<< problem 70 - Totient permutation | Counting fractions - problem 72 >> |
Problem 71: Ordered fractions
(see projecteuler.net/problem=71)
Consider the fraction, dfrac{n}{d}, where n and d are positive integers. If n<d and HCF(n,d)=1, it is called a reduced proper fraction.
If we list the set of reduced proper fractions for d <= 8 in ascending order of size, we get:
dfrac{1}{8}, dfrac{1}{7}, dfrac{1}{6}, dfrac{1}{5}, dfrac{1}{4}, dfrac{2}{7}, dfrac{1}{3}, dfrac{3}{8}, dfrac{2}{5}, dfrac{3}{7}, dfrac{1}{2}, dfrac{4}{7}, dfrac{3}{5}, dfrac{5}{8}, dfrac{2}{3}, dfrac{5}{7}, dfrac{3}{4}, dfrac{4}{5}, dfrac{5}{6}, dfrac{6}{7}, dfrac{7}{8}
It can be seen that dfrac{2}{5} is the fraction immediately to the left of dfrac{3}{7}.
By listing the set of reduced proper fractions for d <= 1,000,000 in ascending order of size, find the numerator of the fraction immediately to the left of dfrac{3}{7}.
My Algorithm
While researching several approaches I read something about Farey sequences (see en.wikipedia.org/wiki/Farey_sequence):
If you know two neighboring fractions dfrac{a}{b} and dfrac{c}{d}, then there exists a so-called mediant m=dfrac{a+c}{b+d}.
dfrac{0}{1} and dfrac{1}{1} are neighboring fractions because the fulfil the condition bc - ad = 1.
My program starts with these two fractions. In each iteration it checks whether the mediant is less or greater than the user input (3/7 by default).
- m < 3/7 → continue searching in interval a/b and m
- m < 3/7 → continue searching in interval m and c/d
- if denominator of m exceeds 1000000 then stop
interval 0/1 and 1/1 => (0+1)/(1+1) = 1/2 which is larger than 3/7
interval 0/1 and 1/2 => (0+1)/(1+2) = 1/3 which is smaller than 3/7
interval 1/3 and 1/2 => (1+1)/(3+2) = 2/5 which is smaller than 3/7
interval 2/5 and 1/2 => (2+1)/(5+2) = 3/7 which is equal to 3/7 (treat the same as if larger)
interval 2/5 and 3/7 => (2+3)/(5+7) = 7/12 ... and so on ...
The algorithm finds the correct result but might be too slow.
Therefore, I stop whenever I my right interval border becomes a/b (which is 3/7 in the example) because after that only the left interval border can change.
Then a much faster explicit computation is possible:
- each iteration would adjust the left border:
leftN += rightN
andleftD += rightD
untilleftN + leftD > limit
. - there would be
1 + (limit - (leftD + rightD)) / rightD
iterations
isLess
that compares two fractions and returns true
if dfrac{a}{b} < dfrac{c}{d}:if a/b < c/d (and all numbers are positive) then ad < cb
This can be done with precise integer arithmetic. However, Hackerrank's denominators are up to 10^15 and therefore the product exceeds 64 bits.
The template
multiply
is a portable way to multiply two 64 bits and return the 128 bit result as two 64 bit variables.Note: GCC supports
__int128
and there is no need to have such a template on Linux/Apple/MinGW.
Interactive test
You can submit your own input to my program and it will be instantly processed at my server:
This is equivalent toecho "1 3 7 8" | ./71
Output:
Note: the original problem's input 1 3 7 1000000
cannot be entered
because just copying results is a soft skill reserved for idiots.
(this interactive test is still under development, computations will be aborted after one second)
My code
… was written in C++11 and can be compiled with G++, Clang++, Visual C++. You can download it, too.
#include <iostream>
// multiply two unsigned number where the result may exceed the largest native data type
template <typename T>
void multiply(T a, T b, T& result_high, T& result_low)
{
const T Shift = 4 * sizeof(a);
const T Mask = T(~0) >> Shift;
auto a_high = a >> Shift;
auto a_low = a & Mask;
auto b_high = b >> Shift;
auto b_low = b & Mask;
auto c_0 = a_low * b_low;
auto c_1a = a_high * b_low;
auto c_1b = a_low * b_high;
auto c_2 = a_high * b_high;
auto carry = ((c_0 >> Shift) + (c_1a & Mask) + (c_1b & Mask)) >> Shift;
result_high = c_2 + (c_1a >> Shift) + (c_1b >> Shift) + carry;
result_low = c_0 + (c_1a << Shift) + (c_1b << Shift);
}
// if a/b < c/d (and all numbers are positive)
// then a*d < c*b
bool isLess(unsigned long long a, unsigned long long b, unsigned long long c, unsigned long long d)
{
// GCC has 128-bit support:
//return (unsigned __int128)x1 * y2 < (unsigned __int128)y1 * x2;
unsigned long long r1_high, r1_low;
unsigned long long r2_high, r2_low;
multiply(a, d, r1_high, r1_low);
multiply(c, b, r2_high, r2_low);
// compare high bits
if (r1_high < r2_high)
return true;
if (r1_high > r2_high)
return false;
// compare low bits
return (r1_low < r2_low);
}
int main()
{
unsigned int tests = 1;
std::cin >> tests;
while (tests--)
{
unsigned int a = 3;
unsigned int b = 7;
unsigned long long limit = 1000000;
std::cin >> a >> b >> limit;
// generate all numbers in the Farey sequence using binary subdivision
// for each step decide whether the number was left or right of the desired fraction
// start with 0/1 and 1/1
unsigned long long leftN = 0;
unsigned long long leftD = 1;
unsigned long long rightN = 1;
unsigned long long rightD = 1;
while (leftD + rightD <= limit)
{
auto mediantN = leftN + rightN;
auto mediantD = leftD + rightD;
// if x1/y1 < x2/y2 (and all numbers are positive)
// then x1*y2 < x2*y1
if (isLess(mediantN, mediantD, a, b))
{
// adjust left border
leftN = mediantN;
leftD = mediantD;
}
else
{
// adjust right border
rightN = mediantN;
rightD = mediantD;
// did we "hit the spot" ?
