"be consistent" PerlMonks

### Divide an array into 2 subsets to verify their sum is equal or not.

 on May 02, 2013 at 07:09 UTC Need Help??
bimleshsharma has asked for the wisdom of the Perl Monks concerning the following question:

I have an array having some interger elements. I have to split into 2 subsets. Need to verify that sum of these 2 subsets are equal or not. Subset size is not matter, it can be element count of 3,2 or 1,4 or any but should be in 2 part. For example...

Ex 1: @a1 = qw(1, 3, 8, 4); This array can be divided into to subsets (1,3,4)=8 and (8)=8, so these two have equal sum.

Ex 2: @a1 = qw(1, 6, 2); This cant be divided into two subsets of equal sum.Because (1,6)!=2 or any combination of a subset is not matching to another subset.

I tried below code but it is working for one set of iteration, actually it should check all possible set of iteration to find possiblities.

```my @array= qw(1 3 5 7);
my @array= qw(1 3 5 7);
&test(\@array);

sub test
{
my (\$s1,\$s2);
my @a=@{\$_[0]};
for (my \$i=0;\$i<=\$#a ;\$i++) {
(\$s1,\$s2)=0;
for (my \$j=0;\$j<=\$#a ;\$j++) {
if (\$i == \$j) {
\$s1=\$s1+ \$a[\$j];
}
else{
\$s2+=\$a[\$j];
}
}

}
print "\n";
}
• Comment on Divide an array into 2 subsets to verify their sum is equal or not.

Replies are listed 'Best First'.
Re: Divide an array into 2 subsets to verify their sum is equal or not.
by choroba (Bishop) on May 02, 2013 at 07:50 UTC
To sum a list, I used the sum from List::Util. I made the following observation: if the array can be split, then the sum of each part is the half of the sum of the whole array. I used vec to generate binary vectors to be used as Indicator function. Checking a half of the possible vectors is enough, the rest is complementary (i.e. the two subsets are swapped).
```#!/usr/bin/perl
use warnings;
use strict;

use List::Util qw(sum);

sub is_divisible {
my \$array = shift;
my \$sum   = sum(@\$array) / 2;

for my \$bitmask (1 .. 2 ** \$#\$array - 1) {
return 1 if sum(map { \$array->[\$_] * vec \$bitmask, \$_, 1} 0 ..
+ \$#\$array) == \$sum;
}

return;
}

my @arrays = (
[qw(1 3 5 7)],
[qw(1 3 8 4)],
[qw(1 6 2)],
[qw(5 5 4 6 2 8 1 9)],
);

for my \$array (@arrays) {
print "@\$array: ", is_divisible(\$array) ? 'yes' : 'no', "\n";
}
لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ
Re: Divide an array into 2 subsets to verify their sum is equal or not.
by BrowserUk (Pope) on May 02, 2013 at 09:51 UTC

choroba's Explore-all-possible-combinations mechanism works okay for smallish sets (max:32 or 64 depending upon your Perl), but will get very slow for anything much larger than 20 or so.

This will very quickly (less than 0.001 of a second) find a solution, if one exists, for sets of 100s or 1000s of elements. :

```#! perl -slw
use strict;
use Time::HiRes qw[ time ];
use List::Util qw[ sum ];

sub partition {
my \$sum = sum @_;
return if \$sum & 1;
\$sum /= 2;
my @s = sort{ \$b <=> \$a } @_;
my @a;
my( \$t, \$n ) = ( 0, -1 );
\$t + \$s[\$n] <= \$sum and \$t+= \$s[\$n] and push @a, \$n while ++\$n < @
+s and \$t <= \$sum;
@a = delete @s[ @a ];
@s = grep defined, @s;
return unless sum( @a ) == sum( @s );
return \@a, \@s;
}

our \$N //= 64;

my( \$a, \$b ) = partition 1,3,5,7;
print "sum( @{ \$a } ) == sum( @{ \$b } )" if \$a;

my @set = map int( rand 100 ), 1 .. \$N;

my \$start = time;
( \$a, \$b ) = partition @set;
printf "Took %f seconds\n", time() - \$start;

if( \$a ) {
printf "(%u) == sum( @{ \$a } ) == sum( @{ \$b } )\n", sum @\$a;
}
else {
print "No solution existed for the \$N element set @set";
}

A few runs:

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
It gives a wrong answer for
```2, 12, 4
لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ

Corrected. Thanks.

