Algorithm-Evolve

 view release on metacpan or  search on metacpan

Changes  view on Meta::CPAN

Revision history for Perl extension Algorithm::Evolve.

0.03	
	- Update selection/replacement dispatch method to play well with DProf,
	  may change back someday..
	- Speedup of A::E::Util::str_agreement at the price of utf8
	- Other misc speedups
	- Change StringEvolver to use blessed hash, for better reuse prospects
	- Added ArrayEvolver, mimicking StringEvolver for array genes
	- Added "Breeding Perls" example

0.02	Mon May 19 2003
	- Added gladitorial selection/replacement
	- Added co-evolving rock-paper-scissors example
	- Lots of POD updates, added POD to StringEvolver.pm
	- Added array utils to Algorithm::Evolve::Util
	- Some switches can be changed dynamically (selection, replacement,
	  parents & children per generation)

examples/ArrayEvolver.pm  view on Meta::CPAN

        mutation_rate  => 0.05,
        crossover_pts  => 2,
        @_
    );
}

sub new {
    my $pkg = shift;
    my $array = shift
        || arr_random($configs{gene_length}, $configs{alphabet});
    return bless { _gene => $array }, $pkg;
}

sub crossover {
    my ($pkg, $c1, $c2) = @_;
    return map { $pkg->new($_) } 
           arr_crossover($c1->gene, $c2->gene, $configs{crossover_pts});
}

sub fitness {
    my $self = shift;

examples/StringEvolver.pm  view on Meta::CPAN

        mutation_rate  => 0.05,
        crossover_pts  => 2,
        @_
    );
}

sub new {
    my $pkg = shift;
    my $string = shift
        || str_random($configs{gene_length}, $configs{alphabet});
    return bless { _gene => $string }, $pkg;
}

sub crossover {
    my ($pkg, $s1, $s2) = @_;
    return map { $pkg->new($_) } 
           str_crossover($s1->gene, $s2->gene, $configs{crossover_pts});
}

sub fitness {
    my $self = shift;

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

###########################

sub debug {
    print @_, "\n" if $DEBUG;
}

sub new {
    my $pkg = shift;

    my $p = bless {
        generations      => 0,
        parents_per_gen  => 2,
        @_
    }, $pkg;
   
    $p->{random_seed}      ||= int(rand $rand_max);
    srand( $p->random_seed );

    $p->{selection}        ||= $p->{replacement};
    $p->{replacement}      ||= $p->{selection};

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

Algorithm::Evolve maintains a population of critter objects to be evolved. You 
may evolve any type of objects you want, provided the class supplies the 
following methods:

=over

=item C<Class-E<gt>new()>

This method will be called as a class method with no arguments. It must return
a blessed critter object. It is recommended that the returned critter's genes
be randomly initialized.

=item C<Class-E<gt>crossover( $critter1, $critter2 )>

This method will also be called as a class method, with two critter objects as 
arguments. It should return a list of two new critter objects based on the 
genes of the passed objects.

=item C<$critter-E<gt>mutate()>