Author: Robert Hyatt
Date: 19:33:04 11/06/03
Go up one level in this thread
On November 06, 2003 at 20:45:57, Dave Gomboc wrote: >On November 06, 2003 at 19:50:09, Robert Hyatt wrote: > >>On November 06, 2003 at 11:23:36, Dave Gomboc wrote: >> >>>On November 06, 2003 at 09:49:33, Robert Hyatt wrote: >>> >>>>On November 06, 2003 at 09:33:28, Renze Steenhuisen wrote: >>>> >>>>>On November 06, 2003 at 08:33:49, Gian-Carlo Pascutto wrote: >>>>> >>>>>>On November 06, 2003 at 05:45:53, Renze Steenhuisen wrote: >>>>>> >>>>>>>Depth-First Algorithms: >>>>>>> AlphaBeta (Fail-hard, Fail-Soft) >>>>>>> MTD(f) >>>>>>> >>>>>>>Best-First Algorithms: >>>>>>> SSS* >>>>>> >>>>>>The distinction between the three (and best-first and depth-first) >>>>>>is very hazy, read "Research re: search and research" by Aske Plaat. >>>>> >>>>>Done that already, but as Aske stated: they search the same nodes, but in a >>>>>different order. >>>>> >>>>>MTD(f) and the others are still DF algorithms, the second list works differently >>>>>(i.e., the order in which the nodes are expanded is different). >>>>> >>>>>Or am I talking rubish? >>>>> >>>>>Renze >>>>> >>>>>PS: Am I missing algorithms (either important or not)? >>>>>PS2: Are Scout and NegaScout equal? >>>> >>>> >>>>They are just variations on the same idea. All fall under the umbrella >>>>of alpha/beta depth-first search... (this is in response to your question >>>>PS2). >>>> >>>>depth-first and breadth-first (best-first is one example of the latter) >>>>are totally unrelated other than the fact they both search a tree. >>> >>>Well, no. Read Plaat's thesis. >>> >>>Dave >> >> >>I have read it. It does _not_ say the two are equivalent in any shape >>or form, except for the actual tree searched in certain circumstances. >>Depth-first and breadth-first are completely different approaches to >>growing a tree, even if on some occasions they grow the _same_ tree. > >In this particular case, the algorithms search the same tree. Therefore, I >think it's reasonable to claim they are they are equivalent in some shape or >form -- not in all shapes and all forms, but at list with respect to the nodes >searched and the order in which they are searched. :-) > >Dave I don't believe that last is correct. IE with respect to order. Particularly comparing members of the breadth-first family to the depth-first family and not just picking one specific algorithm from each.
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.