1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230
|
#!/usr/bin/perl -w
die 'usage: ./shor.pl [number to factor]' unless @ARGV;
use strict;
use warnings;
use Quantum::Entanglement qw(:DEFAULT :complex :QFT);
$Quantum::Entanglement::destroy = 0;
my $num = $ARGV[0];
# do some early die'ing
die "$num is a multiple of two, here I am, brain the size..." unless $num %2;
die "$num is a non-integer, I only have whole numbers of fingers"
unless $num == int($num);
die "$num is less than 15" unless $num >= 15;
print "Performing initial classical steps:\n";
# work out q value
my $q_power = int(2* log($num) / log(2)) +1;
my $q = 2 ** $q_power;
# pick some x so that x is coprime to n.
my $x;
do {
$x = int(rand $num) + 1;
} until ($num % $x != 0 and $x > 2); #ok.. so this misses the point slightly
print "Using q:$q, x:$x\nStarting quantum steps\n";
# fill up a register with integers from 1..q
my $prob = 1/sqrt($q);
my $register1 = entangle(map {$prob, $_} (0..$q-1));
# apply transformation F = x**|a> mod n, store in register 2
# (need to do a p_func to avoid overflow while **)
sub power_mod {
my ($state, $x1, $num1) = @_;
my $rt = 1;
return 1 if $state == 0;
return 1 if $state == 1;
for (1..$state) {
$rt = ($rt * $x1) % $num1;
}
return $rt;
}
print "Performing F = x**|a> mod n\n";
my $register2 = p_func(\&power_mod, $register1, $x, $num);
# We now observe $register2, thus partially collapsing reg1
my $k = "$register2";
print "\$register2 collapsed to $k\n";
print "Finding period of F (this is where you wish for a QCD)\n";
# take a ft of the amplitudes of reg1, placing result in reg3
my $register3 = QFT($register1);
my $lqonr = "$register3"; # observe, this must be multiple of q/r
if ($lqonr == 0) {
print "Got period of '0', halting\n"; exit(0);
}
my $period = int($q / $lqonr + 0.5); # rounding
print "Period of F = x**|a> mod n is $period\n";
# now given the period, we need to work out the factor of n
# work out the two thingies:
if ($period % 2 != 0) {
print "$period is not an even number, doubling to";
$period *=2;
print " $period\n";
}
my $one = $x**($period/2) -1;
my $two = $x**($period/2) +1;
# one and two must have a gcd in common with n, which we now find...
print "$one * $two and $num might share a gcd (classical step)\n";
my ($max1, $max2) = (1,1);
for (2..$num) {
last if $_ > $num;
unless (($num % $_) || ($one % $_)) {
$max1 = $_;
}
unless (($num % $_) || ($two % $_)) {
$max2 = $_;
}
}
print "$max1, $max2 could be factors of $num\n";
__END__;
=head1 NAME
shor - A short demonstration of Quantum::Entanglement
=head1 SYNOPSIS
./shor.pl [number to factor (>14)]
=head1 DESCRIPTION
This program implements Shor's famous algorithm for factoring numbers. A
brief overview of the algorithm is given below.
=head2 The important maths
Given a number B<n> which we are trying to factor, and some other number
which we have guessed, B<x>, we can say that:
x**0 % n == 1 (as x**0 = 1, 1 % n =1)
There will also be some other number, B<r> such that
x**r % n == 1
or, more specifically,
x**(kr) % n ==1
in other words, the function
F(a) = x**a % n
is periodic with period B<r>.
Now, starting from
x**r = 1 % n
x**(2*r/2) = 1 % n
(x**(r/2))**2 - 1 = 0 % n
and, if r is an even number,
(x**(r/2) - 1)*(x**(r/2) + 1) = 0 mod n
or in nice short words, the term on the left is an integer multiple of B<n>.
So long as x**(r/2) != +-1, at least one of the two brackets on the left
must share a factor with B<n>.
Shor's alorithm provides a way to find the periodicity of the function F
and thus a way to calculate two numbers which share a factor with n, it
is then easy to use a classical computer to find the GCD and thus a
factor of B<n>.
=head1 The steps of the algorithm
=head2 1. Remove early trivial cases
We have efficient classical methods for finding that 2 is a factor of 26,
so we do not need to use this method for this.
=head2 2. Pick an integer
Chose a number B<q> so that C<n**2 <= q <= 2n**2>, this is done on a
classical computer. (This is the size we will use for our quantum register.)
=head2 3. Select at random a number coprime to n
Think of some number less than B<n> so that B<n> and B<x> do not share
a common factor (if they do, we already know the answer...).
=head2 4. Fill a quantum register with integers from 0..q-1
This is where we create our first entangled variable, and is the first
non-classical step in this algorithm.
=head2 5. Calculate F, store in a second register
We now calculate C< F(a) = x**a % n> where a represents the superposition
of states in our first register, we store the result of this in our
second register.
=head2 6. Look at register2
We now look at the value of register two and get some value B<k>, this forces
register1 into
a state which can only collapse into values satisfying the equation
x**a % n = k
The probability amplitudes for the remaining states are now all equal to zero,
note that we have not yet looked directly at register1.
=head2 7. Find period of register1
We now apply a fourier transform to the amplitudes of the states in
register1, storing the result as the probability amplitudes for a new
state with the values of register1. This causes there to be a high
probability that the register will collapse to a value which is some
multiple of C<q/r>.
=head2 8. Observe register1
We now observe register1, and use the result to calculate a likely value
for B<r>. From this we can easily calculate two numbers, one of which
will have a factor in common with n, by applying an efficient classical
algoirthm for finding the greatest common denominator, we will be able
to find a value which could be a factor of B<n>.
=head1 Things to remember
This algorithm does not claim to produce a factor of our number the first
time that it is run, there are various conditions which will cause it
to halt mid-way, for instance, the FT step can give a result of 0 which
is clearly useless. The algorithm is better than any known classical one
because the expectation value of the time required to get a correct answer
is still O(n).
This also cannot factor a number which is prime (it being, as it were, prime)
and also cannot factor something which is a prime power (25, say).
=head1 COPYRIGHT
This code is copyright (c) Alex Gough (alex@rcon.org )2001. This is
free software, you may use, modify and redistribute it under the same
terms as Perl itself.
=head1 BUGS
This is slow, being run on classical computers, ah well.
=cut
|