if (rightN == a && rightD == b)
break;
}
}
// now the right border is the fraction we read from input
// and we only have to adjust the left border from here on
//while (leftD + rightD <= limit) // gets the job done, but still slow ...
//{
// leftN += rightN;
// leftD += rightD;
//}
// "instant" result
if (limit >= leftD + rightD)
{
auto difference = limit - (leftD + rightD);
auto repeat = 1 + difference / rightD;
leftN += repeat * rightN;
leftD += repeat * rightD;
}
std::cout << leftN << " " << leftD << std::endl;
}
return 0;
}
This solution contains 14 empty lines, 23 comments and 1 preprocessor command.
Benchmark
The correct solution to the original Project Euler problem was found in less than 0.01 seconds on an Intel® Core™ i7-2600K CPU @ 3.40GHz.
(compiled for x86_64 / Linux, GCC flags: -O3 -march=native -fno-exceptions -fno-rtti -std=gnu++11 -DORIGINAL
)
See here for a comparison of all solutions.
Note: interactive tests run on a weaker (=slower) computer. Some interactive tests are compiled without -DORIGINAL
.
Changelog
March 13, 2017 submitted solution
May 2, 2017 added comments
Hackerrank
see https://www.hackerrank.com/contests/projecteuler/challenges/euler071
My code solves 21 out of 21 test cases (score: 100%)
Difficulty
Project Euler ranks this problem at 10% (out of 100%).
Hackerrank describes this problem as easy.
Note:
Hackerrank has strict execution time limits (typically 2 seconds for C++ code) and often a much wider input range than the original problem.
In my opinion, Hackerrank's modified problems are usually a lot harder to solve. As a rule thumb: brute-force is rarely an option.
Links
projecteuler.net/thread=71 - the best forum on the subject (note: you have to submit the correct solution first)
Code in various languages:
C# www.mathblog.dk/project-euler-71-proper-fractions-ascending-order/ (written by Kristian Edlund)
Python github.com/hughdbrown/Project-Euler/blob/master/euler-071.py (written by Hugh Brown)
Python github.com/nayuki/Project-Euler-solutions/blob/master/python/p071.py (written by Nayuki)
C++ github.com/HaochenLiu/My-Project-Euler/blob/master/071.cpp (written by Haochen Liu)
C++ github.com/Meng-Gen/ProjectEuler/blob/master/71.cc (written by Meng-Gen Tsai)
Java github.com/dcrousso/ProjectEuler/blob/master/PE071.java (written by Devin Rousso)
Java github.com/nayuki/Project-Euler-solutions/blob/master/java/p071.java (written by Nayuki)
Go github.com/frrad/project-euler/blob/master/golang/Problem071.go (written by Frederick Robinson)
Mathematica github.com/nayuki/Project-Euler-solutions/blob/master/mathematica/p071.mathematica (written by Nayuki)
Mathematica github.com/steve98654/ProjectEuler/blob/master/071.nb
Haskell github.com/nayuki/Project-Euler-solutions/blob/master/haskell/p071.hs (written by Nayuki)
Haskell github.com/roosephu/project-euler/blob/master/71.hs (written by Yuping Luo)
Clojure github.com/guillaume-nargeot/project-euler-clojure/blob/master/src/project_euler/problem_071.clj (written by Guillaume Nargeot)
Clojure github.com/rm-hull/project-euler/blob/master/src/euler071.clj (written by Richard Hull)
Scala github.com/samskivert/euler-scala/blob/master/Euler071.scala (written by Michael Bayne)
Rust github.com/gifnksm/ProjectEulerRust/blob/master/src/bin/p071.rs
Those links are just an unordered selection of source code I found with a semi-automatic search script on Google/Bing/GitHub/whatever.
You will probably stumble upon better solutions when searching on your own. Maybe not all linked resources produce the correct result and/or exceed time/memory limits.