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
Re: Divide an array into 2 subsets to verify their sum is equal or not.
by kcott (Chancellor) on May 02, 2013 at 10:35 UTC

G'day bimleshsharma,

Update:

Update2: Original solution substantially rewritten! It had various problems:

Update3: Fixed a bug and added some features:

Update4: Fixed some bugs and changed volume testing.

• Display issue with subset output. Duplicate data due to "@a2 ... @a2" not being "@a1 ... @a2". - FIXED.
• while loop in check_arrays() had flaws. This has been pretty much rewritten. - FIXED.
• Volume testing was decimal-based, now it's octal-based. Previously, --volume_power_max=3 [or --vpm=3] created arrays of up to 10**3 (1,000) elements; now, the value is 8**3 (512) elements. Decimal-based was a bad choice as neither 1 .. 10 nor 1 .. 100 can be split into two equal portions.
• Added a few more tests.

Here's Update4's version of pm_split_equal_sums.pl:

```#!/usr/bin/env perl -l

use strict;
use warnings;

use List::Util qw{first sum};
use Test::More;
use Time::HiRes qw{time};
use Getopt::Long;

my %opt = (
test_more   => 1,
time_hires  => 1,
volume_tests        => 0,
volume_power_max    => 3,
array_limit => 3,
);

GetOptions(map {
join('|' => @{[join '' => /(?>^|_)([a-z])/gi]}, \$_) . ':i' => \\$op
+t{\$_}
} keys %opt);

my \$test_equal_subsets = [
[1, 3, 8, 4],
[1, 3, 5, 7],
[4, 3, 2, 2, 1],
[4, 3, 2, 2, 2, 2, 1],
[5, 5, 4, 6, 2, 8, 1, 9],
[8, 4, 4, 7, 6, 3],
[1, 1],
[2, 2],
[],
[0],
[0, 0],
[0, 0, 0],
[0, 0, 0, 0],
[ (1) x 100 ],
[ 1 .. 1000 ],
];

my \$test_unequal_subsets = [
[1, 6, 2],
[7, 5, 3, 3],
[1, 2 ,3, 7],
[0, 1],
[1, 2],
[1],
[2],
[8, 1, 2, 3],
[ 1 .. 10 ],
[ 1 .. 100 ],
];

if (\$opt{volume_tests}) {
for (1 .. \$opt{volume_power_max}) {
my @volume = map { ((\$_), (\$_)) } 1 .. 8**\$_ / 2;
push @\$test_equal_subsets, [@volume];
push @\$test_unequal_subsets, [@volume, 8**(2 * \$_)];
}
}

if (\$opt{test_more}) {
plan tests => scalar @\$test_equal_subsets + scalar @\$test_unequal_
+subsets;
}

my @expectations = ('Not expecting equal subsets.', 'Expecting equal s
+ubsets.');
my @subsets_data = ([\$test_unequal_subsets, 0, 0], [\$test_equal_subset
+s, 1, 1]);

for (@subsets_data) {
my (\$subsets, \$expect_code, \$expect_name_index) = @\$_;
my \$expect_name = \$expectations[\$expect_name_index];

for (@\$subsets) {
my \$start = time if \$opt{time_hires};

if (\$opt{test_more}) {
is(check_arrays(\$_), \$expect_code, \$expect_name);
}
else {
check_arrays(\$_);
}

printf "Took %f seconds\n", time() - \$start if \$opt{time_hires
+};
}
}

sub check_arrays {
my \$full_array = shift;

print 'Checking: (', array_string(\$full_array), ')';

if (! grep { \$_ } @\$full_array) {
print "\tSubsets: (", array_string(\$full_array), ') and ()';
print "\tSubset sum = 0";
return 1;
}

my \$full_sum = sum @\$full_array;

if (\$full_sum % 2) {
print "\tSubsets not equal: sum of starting array is odd (\$ful
+l_sum).";
return 0;
}

my \$half_sum = \$full_sum / 2;
my @sorted_array = sort { \$b % 2 <=> \$a % 2 || \$b <=> \$a } @\$full_
+array;

if (my \$big = first { \$_ > \$half_sum } @sorted_array) {
print "\tSubsets not equal: element (\$big) larger than sum of
+rest.";
return 0;
}

my (@a1, @a2);
my \$total = 0;

while (@sorted_array) {
push @a1, shift @sorted_array;
\$total += \$a1[\$#a1];

@sorted_array = map { \$total + \$_ <= \$half_sum
? do { push @a1, \$_; \$total += \$_; () } : \$_
} @sorted_array;

if (\$total == \$half_sum) {
(@a2, @sorted_array) = (@a2, @sorted_array);
}
else {
push @a2, pop @a1 if @a1;
}
}

if (\$total == \$half_sum) {
print "\tSubsets: (", array_string([sort { \$a <=> \$b } @a1]),
+')';
print "\t     and (", array_string([sort { \$a <=> \$b } @a2]),
+')';
print "\tSubset sum = \$half_sum";
return 1;
}
else {
print "\tSubsets not equal: no solution found.";
return 0
}
}

sub array_string {
my \$array = shift;

return join(', ' => @\$array > 3 * \$opt{array_limit}
? ( @\$array[0 .. \$opt{array_limit} - 1],
" ... [snip: @{[@\$array - 2 * \$opt{array_limit}]} elements
+] ...",
@\$array[@\$array - \$opt{array_limit} .. \$#\$array] )
: @\$array);
}

Here's a test run. Note that this uses --vpm=8 and final volume test "Took 89.489836 seconds" — you might want to start with a lower value.

