Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Search algorithms

Author: Robert Hyatt

Date: 17:38:25 11/09/03

Go up one level in this thread


On November 09, 2003 at 12:01:59, Dave Gomboc wrote:

>On November 08, 2003 at 12:43:29, Robert Hyatt wrote:
>
>>On November 07, 2003 at 14:16:16, Dave Gomboc wrote:
>>
>>>On November 06, 2003 at 22:42:14, Robert Hyatt wrote:
>>>
>>>>On November 06, 2003 at 22:33:04, Robert Hyatt wrote:
>>>>
>>>>>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.
>>>>
>>>>
>>>>BTW,  I hope you don't try to convince me all sort algorithms are
>>>>equivalent, just because they take the same list and produce the
>>>>same final result.  :)
>>>
>>>Well, what is correct is that the node expansions are done in the same order.
>>>
>>>Dave
>>
>>
>>Again, what two algorithms specifically are you comparing?  Best first
>>and depth-first _never_ expand the nodes in the same order, except for
>>trees where each side has only one legal move at every position...
>
>Replied in other thread.
>
>Dave


I'll go back and re-read this stuff.  One thing is crystal clear.  Best
first is a breadth-first algorithm.  mtd(f) is clearly a depth-first
algorithm based on alpha/beta.  I don't see how the two can ever be
comparable in the tree size searched, nor in the order the nodes are
expanded.  The two search strategies (depth-first vs breadth-first) are
simply too different...

At least in any book I own.



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.