Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Search algorithms in chess programs

Author: José Carlos

Date: 06:17:23 12/13/01

Go up one level in this thread


On December 13, 2001 at 03:47:39, Uri Blass wrote:

>On December 13, 2001 at 03:33:07, Wylie Garvin wrote:
>
>>On December 12, 2001 at 23:27:58, Russell Reagan wrote:
>>
>>>Depth-first search is usually what is used in chess programs (at least all that
>>>I have seen source for). What are the drawbacks of other search algorithms such
>>>as breadth-first and best-first that prevent them from being more widely used in
>>>chess programs?
>>
>>speed and space requirements.  Keep in mind that since nearly all chess programs
>>use iterative deepening on their depth first searches
>
>Nearly all?
>I doubt it.
>There are a lot of weak chess programs that do not use this technique.
>
>One of them (TSCP) even could win a tournament of weak programs that Dan Corbit
>made.

  TSCP _does_ iterative deepening. Iterative deepening means: first search with
depth = 1, then search with depth = 2, ... and so on.
  I don't know of any program that do fixed ply search with no iterative
deepeing. Sure there must be some, but not that I know of.

  José C.

>, and since chess has a
>>high branching factor, even with all the pruning techniques (well >= 2 anyway),
>>this means they are essentially getting a breadth-first search in depth-first
>>space, O(ply).  With selective search not all lines have the same "depth", but
>>the principle still applies.
>>
>>wylie
>
>What is the meaning of breadth first and best first search?
>
>Uri



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.