Solution to Project Euler problem 75

db4
Aaron Schaefer 2008-02-02 17:22:20 -05:00
parent a0dad18f4f
commit 8b207d1f48
4 changed files with 98 additions and 16 deletions

View File

@ -1,7 +1,7 @@
! Copyright (c) 2008 Aaron Schaefer.
! See http://factorcode.org/license.txt for BSD license.
USING: arrays combinators.lib kernel math math.matrices math.ranges namespaces
sequences ;
USING: arrays combinators.lib kernel math math.ranges namespaces
project-euler.common sequences ;
IN: project-euler.039
! http://projecteuler.net/index.php?section=problems&id=39
@ -21,6 +21,7 @@ IN: project-euler.039
! --------
! Algorithm adapted from http://mathworld.wolfram.com/PythagoreanTriple.html
! Identical implementation as problem #75
! Basically, this makes an array of 1000 zeros, recursively creates primitive
! triples using the three transforms and then increments the array at index
@ -39,18 +40,6 @@ SYMBOL: p-count
max-p 1- over <range> p-count get
[ [ 1+ ] change-nth ] curry each ;
: transform ( triple matrix -- new-triple )
[ 1array ] dip m. first ;
: u-transform ( triple -- new-triple )
{ { 1 2 2 } { -2 -1 -2 } { 2 2 3 } } transform ;
: a-transform ( triple -- new-triple )
{ { 1 2 2 } { 2 1 2 } { 2 2 3 } } transform ;
: d-transform ( triple -- new-triple )
{ { -1 -2 -2 } { 2 1 2 } { 2 2 3 } } transform ;
: (count-perimeters) ( seq -- )
dup sum max-p < [
dup sum adjust-p-count

View File

@ -0,0 +1,78 @@
! Copyright (c) 2008 Aaron Schaefer.
! See http://factorcode.org/license.txt for BSD license.
USING: arrays combinators.lib kernel math math.ranges namespaces
project-euler.common sequences ;
IN: project-euler.075
! http://projecteuler.net/index.php?section=problems&id=75
! DESCRIPTION
! -----------
! It turns out that 12 cm is the smallest length of wire can be bent to form a
! right angle triangle in exactly one way, but there are many more examples.
! 12 cm: (3,4,5)
! 24 cm: (6,8,10)
! 30 cm: (5,12,13)
! 36 cm: (9,12,15)
! 40 cm: (8,15,17)
! 48 cm: (12,16,20)
! In contrast, some lengths of wire, like 20 cm, cannot be bent to form a right
! angle triangle, and other lengths allow more than one solution to be found;
! for example, using 120 cm it is possible to form exactly three different
! right angle triangles.
! 120 cm: (30,40,50), (20,48,52), (24,45,51)
! Given that L is the length of the wire, for how many values of L ≤ 1,000,000
! can exactly one right angle triangle be formed?
! SOLUTION
! --------
! Algorithm adapted from http://mathworld.wolfram.com/PythagoreanTriple.html
! Identical implementation as problem #39
! Basically, this makes an array of 1000000 zeros, recursively creates
! primitive triples using the three transforms and then increments the array at
! index [a+b+c] by one for each triple's sum AND its multiples under 1000000
! (to account for non-primitive triples). The answer is just the number of
! indexes that equal one.
SYMBOL: p-count
<PRIVATE
: max-p ( -- n )
p-count get length ;
: adjust-p-count ( n -- )
max-p 1- over <range> p-count get
[ [ 1+ ] change-nth ] curry each ;
: (count-perimeters) ( seq -- )
dup sum max-p < [
dup sum adjust-p-count
[ u-transform ] keep [ a-transform ] keep d-transform
[ (count-perimeters) ] 3apply
] [
drop
] if ;
: count-perimeters ( n -- )
0 <array> p-count set { 3 4 5 } (count-perimeters) ;
PRIVATE>
: euler075 ( -- answer )
[
1000000 count-perimeters p-count get [ 1 = ] count
] with-scope ;
! [ euler075 ] 100 ave-time
! 1873 ms run / 123 ms GC ave time - 100 trials
MAIN: euler075

View File

@ -1,5 +1,6 @@
USING: arrays combinators.lib kernel math math.functions math.miller-rabin
math.parser math.primes.factors math.ranges namespaces sequences sorting ;
math.matrices math.parser math.primes.factors math.ranges namespaces
sequences sorting ;
IN: project-euler.common
! A collection of words used by more than one Project Euler solution
@ -16,6 +17,7 @@ IN: project-euler.common
! propagate-all - #18, #67
! sum-proper-divisors - #21
! tau* - #12
! [uad]-transform - #39, #75
: nth-pair ( n seq -- nth next )
@ -45,6 +47,9 @@ IN: project-euler.common
dup perfect-square? [ sqrt >fixnum neg , ] [ drop ] if
] { } make sum ;
: transform ( triple matrix -- new-triple )
[ 1array ] dip m. first ;
PRIVATE>
: cartesian-product ( seq1 seq2 -- seq1xseq2 )
@ -101,3 +106,12 @@ PRIVATE>
dup sqrt >fixnum [1,b] [
dupd mod zero? [ [ 2 + ] dip ] when
] each drop * ;
! These transforms are for generating primitive Pythagorean triples
: u-transform ( triple -- new-triple )
{ { 1 2 2 } { -2 -1 -2 } { 2 2 3 } } transform ;
: a-transform ( triple -- new-triple )
{ { 1 2 2 } { 2 1 2 } { 2 2 3 } } transform ;
: d-transform ( triple -- new-triple )
{ { -1 -2 -2 } { 2 1 2 } { 2 2 3 } } transform ;

View File

@ -12,7 +12,8 @@ USING: definitions io io.files kernel math math.parser project-euler.ave-time
project-euler.029 project-euler.030 project-euler.031 project-euler.032
project-euler.033 project-euler.034 project-euler.035 project-euler.036
project-euler.037 project-euler.038 project-euler.039 project-euler.067
project-euler.134 project-euler.169 project-euler.173 project-euler.175 ;
project-euler.075 project-euler.134 project-euler.169 project-euler.173
project-euler.175 ;
IN: project-euler
<PRIVATE