Author: Anthony Cozzie
Date: 06:16:21 06/28/04
Go up one level in this thread
>The basic idea (which actually dates to my senior year in college, when I took >CAD tools, 18760) is to use simulated annealing. Simulated annealing makes >random changes and accepts them probablistically: > >(V1 > V0) : accept change >(V1 < V0) : accept if rand() > exp((V1 - V0)/T) > >Where T is the temperature, and it gradually gets smaller during the anneal. >This has the effect of ignoring small changes initially, and concentrating on >the big picture. This is the best link I can find in a few minutes of google >search; I will explain more if people are still unclear: > >http://members.aol.com/btluke/simann1.htm Amazingly I forgot to mention that it uses comparison training, that is making sure the evaluation of the position after the best move is higher than the lesser moves. anthony
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.