```\$ pm_split_equal_sums.pl --tm=1 --th=1 --vt=1 --vpm=8 --al=3
1..41
Checking: (1, 6, 2)
Subsets not equal: sum of starting array is odd (9).
ok 1 - Not expecting equal subsets.
Took 0.000318 seconds
Checking: (7, 5, 3, 3)
Subsets not equal: no solution found.
ok 2 - Not expecting equal subsets.
Took 0.000233 seconds
Checking: (1, 2, 3, 7)
Subsets not equal: sum of starting array is odd (13).
ok 3 - Not expecting equal subsets.
Took 0.000169 seconds
Checking: (0, 1)
Subsets not equal: sum of starting array is odd (1).
ok 4 - Not expecting equal subsets.
Took 0.000165 seconds
Checking: (1, 2)
Subsets not equal: sum of starting array is odd (3).
ok 5 - Not expecting equal subsets.
Took 0.000167 seconds
Checking: (1)
Subsets not equal: sum of starting array is odd (1).
ok 6 - Not expecting equal subsets.
Took 0.000165 seconds
Checking: (2)
Subsets not equal: element (2) larger than sum of rest.
ok 7 - Not expecting equal subsets.
Took 0.000177 seconds
Checking: (8, 1, 2, 3)
Subsets not equal: element (8) larger than sum of rest.
ok 8 - Not expecting equal subsets.
Took 0.000179 seconds
Checking: (1, 2, 3,  ... [snip: 4 elements] ..., 8, 9, 10)
Subsets not equal: sum of starting array is odd (55).
ok 9 - Not expecting equal subsets.
Took 0.000179 seconds
Checking: (1, 2, 3,  ... [snip: 94 elements] ..., 98, 99, 100)
Subsets not equal: no solution found.
ok 10 - Not expecting equal subsets.
Took 0.000817 seconds
Checking: (1, 1, 2, 2, 3, 3, 4, 4, 64)
Subsets not equal: element (64) larger than sum of rest.
ok 11 - Not expecting equal subsets.
Took 0.000193 seconds
Checking: (1, 1, 2,  ... [snip: 59 elements] ..., 32, 32, 4096)
Subsets not equal: element (4096) larger than sum of rest.
ok 12 - Not expecting equal subsets.
Took 0.000379 seconds
Checking: (1, 1, 2,  ... [snip: 507 elements] ..., 256, 256, 262144)
Subsets not equal: element (262144) larger than sum of rest.
ok 13 - Not expecting equal subsets.
Took 0.000769 seconds
Checking: (1, 1, 2,  ... [snip: 4091 elements] ..., 2048, 2048, 167772
+16)
Subsets not equal: element (16777216) larger than sum of rest.
ok 14 - Not expecting equal subsets.
Took 0.004773 seconds
Checking: (1, 1, 2,  ... [snip: 32763 elements] ..., 16384, 16384, 107
+3741824)
Subsets not equal: element (1073741824) larger than sum of rest.
ok 15 - Not expecting equal subsets.
Took 0.035878 seconds
Checking: (1, 1, 2,  ... [snip: 262139 elements] ..., 131072, 131072,
