Semeais can, in principle, be solved by expert systems, except that
semeais with (basic) kos in general are a complex open research problem
(my current estimate for 2-groups semeais is that there must be hundreds
of different types!). Currently expert systems are a bit out of fashion
but at least you might use a minimal knowledge approach:
1) Classify functionally equivalent moves.
2) Represent each class as one move / by one representative move.
Examples:
(A)
# # # # # # # # # #
. . . . . . . . . .
O O O O O O O O O O
You do not need 10! sequences here but only 1.
(B)
. . . . . . . . . .
# # # # # # # # # #
. . . . . . . . . .
O O O O O O O O O O
. . . . . . . . . .
Classify this into three move classes: outside of Black, outside of
White, inside. Without expert knowledge, this is still too complex.
Maybe you are lucky and a classified local simulation finds the correct
answer. Expert knowledge surely does help a lot and reduces the
sequences from 30! via 3^10 to 1.
(C)
. . . . . . . . . .
. # . . . . . . # .
# # # # # # # # # #
. . . . . . . . . .
O O O O O O O O O O
. . O . O . O . . .
. . . . . . . . . .
Paired approach defects and eye-threatening potential benefit from more
expert knowledge, i.e., classify into more (sub)types.
Suggestion for MC:
1) Identify semeais by expert knowledge.
2) Divide the board into each semeai / non-semeai region.
3) Solve / represent each semeai by expert knowledge.
4) Do global MC with the change of special treatment for each semeai
region. E.g., know the best black / white move in each semeai region and
treat that region like just one intersection for the respective player
to move (alternatively: as two intersections for either player).
For other problems like life and death, endgame or game's first move, a
similar Local Expert Knowledge approach could be used.
There are complications of course:
- dynamic region update
- multiple threats on several regions
- ko
--
robert jasiek
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