Heatmap
Please click on a problem's number to open my solution to that problem:
green | solutions solve the original Project Euler problem and have a perfect score of 100% at Hackerrank, too | |
yellow | solutions score less than 100% at Hackerrank (but still solve the original problem easily) | |
gray | problems are already solved but I haven't published my solution yet | |
blue | solutions are relevant for Project Euler only: there wasn't a Hackerrank version of it (at the time I solved it) or it differed too much | |
orange | problems are solved but exceed the time limit of one minute or the memory limit of 256 MByte | |
red | problems are not solved yet but I wrote a simulation to approximate the result or verified at least the given example - usually I sketched a few ideas, too | |
black | problems are solved but access to the solution is blocked for a few days until the next problem is published | |
[new] | the flashing problem is the one I solved most recently |
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 | 231 | 232 | 233 | 234 | 235 | 236 | 237 | 238 | 239 | 240 | 241 | 242 | 243 | 244 | 245 | 246 | 247 | 248 | 249 | 250 |
251 | 252 | 253 | 254 | 255 | 256 | 257 | 258 | 259 | 260 | 261 | 262 | 263 | 264 | 265 | 266 | 267 | 268 | 269 | 270 | 271 | 272 | 273 | 274 | 275 |
276 | 277 | 278 | 279 | 280 | 281 | 282 | 283 | 284 | 285 | 286 | 287 | 288 | 289 | 290 | 291 | 292 | 293 | 294 | 295 | 296 | 297 | 298 | 299 | 300 |
301 | 302 | 303 | 304 | 305 | 306 | 307 | 308 | 309 | 310 | 311 | 312 | 313 | 314 | 315 | 316 | 317 | 318 | 319 | 320 | 321 | 322 | 323 | 324 | 325 |
326 | 327 | 328 | 329 | 330 | 331 | 332 | 333 | 334 | 335 | 336 | 337 | 338 | 339 | 340 | 341 | 342 | 343 | 344 | 345 | 346 | 347 | 348 | 349 | 350 |
351 | 352 | 353 | 354 | 355 | 356 | 357 | 358 | 359 | 360 | 361 | 362 | 363 | 364 | 365 | 366 | 367 | 368 | 369 | 370 | 371 | 372 | 373 | 374 | 375 |
376 | 377 | 378 | 379 | 380 | 381 | 382 | 383 | 384 | 385 | 386 | 387 | 388 | 389 | 390 | 391 | 392 | 393 | 394 | 395 | 396 | 397 | 398 | 399 | 400 |
401 | 402 | 403 | 404 | 405 | 406 | 407 | 408 | 409 | 410 | 411 | 412 | 413 | 414 | 415 | 416 | 417 | 418 | 419 | 420 | 421 | 422 | 423 | 424 | 425 |
426 | 427 | 428 | 429 | 430 | 431 | 432 | 433 | 434 | 435 | 436 | 437 | 438 | 439 | 440 | 441 | 442 | 443 | 444 | 445 | 446 | 447 | 448 | 449 | 450 |
451 | 452 | 453 | 454 | 455 | 456 | 457 | 458 | 459 | 460 | 461 | 462 | 463 | 464 | 465 | 466 | 467 | 468 | 469 | 470 | 471 | 472 | 473 | 474 | 475 |
476 | 477 | 478 | 479 | 480 | 481 | 482 | 483 | 484 | 485 | 486 | 487 | 488 | 489 | 490 | 491 | 492 | 493 | 494 | 495 | 496 | 497 | 498 | 499 | 500 |
501 | 502 | 503 | 504 | 505 | 506 | 507 | 508 | 509 | 510 | 511 | 512 | 513 | 514 | 515 | 516 | 517 | 518 | 519 | 520 | 521 | 522 | 523 | 524 | 525 |
526 | 527 | 528 | 529 | 530 | 531 | 532 | 533 | 534 | 535 | 536 | 537 | 538 | 539 | 540 | 541 | 542 | 543 | 544 | 545 | 546 | 547 | 548 | 549 | 550 |
551 | 552 | 553 | 554 | 555 | 556 | 557 | 558 | 559 | 560 | 561 | 562 | 563 | 564 | 565 | 566 | 567 | 568 | 569 | 570 | 571 | 572 | 573 | 574 | 575 |
576 | 577 | 578 | 579 | 580 | 581 | 582 | 583 | 584 | 585 | 586 | 587 | 588 | 589 | 590 | 591 | 592 | 593 | 594 | 595 | 596 | 597 | 598 | 599 | 600 |
601 | 602 | 603 | 604 | 605 | 606 | 607 | 608 | 609 | 610 | 611 | 612 | 613 | 614 | 615 | 616 |
I scored 13526 points (out of 15700 possible points, top rank was 17 out of ≈60000 in August 2017) at Hackerrank's Project Euler+.
My username at Project Euler is stephanbrumme while it's stbrumme at Hackerrank.
Look at my progress and performance pages to get more details.
Copyright
I hope you enjoy my code and learn something - or give me feedback how I can improve my solutions.
All of my solutions can be used for any purpose and I am in no way liable for any damages caused.
You can even remove my name and claim it's yours. But then you shall burn in hell.
The problems and most of the problems' images were created by Project Euler.
Thanks for all their endless effort !!!
<< problem 70 - Totient permutation | Counting fractions - problem 72 >> |