view release on metacpan or  search on metacpan

README  view on Meta::CPAN

This module is intended to be a useful tool for quick and easy implementation
of evolutionary algorithms. It aims to be flexible, yet simple. For this
reason, it is not a comprehensive implementation of all possible evolutionary
algorithm configurations. The flexibility of Perl allows the evolution of
any type of object conceivable: a simple string or array, a deeper structure
like a hash of arrays, or even a complex object like graph object from another
CPAN module, etc. 

It's also worth mentioning that evolutionary algorithms are generally very
CPU-intensive. There are a great deal of calls to C<rand()> and a lot of
associated floating-point math. If you want a lightning-fast framework, then
searching CPAN at all is probably a bad place to start. However, this doesn't
mean that I've ignored efficiency. The fitness function is often the biggest
bottleneck.

INSTALLATION

To install this module type the following:

   perl Makefile.PL
   make

lib/Algorithm/Evolve.pm  view on Meta::CPAN

}

##########################################
##########################################
##########################################
1;
__END__

=head1 NAME

Algorithm::Evolve - An extensible and generic framework for executing 
evolutionary algorithms

=head1 SYNOPSIS

    #!/usr/bin/perl -w
    use Algorithm::Evolve;
    use MyCritters;     ## Critter class providing appropriate methods
    
    sub callback {
        my $p = shift;  ## get back the population object

lib/Algorithm/Evolve.pm  view on Meta::CPAN

This module is intended to be a useful tool for quick and easy implementation
of evolutionary algorithms. It aims to be flexible, yet simple. For this
reason, it is not a comprehensive implementation of all possible evolutionary
algorithm configurations. The flexibility of Perl allows the evolution of
any type of object conceivable: a simple string or array, a deeper structure
like a hash of arrays, or even something as complex as graph object from
another CPAN module, etc. 

It's also worth mentioning that evolutionary algorithms are generally very
CPU-intensive. There are a great deal of calls to C<rand()> and a lot of
associated floating-point math. If you want a lightning-fast framework, then
searching CPAN at all is probably a bad place to start. However, this doesn't
mean that I've ignored efficiency. The fitness function is often the biggest
bottleneck.

=head2 Framework Overview

The configurable parts of an evolutionary algorithm can be split up into two 
categories:

=over