+68719476736)
Subsets not equal: element (68719476736) larger than sum of rest.
ok 16 - Not expecting equal subsets.
Took 0.294454 seconds
Checking: (1, 1, 2,  ... [snip: 2097147 elements] ..., 1048576, 104857
+6, 4398046511104)
Subsets not equal: element (4398046511104) larger than sum of rest
+.
ok 17 - Not expecting equal subsets.
Took 2.395664 seconds
Checking: (1, 1, 2,  ... [snip: 16777211 elements] ..., 8388608, 83886
+08, 281474976710656)
Subsets not equal: element (281474976710656) larger than sum of re
+st.
ok 18 - Not expecting equal subsets.
Took 19.422507 seconds
Checking: (1, 3, 8, 4)
Subsets: (1, 3, 4)
and (8)
Subset sum = 8
ok 19 - Expecting equal subsets.
Took 0.000223 seconds
Checking: (1, 3, 5, 7)
Subsets: (1, 7)
and (3, 5)
Subset sum = 8
ok 20 - Expecting equal subsets.
Took 0.000201 seconds
Checking: (4, 3, 2, 2, 1)
Subsets: (1, 2, 3)
and (2, 4)
Subset sum = 6
ok 21 - Expecting equal subsets.
Took 0.000203 seconds
Checking: (4, 3, 2, 2, 2, 2, 1)
Subsets: (1, 3, 4)
and (2, 2, 2, 2)
Subset sum = 8
ok 22 - Expecting equal subsets.
Took 0.000208 seconds
Checking: (5, 5, 4, 6, 2, 8, 1, 9)
Subsets: (1, 5, 5, 9)
and (2, 4, 6, 8)
Subset sum = 20
ok 23 - Expecting equal subsets.
Took 0.000209 seconds
Checking: (8, 4, 4, 7, 6, 3)
Subsets: (3, 6, 7)
and (4, 4, 8)
Subset sum = 16
ok 24 - Expecting equal subsets.
Took 0.000206 seconds
Checking: (1, 1)
Subsets: (1)
and (1)
Subset sum = 1
ok 25 - Expecting equal subsets.
Took 0.000197 seconds
Checking: (2, 2)
Subsets: (2)
and (2)
Subset sum = 2
ok 26 - Expecting equal subsets.
Took 0.000195 seconds
Checking: ()
Subsets: () and ()
Subset sum = 0
ok 27 - Expecting equal subsets.
Took 0.000168 seconds
Checking: (0)
Subsets: (0) and ()
Subset sum = 0
ok 28 - Expecting equal subsets.
Took 0.000168 seconds
Checking: (0, 0)
Subsets: (0, 0) and ()
Subset sum = 0
ok 29 - Expecting equal subsets.
Took 0.000168 seconds
Checking: (0, 0, 0)
Subsets: (0, 0, 0) and ()
Subset sum = 0
ok 30 - Expecting equal subsets.
Took 0.000170 seconds
Checking: (0, 0, 0, 0)
Subsets: (0, 0, 0, 0) and ()
Subset sum = 0
ok 31 - Expecting equal subsets.
Took 0.000181 seconds
Checking: (1, 1, 1,  ... [snip: 94 elements] ..., 1, 1, 1)
Subsets: (1, 1, 1,  ... [snip: 44 elements] ..., 1, 1, 1)
and (1, 1, 1,  ... [snip: 44 elements] ..., 1, 1, 1)
Subset sum = 50
ok 32 - Expecting equal subsets.
Took 0.000323 seconds
Checking: (1, 2, 3,  ... [snip: 994 elements] ..., 998, 999, 1000)
