Author: Gian-Carlo Pascutto
Date: 15:21:32 01/07/01
Go up one level in this thread
On January 07, 2001 at 18:02:51, José Carlos wrote: >On January 07, 2001 at 16:38:50, martin fierz wrote: > >>hi, >> >>i just implemented an MTD(f) search for my checkers program instead of my usual >>PVS/NegaScout - it seems to be about same good. in my hashtable i only store >>the value and valuetype, and i noticed that in aske plaat's MTD(f) >>description he stores both upper and lower bounds in the hashtable. i >>couldn't figure out a reason to do this - can somebody enlighten me? > >I don't do MTD(f) in my program, but I can figure why it is useful to store >both bounds: in MTD(f), most of your searchs are null-window, so you are >failing high/low most of the time. The catch here is that it keeps failing in the same direction most of the time, so one bound will often be enough. You might want to experiment whether the memory saved by only storing one bound offsets the additional work if the wrong bound is stored. If you use convergence-accelerating heuristics, dual bounds will probably be better 'cos you will be failing in both directions then. -- GCP
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.