line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
1
|
|
|
1
|
|
23118
|
use strict; #-*-CPerl-*- |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
29
|
|
2
|
1
|
|
|
1
|
|
5
|
use warnings; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
34
|
|
3
|
|
|
|
|
|
|
|
4
|
1
|
|
|
1
|
|
4
|
use lib qw( ../../../lib ); |
|
1
|
|
|
|
|
5
|
|
|
1
|
|
|
|
|
5
|
|
5
|
|
|
|
|
|
|
|
6
|
|
|
|
|
|
|
=encoding utf8 |
7
|
|
|
|
|
|
|
|
8
|
|
|
|
|
|
|
=head1 NAME |
9
|
|
|
|
|
|
|
|
10
|
|
|
|
|
|
|
Algorithm::Evolutionary::Utils - Collection of functions used in all kind of evolutionary algorithms.. |
11
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
=head1 SYNOPSIS |
13
|
|
|
|
|
|
|
|
14
|
|
|
|
|
|
|
use Algorithm::Evolutionary::Utils qw(entropy genotypic_entropy hamming consensus average random_bitstring random_number_array decode_string vector_compare ); |
15
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
my $this_entropy = entropy( $population ); |
17
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
#Computes consensus sequence (for binary chromosomes |
19
|
|
|
|
|
|
|
my $this_consensus = consensus( $population); |
20
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
=head1 DESCRIPTION |
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
Miscellaneous class that contains functions that might be useful |
24
|
|
|
|
|
|
|
somewhere else, especially when computing EA statistics. |
25
|
|
|
|
|
|
|
|
26
|
|
|
|
|
|
|
=cut |
27
|
|
|
|
|
|
|
|
28
|
|
|
|
|
|
|
|
29
|
|
|
|
|
|
|
=head1 METHODS |
30
|
|
|
|
|
|
|
|
31
|
|
|
|
|
|
|
=cut |
32
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
package Algorithm::Evolutionary::Utils; |
34
|
|
|
|
|
|
|
|
35
|
1
|
|
|
1
|
|
106
|
use Exporter; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
78
|
|
36
|
|
|
|
|
|
|
our @ISA = qw(Exporter); |
37
|
|
|
|
|
|
|
|
38
|
1
|
|
|
1
|
|
631
|
use version; our $VERSION = qv("3.403"); |
|
1
|
|
|
|
|
2003
|
|
|
1
|
|
|
|
|
5
|
|
39
|
|
|
|
|
|
|
|
40
|
|
|
|
|
|
|
our @EXPORT_OK = qw( entropy genotypic_entropy consensus hamming |
41
|
|
|
|
|
|
|
random_bitstring random_number_array average |
42
|
|
|
|
|
|
|
parse_xml decode_string vector_compare); |
43
|
|
|
|
|
|
|
|
44
|
1
|
|
|
1
|
|
117
|
use Carp; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
82
|
|
45
|
1
|
|
|
1
|
|
679
|
use String::Random; |
|
1
|
|
|
|
|
4172
|
|
|
1
|
|
|
|
|
72
|
|
46
|
1
|
|
|
1
|
|
298
|
use XML::Parser; |
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
47
|
|
|
|
|
|
|
use Statistics::Basic qw(mean); |
48
|
|
|
|
|
|
|
|
49
|
|
|
|
|
|
|
=head2 entropy( $population) |
50
|
|
|
|
|
|
|
|
51
|
|
|
|
|
|
|
Computes the entropy using the well known Shannon's formula: L |
52
|
|
|
|
|
|
|
'to avoid botching highlighting |
53
|
|
|
|
|
|
|
|
54
|
|
|
|
|
|
|
=cut |
55
|
|
|
|
|
|
|
|
56
|
|
|
|
|
|
|
sub entropy { |
57
|
|
|
|
|
|
|
my $population = shift; |
58
|
|
|
|
|
|
|
my %frequencies; |
59
|
|
|
|
|
|
|
