Hi all,
after I had implemented the scheme I thought of in this post -
http://tech.groups.yahoo.com/group/fc-solve-discuss/message/1135 -
I found out that the new depth_dbm_fc_solver traversed more boards than
the normal dbm_fc_solver for one of the impossible deals. That indicated
that some deals were probably counted twice, so I decided to investigate.
What seems to have been the problem is the fact that some irreversible moves
actually consist of two irreversible moves. I.e: if we move the 2H from its original
location under a non-parent card to the foundations, then it would be equivalent to two
irreversible moves, because we may reach this in a different case by first moving it into
a freecell (one irreversible move) and then moving it to the foundation (another
irreversible move). So I need to account for those in both the calculation of the
rank of irreversibility of a single move, and the number of irreversible moves performed
by the Horne's Prune's process.
After I fixed this problem, the number of derived states and their contents was identical
between the depth_dbm_fc_solver and the dbm_fc_solver.
Regards,
Shlomi Fish
--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
Interview with Ben Collins-Sussman - http://shlom.in/sussman
If God exists and is the ego‐maniacal, sadistic and helpless creature that is
described in the Old Testament, then we’re in deep trouble.
Please reply to list if it's a mailing list post - http://shlom.in/reply .
Received on Tue Aug 14 2012 - 06:14:17 IDT