[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: computer-go: a new fractal algorithm
Tapani Raiko wrote:
>
> > > the big problem seems to be how to reduce the board size without losing
> > > importnt information. If a reduced group becomes dead (because there isn't
> > > enough space to make it alive), then the reduction fails...
>
> This is a good point. The reduced shapes don't follow the same rules as
> mere stones. Perhaps the reduction might work somewhat, if there were also
> special stones to represent an eye and so on.
>
> By the way: try to imagine a board with 31 distinct living groups
> reduced to 5x5 board!!! What use is there from having solved 5x5 go?
>
> But of course the weirdest ideas are usually the best way to good ideas...
> :)
Of course the ultimate 5x5 board is useless for tactical needs. But for
strategic ones ?
The assumption I make in fractal go is that master 19x19 games, when "reduced",
even to as little as 5x5, still have some meaning : first occupy one corner,
then extend to a border, invade opponent groups, and so on...
Also don't forget that the fractal go algorithm is not a one-step reduction from
19x19 to 5x5 : there are several intermediate sizes, and at each size you
perform a "local "search" (with a small branching factor) to get the "best" move
for this size.
--
____________
/ Linux now !\ ("`-/")_.-'"``-._ _
,,, O \____________/ o @ @ `; -._ )-;-,_))
/'^'\ o° / catch it ?\O° (v_,)' _ )`-.\ `---'
( O O ) \___________/ __.- _..-_/ / ((.'
+---oOOo-(_)-oOOo-----------------((,.-'---((,/---------+
| Serge Boisse |
| SERVICE TECHNIQUE DE LA NAVIGATION AERIENNE (STNA) |
| ODS FRANCE project, http://www.stna.dgac.fr/phidias |
| tel: (0)562 14 5731 mailto:boisse@xxxxxxxxxxxxxxxxx |
| homepages: http://www.multimania.com/boisse |
| and: http://www.multimania.com/unitedplanet |
+-----------Oooo----------------------------------------+
oooO ( )
( ) ) /
\ ( (_/
\_)