map( (defined $_->{'_fitness'})?$frequencies{$_->{'_fitness'}}++:1, @$population ); |
60
|
|
|
|
|
|
|
my $entropy = 0; |
61
|
|
|
|
|
|
|
my $gente = scalar(@$population); # Population size |
62
|
|
|
|
|
|
|
for my $f ( keys %frequencies ) { |
63
|
|
|
|
|
|
|
my $this_freq = $frequencies{$f}/$gente; |
64
|
|
|
|
|
|
|
$entropy -= $this_freq*log( $this_freq ); |
65
|
|
|
|
|
|
|
} |
66
|
|
|
|
|
|
|
return $entropy; |
67
|
|
|
|
|
|
|
} |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
=head2 genotypic_entropy( $population) |
70
|
|
|
|
|
|
|
|
71
|
|
|
|
|
|
|
Computes the entropy using the well known Shannon's formula: |
72
|
|
|
|
|
|
|
L 'to avoid botching |
73
|
|
|
|
|
|
|
highlighting; in this case we use chromosome frequencies instead of |
74
|
|
|
|
|
|
|
fitness. |
75
|
|
|
|
|
|
|
|
76
|
|
|
|
|
|
|
=cut |
77
|
|
|
|
|
|
|
|
78
|
|
|
|
|
|
|
sub genotypic_entropy { |
79
|
|
|
|
|
|
|
my $population = shift; |
80
|
|
|
|
|
|
|
my %frequencies; |
81
|
|
|
|
|
|
|
map( $frequencies{$_->{'_str'}}++, @$population ); |
82
|
|
|
|
|
|
|
my $entropy = 0; |
83
|
|
|
|
|
|
|
my $gente = scalar(@$population); # Population size |
84
|
|
|
|
|
|
|
for my $f ( keys %frequencies ) { |
85
|
|
|
|
|
|
|
my $this_freq = $frequencies{$f}/$gente; |
86
|
|
|
|
|
|
|
$entropy -= $this_freq*log( $this_freq ); |
87
|
|
|
|
|
|
|
} |
88
|
|
|
|
|
|
|
return $entropy; |
89
|
|
|
|
|
|
|
} |
90
|
|
|
|
|
|
|
|
91
|
|
|
|
|
|
|
=head2 hamming( $string_a, $string_b ) |
92
|
|
|
|
|
|
|
|
93
|
|
|
|
|
|
|
Computes the number of bit positions that are different among two strings, the well known Hamming distance. |
94
|
|
|
|
|
|
|
|
95
|
|
|
|
|
|
|
=cut |
96
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
sub hamming { |
98
|
|
|
|
|
|
|
my ($string_a, $string_b) = @_; |
99
|
|
|
|
|
|
|
return ( ( $string_a ^ $string_b ) =~ tr/\1//); |
100
|
|
|
|
|
|
|
} |
101
|
|
|
|
|
|
|
|
102
|
|
|
|
|
|
|
=head2 consensus( $population, $rough = 0 ) |
103
|
|
|
|
|
|
|
|
104
|
|
|
|
|
|
|
Consensus sequence representing the majority value for each bit; |
105
|
|
|
|
|
|
|
returns the consensus binary string. If "rough", then the bit is set only if the |
106
|
|
|
|
|
|
|
difference is bigger than 0.2 (60/40 proportion). Otherwise, it is set to C<-> |
107
|
|
|
|
|
|
|
|
108
|
|
|
|
|
|
|
=cut |
109
|
|
|
|
|
|
|
|
110
|
|
|
|
|
|
|
sub consensus { |
111
|
|
|
|
|
|
|
my $population = shift; |
112
|
|
|
|
|
|
|
my $rough = shift; |
113
|
|
|
|
|
|
|
my @frequencies; |
114
|
|
|
|
|
|
|
for ( @$population ) { |
115
|
|
|
|
|
|
|
for ( my $i = 0; $i < length($_->{'_str'}); $i ++ ) { |
116
|
|
|
|
|
|
|
if ( !$frequencies[$i] ) { |
117
|
|
|
|
|
|
|
$frequencies[$i]={ 0 => 0, |
118
|
|
|
|
|
|
|
1 => 0}; |
119
|
|
|
|
|
|
|
} |
120
|
|
|
|
|
|
|
$frequencies[$i]->{substr($_->{'_str'}, $i, 1)}++; |
121
|
|
|
|
|
|
|
} |
122
|
|
|
|
|
|
|
} |
123
|
|
|
|
|
|
|
my $consensus; |
124
|
|
|
|
|
|
|
for my $f ( @frequencies ) { |
125
|
|
|
|
|
|
|
if ( !$rough ) { |
126
|
|
|
|
|
|
|
if ( $f->{'0'} > $f->{'1'} ) { |
