View Single Post
Old 04-28-2007, 01:53 PM   #20 (permalink)
zakafellida
Upright
 
Im not sure to have understood everything written in this post and i might just say a solution that was posted before but ill still give it a try...

On a 6x6 board, a shape is average 3x3 and there's 15 shapes (supposition). That means there is 16 possibilities for each. Now if the solver can make 10k tries (with backtracking)/second, that would take 20828286 years.

Inspired by rsl12's post (checkerboard with black = x mod 2 = 0... ), i though that first, the program make 1 try, only one. Then, since every shape has 16 positions but they are grouped in 2 sections : those that activate X black checker and and Y white ones and those that activate Z black checker and and T white ones. Of course, it dosent work with shapes containing a pair number of squares.

But still, its not 15^16 anymore but 15^12, giving 411 years instead. Thats a 50677x improvement!

But then, if we consider a 4 color scheme, only 1/4 can be divided by 4 (thereforth unaffected) and the numbers of positions availible because divided by 4 instead 2.

So it becomes 15^6,75, giving (only...) 2h45! Yeah i know its still alot but thats still infinitly faster then backtracking only. Just for the sake of a number, with a 8 color scheme it would take 15^3,75, giving a mere...3 seconds.

My intuition tells me i made a mistake somewhere on how to calculate the amount of time taken, but if im not mistaken you puzzle there should take no more then 1/2 day. If that isn't fast enough just add more color schemes (a good number is Roundup((row+colum)/2*1,25). And if your computer can do more then 10000 tries/second...

However, im not a very experimented coder and visual basic isn't the fastest thing ever so i didn't manage to succeed doing this code...
zakafellida is offline  
 

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