01 May 2015

Flood It: Next Results

I wrote up the more general algorithm I mentioned before:  Find a lower bound to remove each color and then compute an overall lower bound.  I found a bug in the algorithm from the previous post, which pretty much invalidates those results.  For the more general algorithm we get the following (at least until we find the next bug...).

14x14 board: 5,776 nodes evaluated,1,795 expanded.  In 44 seconds.
21x12 board: 109,246 nodes evaluated, 36,579 expanded.  In 26 minutes.

Looks like the next step might be to optimize the performance of the new algorithm a bit.


Post a Comment

<< Home