Subsets: (1, 3, 5,  ... [snip: 495 elements] ..., 995, 997, 999)
and (2, 4, 6,  ... [snip: 493 elements] ..., 996, 998, 1000)
Subset sum = 250250
ok 33 - Expecting equal subsets.
Took 0.001866 seconds
Checking: (1, 1, 2, 2, 3, 3, 4, 4)
Subsets: (1, 1, 2, 3, 3)
and (2, 4, 4)
Subset sum = 10
ok 34 - Expecting equal subsets.
Took 0.000212 seconds
Checking: (1, 1, 2,  ... [snip: 58 elements] ..., 31, 32, 32)
Subsets: (1, 1, 3,  ... [snip: 27 elements] ..., 29, 31, 31)
and (2, 2, 4,  ... [snip: 25 elements] ..., 30, 32, 32)
Subset sum = 528
ok 35 - Expecting equal subsets.
Took 0.000310 seconds
Checking: (1, 1, 2,  ... [snip: 506 elements] ..., 255, 256, 256)
Subsets: (1, 1, 3,  ... [snip: 251 elements] ..., 253, 255, 255)
and (2, 2, 4,  ... [snip: 249 elements] ..., 254, 256, 256)
Subset sum = 32896
ok 36 - Expecting equal subsets.
Took 0.001029 seconds
Checking: (1, 1, 2,  ... [snip: 4090 elements] ..., 2047, 2048, 2048)
Subsets: (1, 1, 3,  ... [snip: 2043 elements] ..., 2045, 2047, 204
+7)
and (2, 2, 4,  ... [snip: 2041 elements] ..., 2046, 2048, 204
+8)
Subset sum = 2098176
ok 37 - Expecting equal subsets.
Took 0.006883 seconds
Checking: (1, 1, 2,  ... [snip: 32762 elements] ..., 16383, 16384, 163
+84)
Subsets: (1, 1, 3,  ... [snip: 16379 elements] ..., 16381, 16383,
+16383)
and (2, 2, 4,  ... [snip: 16377 elements] ..., 16382, 16384,
+16384)
Subset sum = 134225920
ok 38 - Expecting equal subsets.
Took 0.054265 seconds
Checking: (1, 1, 2,  ... [snip: 262138 elements] ..., 131071, 131072,
+131072)
Subsets: (1, 1, 3,  ... [snip: 131067 elements] ..., 131069, 13107
+1, 131071)
and (2, 2, 4,  ... [snip: 131065 elements] ..., 131070, 13107
+2, 131072)
Subset sum = 8590000128
ok 39 - Expecting equal subsets.
Took 0.449057 seconds
Checking: (1, 1, 2,  ... [snip: 2097146 elements] ..., 1048575, 104857
+6, 1048576)
Subsets: (1, 1, 3,  ... [snip: 1048571 elements] ..., 1048573, 104
+8575, 1048575)
and (2, 2, 4,  ... [snip: 1048569 elements] ..., 1048574, 104
+8576, 1048576)
Subset sum = 549756338176
ok 40 - Expecting equal subsets.
Took 3.644046 seconds
Checking: (1, 1, 2,  ... [snip: 16777210 elements] ..., 8388607, 83886
+08, 8388608)
Subsets: (1, 1, 3,  ... [snip: 8388603 elements] ..., 8388605, 838
+8607, 8388607)
and (2, 2, 4,  ... [snip: 8388601 elements] ..., 8388606, 838
+8608, 8388608)
Subset sum = 35184376283136
ok 41 - Expecting equal subsets.
Took 89.489836 seconds

