Author: Heiner Marxen
Date: 12:46:48 06/01/02
Go up one level in this thread
On June 01, 2002 at 15:00:48, Heiner Marxen wrote: >On June 01, 2002 at 13:07:36, Dan Andersson wrote: > >>>Why do you think that draws (even with large/long cycles) are recognized >>>easily with PN^2 ? Do you have experience? Data to share? Theoretic >>>insight? >>As the problems with detecting long draws are depth based. A best first search >>seems appropriate. But to avoid duplicate work, a cache of already evaluated >>nodes need to be probed also. A modified PN search called Proof-Set Search by >>Martin Müller might be applicable. > >Thanks! Can be downloaded (ps orf pdf) via > http://web.cs.ualberta.ca/~mmueller/publications.html > >I havn't yet studied it thoroughly, but it sounds quite interesting >after a short look into it. Also, Martin Müller seems to have more >to say about advanced search methods, maybe I'll read his other papers also. After reading more of it I have to add that unfortunately PSS (Proof-Set Search) does not (yet) work for cyclic graphs, which is exactly the case I try to tackle. Still, it can be an improvement over standard PN-search, so it may be interesting for those who already do PN-search. >Many thanks for the pointer! That is still valid! :-) Cheers, Heiner
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.