo, pedro (at) geneura.ugr.es. Patient users that have submitted bugs include jamarier. Bugs, requests and any kind of comment are welcome.

Examples

There are a few examples in the examples subdirectory, which should have been included with your CPAN bundle. For instance, check out tide_float.pl, an example of floating point vector optimization, or cd examples; run_easy_ga.pl p_peaks.yaml, which should run an example of a simple GA on the P_Peaks deceptive function.

Some other examples are installed: check out tide_bitstring.pl, tide_float.pl and canonical-genetic-algorithm.pl, which you can run and play with to get a taste of what EA programming is like, and then ammend, add and modify at leisure to create your own evolutionary algorithms.

SEE ALSO

Algorithm::Evolutionary::Op::Base.

Algorithm::Evolutionary::Individual::Base.

Algorithm::Evolutionary::Fitness::Base.

Algorithm::Evolutionary::Experiment.

Algorithm::Evolutionary::CanonicalGA.

XML for an explanation of the XML format used

POE::Component::Algorithm::Evolutionary if you want to mix evolutionary algorithms with anything else easily

You might be interested in one of the other GA modules out there, such as AI::Genetic::Pro

Copyright

  This file is released under the GPL. See the LICENSE file included in this distribution,
  or go to http://www.fsf.org/licenses/gpl.txt

  CVS Info: $Date: 2009/07/22 10:45:36 $ 
  $Header: /cvsroot/opeal/Algorithm-Evolutionary/lib/Algorithm/Evolutionary.pm,v 2.17 2009/07/22 10:45:36 jmerelo Exp $ 
  $Author: jmerelo $ 
  $Revision: 2.17 $

cpanm Algorithm::Evolutionary

perl -MCPAN -e shell
install Algorithm::Evolutionary