127
|
|
|
|
|
|
|
$consensus.='0'; |
128
|
|
|
|
|
|
|
} else { |
129
|
|
|
|
|
|
|
$consensus.='1'; |
130
|
|
|
|
|
|
|
} |
131
|
|
|
|
|
|
|
} else { |
132
|
|
|
|
|
|
|
my $total = $f->{'0'} + $f->{'1'}; |
133
|
|
|
|
|
|
|
my $difference = (abs( $f->{'0'} - $f->{'1'} ))/$total; |
134
|
|
|
|
|
|
|
if ( $difference < 0.2 ) { |
135
|
|
|
|
|
|
|
$consensus .= '-'; |
136
|
|
|
|
|
|
|
} else { |
137
|
|
|
|
|
|
|
if ( $f->{'0'} > $f->{'1'} ) { |
138
|
|
|
|
|
|
|
$consensus.='0'; |
139
|
|
|
|
|
|
|
} else { |
140
|
|
|
|
|
|
|
$consensus.='1'; |
141
|
|
|
|
|
|
|
} |
142
|
|
|
|
|
|
|
} |
143
|
|
|
|
|
|
|
} |
144
|
|
|
|
|
|
|
} |
145
|
|
|
|
|
|
|
return $consensus; |
146
|
|
|
|
|
|
|
} |
147
|
|
|
|
|
|
|
|
148
|
|
|
|
|
|
|
=head2 average( $population ) |
149
|
|
|
|
|
|
|
|
150
|
|
|
|
|
|
|
Computes an average of population fitness |
151
|
|
|
|
|
|
|
|
152
|
|
|
|
|
|
|
=cut |
153
|
|
|
|
|
|
|
|
154
|
|
|
|
|
|
|
sub average { |
155
|
|
|
|
|
|
|
my $population = shift; |
156
|
|
|
|
|
|
|
my @frequencies; |
157
|
|
|
|
|
|
|
my @fitnesses = map( $_->{'_fitness'}, @$population ); |
158
|
|
|
|
|
|
|
return mean( @fitnesses ); |
159
|
|
|
|
|
|
|
|
160
|
|
|
|
|
|
|
} |
161
|
|
|
|
|
|
|
|
162
|
|
|
|
|
|
|
=head2 random_bitstring( $bits[, $chromify = 0 ] ) |
163
|
|
|
|
|
|
|
|
164
|
|
|
|
|
|
|
Returns a random bitstring with the stated number of bits. Useful for |
165
|
|
|
|
|
|
|
testing, mainly. Returns a "chromosome-like" object if $chromify is 1, |
166
|
|
|
|
|
|
|
mainly putting it into a hashref with the C<_str> key. |
167
|
|
|
|
|
|
|
|
168
|
|
|
|
|
|
|
=cut |
169
|
|
|
|
|
|
|
|
170
|
|
|
|
|
|
|
sub random_bitstring { |
171
|
|
|
|
|
|
|
my $bits = shift || croak "No bits!"; |
172
|
|
|
|
|
|
|
my $chromify = shift || 0; |
173
|
|
|
|
|
|
|
my $generator = new String::Random; |
174
|
|
|
|
|
|
|
my $regex = "\[01\]{$bits}"; |
175
|
|
|
|
|
|
|
my $this_string = $generator->randregex($regex); |
176
|
|
|
|
|
|
|
return $chromify?{_str => $this_string}:$this_string; |
177
|
|
|
|
|
|
|
} |
178
|
|
|
|
|
|
|
|
179
|
|
|
|
|
|
|
=head2 random_number_array( $dimensions [, $min = -1] [, $range = 2] ) |
180
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
Returns a random number array with the stated length. Useful for testing, mainly. |
182
|
|
|
|
|
|
|
|
183
|
|
|
|
|
|
|
=cut |
184
|
|
|
|
|
|
|
|
185
|
|
|
|
|
|
|
sub random_number_array { |
186
|
|
|
|
|
|
|
my $dimensions = shift || croak "Null dimension!"; |
187
|
|
|
|
|
|
|
my $min = shift || -1; |
188
|
|
|
|
|
|
|
my $range = shift || 2; |
189
|
|
|
|
|
|
|
|
190
|
|
|
|
|
|
|
my @array; |
191
|
|
|
|
|
|
|
for ( my $i = 0; $i < $dimensions; $i ++ ) { |
192
|
|
|
|
|
|
|
push @array, $min + rand($range); |
193
|
|
|
|
|
|
|
} |
194
|
|
|
|
|
|
|
return @array; |
195
|
|
|
|
|
|
|
} |
196
|
|
|
|
|
|
|
|
197
|
|
|
|
|
|
|
|
198
|
|
|
|
|
|
|
=head2 decode_string( $chromosome, $gene_size, $min, $range ) |
199
|
|
|
|
|
|
|
|
200
|
|
|
|
|
|
|
