Блог пользователя Swistakk

Автор Swistakk, история, 7 лет назад, По-английски

Hey, I recently managed to play with some small magnet balls and created a simple riddle:

https://www.dropbox.com/s/raleq3lsuhjr4pz/Riddle.mp4?dl=0

Question is: how many balls are there?

I didn't do anything sly, interior is filled as tight as it can be. Of course it is rather simple, but I find it entertaining. Please use 'hide' tags for answers.

UPD: Congratulations to lmn0x4F on being first one to give correct answer :)! If you want to read my explanation and additional comment then look at my answer to rng_58's comment.

  • Проголосовать: нравится
  • +85
  • Проголосовать: не нравится

»
7 лет назад, # |
  Проголосовать: нравится +28 Проголосовать: не нравится

I guess i'll try my luck, it's probably incorrect though XD

Spoiler
»
7 лет назад, # |
  Проголосовать: нравится +56 Проголосовать: не нравится
Spoiler

:p

»
7 лет назад, # |
Rev. 2   Проголосовать: нравится 0 Проголосовать: не нравится

My very simple guess:

Спойлер
»
7 лет назад, # |
Rev. 2   Проголосовать: нравится +22 Проголосовать: не нравится
Spoiler
  • »
    »
    7 лет назад, # ^ |
      Проголосовать: нравится +47 Проголосовать: не нравится

    Task: How many balls in Michelangelo's Swistakk's sculpture?

    Solution
  • »
    »
    7 лет назад, # ^ |
      Проголосовать: нравится 0 Проголосовать: не нравится
    My explanation and additional comment
»
7 лет назад, # |
Rev. 13   Проголосовать: нравится 0 Проголосовать: не нравится

May be

Answer
Reason
»
7 лет назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

I have a range

Spoiler
»
7 лет назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

I think

Spoiler
»
7 лет назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

Auto comment: topic has been updated by Swistakk (previous revision, new revision, compare).

»
7 лет назад, # |
Rev. 2   Проголосовать: нравится +56 Проголосовать: не нравится

By the way, it reminds me one of my favorite games in my babyhood.

I had lots of silver balls:

I also had Go's stones. I didn't use the stones in this game, I only used the lids:

Put n silver balls into the lid, stir them, and wait until they stop. Observe the pattern formed by the balls.

Formally, the surface of the lid can be approximated as z = k(x2 + y2), so we want to find (x1, y1), ..., (xn, yn) that minimizes the sum of xi2 + yi2 under the constraint that (xi - xj)2 + (yi - yj)2 ≥ 1. Guess what happens when n = 7. Can you guess the patterns for bigger n?

  • »
    »
    7 лет назад, # ^ |
    Rev. 3   Проголосовать: нравится 0 Проголосовать: не нравится

    Actually, with surface equation as it is right now, minimizing potential energy is equivalent to minizing sum of not xi2 + yi2, but I guess that answers don't change. Does the surface have shape of cone (current equation describes it) or paraboloid (what would correspond to minimizing sum of xi2 + yi2)? Moreover actual constraint is (xi - xj)2 + (yi - yj)2 + (zi - zj)2 ≥ 1, you forgot about z. Actually for n = 7 I have 3 possible candidates for answer, but it is hard to check which one is best, so I will just say that I bet on

    Spoiler
    • »
      »
      »
      7 лет назад, # ^ |
        Проголосовать: нравится 0 Проголосовать: не нравится

      Sorry it was like a very flat paraboloid. I guess this game is very boring for cones. k is very small and z term can be almost ignored (otherwise the pattern very heavily depends on the value of k).

      Yes, for n = 7 you are correct, though I used to play it like 15-20 years ago and I can't confirm the answer for other n now.

      • »
        »
        »
        »
        7 лет назад, # ^ |
          Проголосовать: нравится 0 Проголосовать: не нравится

        If we ignore z term then this becomes very intuitive (however I am not sure how to prove this in two lines) and in fact all of my 3 candidates for answer collapse to the same one ;p.

  • »
    »
    7 лет назад, # ^ |
      Проголосовать: нравится 0 Проголосовать: не нравится

    I guess that for n ≥ 7 we can get a lot of different patterns as answers for a fixed n. Physics only guarantees to find a local minimum instead of global minimum and I think many "downward closed" patterns on triangular lattice can be local optimums.

    Another example of geometric optimization problem where physics can help us that I saw was Steiner tree on a plane. Guy that was showing it had two glass tiles placed close and parallel to each other connected by short sticks in few places. He then sinked that in some soapish liquid and after taking it out that liquid created network of walls between these sticks that was sometimes a Steiner tree (and sometimes some other local minimum :P)

    Would be funny to use physics to (NP-?)hard problems, however it seems that physics can only find local minimums and not global ones. And if there is only one local minimum then probably it can be found by local search/gradient descent or similar techniques?

    • »
      »
      »
      7 лет назад, # ^ |
        Проголосовать: нравится +9 Проголосовать: не нравится

      Actually, physics can even solve problems harder than NP, so even if P=NP the quantum computation will be more effective than classical in some cases. An example that is already achieved by humans is boson sampling, that was done using integrated photonics. I know that some scientific teams try to use the same method for problems like finding exit from a maze.

      I have also heard of a quantum computing circuit that can effectively simulate the dynamics of a particle, which is a computationally-heavy problem otherwise.

  • »
    »
    7 лет назад, # ^ |
      Проголосовать: нравится 0 Проголосовать: не нравится

    In physics, you usually use Monte Carlo method (a Marathon Match veteran would probably upgrade this to simulated annealing) to simulate the process of reaching equilibrium. If you take a bigger space (or even infinite and then talk about density of balls instead of their count) you may talk about phase transition. For counts/densities below threshold you will have mor or less chaos (i.e. liquid) while for higher densities you end up with order (i.e. solid).

    The order may take various forms, e.g. triangle or square lattice. In finite case, there will probably be clusters of densely packed balls. In infinite case there should appear translational symmetry.

    If the z-coordinate differences are small then it may be neglected in the radius condition and only taken into account as some potential energy that drives the balls off the edge.