-- Ken

This is an NP-complete problem

Only true if the OP was looking for an optimum solution. He isn't:

Subset size is not matter,

Behooves you to read the actual question.

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
Applies to BrowserUK's solution as well.

Look again :)

```#! perl -slw
use strict;
use Time::HiRes qw[ time ];
use List::Util qw[ sum ];

sub partition {
my \$sum = sum @_;
return if \$sum & 1;
\$sum /= 2;
my @s = sort{ \$b <=> \$a } @_;
my @a;
my( \$t, \$n ) = ( 0, -1 );
\$t + \$s[\$n] <= \$sum and \$t+= \$s[\$n] and push @a, \$n while ++\$n < @
+s and \$t <= \$sum;
@a = delete @s[ @a ];
@s = grep defined, @s;
return unless sum( @a ) == sum( @s );
return \@a, \@s;
}

my \$start = time;
my( \$a, \$b ) = partition 8, 4, 4, 7, 6, 3;
my @set = map int( rand 100 ), 1 .. \$N;
printf "Took %f seconds\n", time() - \$start;
if( \$a ) {
printf "(%u) == sum( @{ \$a } ) == sum( @{ \$b } )\n", sum @\$a;
}
else {
print "No solution existed for 8, 4, 4, 7, 6, 3";
}
__END__
No solution existed for 8, 4, 4, 7, 6, 3
Took 0.000258 seconds

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
div
> This is an NP-complete problem http://en.wikipedia.org/wiki/Knapsack_problem

Yes, to be precise a sub class known as "Partition Problem".

See WP article for some efficient algorithms and further links.

I wonder who and why is posting well known scientific problems w/o references ...?

Cheers Rolf

( addicted to the Perl Programming Language)

++ Thanks. I've rewritten the solution to handle those sorts of cases.

-- Ken

[ 7, 5, 3, 3]

;^)

...roboticus

When your only tool is a hammer, all problems look like your thumb.

++ Thanks. That node was something of a minor disaster. I've rewritten the solution.

-- Ken

A couple of questions:

1. Why pass in a reference if the first thing you are going to do is copy the reference array to a local array?
```sub check_arrays {
my @full_array = @{shift()};
2. Why make a local copy of the array at all, when all the uses (join, sum, sort) of it require you to pass a list?

Ie. Why not my \$sum = sum @\$ref; etc.

3. Isn't re-summing your partial array over and over wildly inefficient?

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.

++ Thanks. All your points are perfectly valid. There were other issues as well. I've substantially rewritten the solution.

-- Ken

Re: Divide an array into 2 subsets to verify their sum is equal or not.
by hdb (Monsignor) on May 02, 2013 at 09:47 UTC

Using recursion:

```use strict;
use warnings;
use List::Util qw/sum/;

# finds one solution
sub findsum {
my (\$target, @array) = @_;
while( @array ) {
my \$cand = shift @array;
return () if \$cand > \$target;
return ( \$cand ) if \$cand==\$target;
my @sol = findsum( \$target-\$cand, @array );
return (\$cand, @sol) if @sol;
}
return ();
}

my @array = qw(1 3 5 7);
my \$total = sum(@array);

die "Odd total \$total cannot be split!\n" if \$total % 2;

my @sol = findsum( \$total/2, sort @array );
if( @sol ) {
print "Solution: ",join( ",", @sol), "\n";
} else {
print "No solution.";
}

Re: Divide an array into 2 subsets to verify their sum is equal or not.
by davido (Cardinal) on May 03, 2013 at 16:10 UTC

Here's my try, using Algorithm::Bucketizer:

```use strict;
use warnings;
use Algorithm::Bucketizer;
use List::Util qw( sum );
use POSIX qw( ceil );

my @arrays = (
[ qw( 1 3 8 4 ) ],
[ qw( 1 6 2   ) ],
[ qw( 1 3 5 7 ) ],
);

foreach my \$array ( @arrays ) {
print "( @{\$array} ) can ",
can_evenly_distribute( @{\$array} ) ? '' : 'not ',
"be evenly distributed.\n";
}

sub can_evenly_distribute {
my @elements = @_;
my \$b_size = ceil( sum( @elements ) / 2 );

my \$b = Algorithm::Bucketizer->new(
bucketsize => \$b_size,
algorithm  => 'retry'
);

\$b->add_item( \$_, \$_ ) foreach @elements;

\$b->optimize( algorithm => 'random', maxrounds => @elements * 10 );

my @buckets = \$b->buckets;

return    @buckets == 2
&& sum( \$buckets[0]->items ) == sum( \$buckets[1]->items );

}

I haven't found any cases where it fails to return the correct answer... but having just said that, someone will probably find one. ;)

Dave

Create A New User
Node Status?
node history
Node Type: perlquestion [id://1031699]
Approved by Corion
Front-paged by kcott
help
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others surveying the Monastery: (6)
As of 2018-12-11 10:16 GMT
Sections?
Information?
Find Nodes?
Leftovers?
Voting Booth?
How many stories does it take before you've heard them all?

Results (53 votes). Check out past polls.

Notices?