 Show Posts
|
|
Pages: [1] 2 3 ... 8
|
|
1
|
Playing the Game / Custom Puzzle Levels / Re: Curly's Puzzles
|
on: April 14, 2013, 08:27:51 PM
|
For Signal to Noise Ratio ( pivotis): hogirorThat was tough, very tough. I started out with one approach, but could not get it to handle all special cases (quite a view for this level). I tried another approach, but did not go anywhere with that either. Then I reverted back to the first approach again, and painstakingly got everything to work. I have no room to spare.  The solution I settled on is a parallel solution: all cycles are detected in parallel, and the smallest cycle that has not failed yet is continuously selected as solution. The hardest trick was to find out which one is the smallest: the valve head of the 15 cylinder engine that forms the solution. The alternative I tried was a sequential approach: try the 1 cycle first, if that fails, try the 2 cycle, etc, with both 0 and F as special cases.
I didn't expect it to be that tough when I created it but I spent almost a week trying to solve it without success before posting it. I'm glad you didn't find it easy.
My attempts at a sequential solution always ran out of space. I knew there would be a parallel solution but I didn't know how to build that. I had to study yours for ages to work out how it works. Very cool. I even like the little bit that multiplexes the input and output streams to save some extra space. |
|
|
|
|
7
|
Playing the Game / Custom Puzzle Levels / Re: Curly's Puzzles
|
on: February 07, 2013, 05:04:47 AM
|
For Logic Problem ( gorecop): jupafotEdit: In this cleaned up version I accidently used door keys from the sandbox: cokudad. I'll remove the door keys one day.
Actually both of your versions use door keys at the bottom! But I guess I can let you away with that.
Here's my permutation debedid which shares 7 of the 16 inputs with yours. |
|
|
|
|
8
|
Playing the Game / Custom Puzzle Levels / Re: Curly's Puzzles
|
on: February 06, 2013, 07:09:37 PM
|
Your progress encouraged me to finish writing the program I'd started but I found 5178 (instead of 1578) combinations so perhaps you made a typo before. Mine runs in 0.6 seconds  I find all the valid inputs for each of the 8 machines separately first, and then work through the full permutations pruning anything that would give duplicate inputs or outputs.
It'll be interesting to see whether you pick a combination that helps on the input side or the output side. Perhaps there is an ideal one that makes both easy. My solution wasn't built this way and so isn't pretty but perhaps tonight I'll find a nicer one. |
|
|
|
|
9
|
Playing the Game / Custom Puzzle Levels / Re: Curly's Puzzles
|
on: February 06, 2013, 06:19:52 PM
|
Then the chances are good that each solution will be quite different 
My original plan was for a puzzle which had just a single solution but I guess 1578 out of 16! is hard enough to ensure nobody will find a solution without killing some brain cells.
|
|
|
|
|
10
|
Playing the Game / Custom Puzzle Levels / Re: Curly's Puzzles
|
on: February 06, 2013, 09:09:10 AM
|
Have you tested a solution that works? I've programmed a simulation of all possible input permutations and cannot reproduce the set of outputs, even accounting for timing differences in the timing-sensitive inputs.
Yes, I have a working solution. Does your program simulate Rubicon physics, or does it just do each of the 8 calculations? If the latter then I think I know what you have missed. It's not meant to be easy  |
|
|
|
|
15
|
General / Support / Re: Bug
|
on: January 11, 2013, 10:36:03 PM
|
But having saved it it now behaves as expected.
The one I still have open in my editor is broken. |
|
|
|
|
|
Rubicon Forum
