Hi all,
On Sun, 17 Jun 2012 08:30:53 +0300
Shlomi Fish <shlomif_at_shlomifish.org> wrote:
> Hi all,
>
> as I’ve been playing some Baker’s Dozen (see
> http://en.wikipedia.org/wiki/Baker%27s_Dozen_%28solitaire%29 ) lately,
> I came to a realisation of one way we could prune the moves there, and in
> other variants where empty columns cannot be filled by any card, and which I
> believe would still guarantee a correct verdict.
>
> What I thought of is that there is no point in moving the last card in a column
> to a parent on a different column, because then the column won't be able to be filled,
> and will be left to disuse. Furthermore, after moved to a parent, the card might block
> other cards that can be placed on the parent.
>
OK, so I implemented this prune in commit
7cd599215f83a82241bb10abd97434903e8bf88b in the repository
(titled "Prune for tests__is_filled_by_none().") and bechmarked the first
two-hundred (200) PySolFC Baker’s Dozen deals. The results are that:
1. With the prune it now runs in 742 wallclock seconds instead of 938s (a 26% improvement).
2. It now solves 127 deals instead of 89.
3. It's not all roses, because some of the previously solved deals are now intractable.
4. Only one deal was found to be unsolvable in both cases, and it's the same deal.
---------------
Note that this is with the default Freecell Solver configuration.
So overall I am happy - it did not require a lot of coding work, and it yielded good results.
Regards,
Shlomi Fish
> A variation of this prune for variants where only kings can fill empty columns is that if
> there are enough columns to provide hosting for all the kings that are active, then there
> is no point in clearing a column of cards.
>
> Does this sound reasonable to you?
>
> Regards,
>
> Shlomi Fish
>
>
--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
Escape from GNU Autohell - http://www.shlomifish.org/open-source/anti/autohell/
He has a high degree of idealism, a high degree of stubbornness, and an even
higher degree of inability to distiniguish between the two.
Please reply to list if it's a mailing list post - http://shlom.in/reply .
Received on Tue Jun 19 2012 - 05:25:03 IDT