Decodes to a vector, each one of whose components ranges between $min |
201
|
|
|
|
|
|
|
and $max. Returns that vector. |
202
|
|
|
|
|
|
|
|
203
|
|
|
|
|
|
|
It does not work for $gene_size too big. Certainly not for 64, maybe for 32. |
204
|
|
|
|
|
|
|
|
205
|
|
|
|
|
|
|
=cut |
206
|
|
|
|
|
|
|
|
207
|
|
|
|
|
|
|
sub decode_string { |
208
|
|
|
|
|
|
|
my ( $chromosome, $gene_size, $min, $range ) = @_; |
209
|
|
|
|
|
|
|
|
210
|
|
|
|
|
|
|
my @output_vector; |
211
|
|
|
|
|
|
|
my $max_range = eval "0b"."1"x$gene_size; |
212
|
|
|
|
|
|
|
for (my $i = 0; $i < length($chromosome)/$gene_size; $i ++ ) { |
213
|
|
|
|
|
|
|
my $substr = substr( $chromosome, $i*$gene_size, $gene_size ); |
214
|
|
|
|
|
|
|
push @output_vector, (($range - $min) * eval("0b$substr") / $max_range) + $min; |
215
|
|
|
|
|
|
|
} |
216
|
|
|
|
|
|
|
return @output_vector; |
217
|
|
|
|
|
|
|
} |
218
|
|
|
|
|
|
|
|
219
|
|
|
|
|
|
|
=head2 vector_compare( $vector_1, $vector_2 ) |
220
|
|
|
|
|
|
|
|
221
|
|
|
|
|
|
|
Compares vectors, returns 1 if 1 dominates 2, -1 if it's the other way |
222
|
|
|
|
|
|
|
round, and 0 if neither dominates the other. Both vectors are supposed |
223
|
|
|
|
|
|
|
to be numeric. Returns C<0> if neither is bigger, and they are not |
224
|
|
|
|
|
|
|
equal. Fails if the length is not the same. None of the combinations |
225
|
|
|
|
|
|
|
above, returns C. |
226
|
|
|
|
|
|
|
|
227
|
|
|
|
|
|
|
=cut |
228
|
|
|
|
|
|
|
|
229
|
|
|
|
|
|
|
sub vector_compare { |
230
|
|
|
|
|
|
|
my ( $vector_1, $vector_2 ) = @_; |
231
|
|
|
|
|
|
|
|
232
|
|
|
|
|
|
|
if ( scalar @$vector_1 != scalar @$vector_2 ) { |
233
|
|
|
|
|
|
|
croak "Different lengths, can't compare\n"; |
234
|
|
|
|
|
|
|
} |
235
|
|
|
|
|
|
|
|
236
|
|
|
|
|
|
|
my $length = scalar @$vector_1; |
237
|
|
|
|
|
|
|
my @results = map( $vector_1->[$_] <=> $vector_2->[$_], 0..($length-1)); |
238
|
|
|
|
|
|
|
my %comparisons; |
239
|
|
|
|
|
|
|
map( $comparisons{$_}++, @results ); |
240
|
|
|
|
|
|
|
if ( $comparisons{1} && !$comparisons{-1} ) { |
241
|
|
|
|
|
|
|
return 1; |
242
|
|
|
|
|
|
|
} |
243
|
|
|
|
|
|
|
if ( !$comparisons{1} && $comparisons{-1} ) { |
244
|
|
|
|
|
|
|
return -1; |
245
|
|
|
|
|
|
|
} |
246
|
|
|
|
|
|
|
if ( defined $comparisons{0} && $comparisons{0} == $length ) { |
247
|
|
|
|
|
|
|
return 0; |
248
|
|
|
|
|
|
|
} |
249
|
|
|
|
|
|
|
return undef; |
250
|
|
|
|
|
|
|
} |
251
|
|
|
|
|
|
|
|
252
|
|
|
|
|
|
|
=head1 SEE ALSO |
253
|
|
|
|
|
|
|
|
254
|
|
|
|
|
|
|
This is a spin off from L so it's worth the while to check it out. And the spinning was due to finding I needed to include it in examples for the much simpler L. |
255
|
|
|
|
|
|
|
|
256
|
|
|
|
|
|
|
=head1 Copyright |
257
|
|
|
|
|
|
|
|
258
|
|
|
|
|
|
|
This file is released under the GPL. See the LICENSE file included in this distribution, |
259
|
|
|
|
|
|
|
or go to http://www.fsf.org/licenses/gpl.txt |
260
|
|
|
|
|
|
|
|
261
|
|
|
|
|
|
|
=cut |
262
|
|
|
|
|
|
|
|
263
|
|
|
|
|
|
|
"Still there?"; |