Hi all,
I would like to consult with you about pruning deep positions in Freecell games
(especially those with only 2 available freecells). As the game progresses
more cards are placed on sequences of alternating colours and descending ranks,
and as a result many empty (or potentially empty) columns are available.
Then cards can be moved from columns into sequences into these free columns
using atomic reversible moves (see
http://tech.groups.yahoo.com/group/fc-solve-discuss/message/1093 )
yielding an exponential (or close to it) number of derived states in the same
fully-connected-component, only to see if we can perform some irreversible
moves eventually. As a result, the memory and the time consumption required to
traverse all these states increases considerably.
My question is can anyone think of any way to prune unnecessary states which
yield no gain and only yield an enumeration of the exit-points of the
irreversible moves which can actually bring progress? (Instead of having the
solver beat around the bush.)
I have some initial ideas and leads for that, like finding which possible
irreversible moves can be done and trying to form a game layout that will make
them happen, but I would appreciate some other insights.
Regards,
Shlomi Fish
--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
Parody of "The Fountainhead" - http://shlom.in/towtf
I hope that if it had not been clear before, it isn’t less clear now.
— One of Shlomi Fish’s Technion Lecturers
Please reply to list if it's a mailing list post - http://shlom.in/reply .
Received on Tue Mar 06 2012 - 12:21:22 IST