BCS Wrote: > Reply to Jason, > > > My hobby project is a multi-threaded game-playing AI. My current > > scheme uses a shared search tree using lockless updates with search > > results. > > > > As in threaded min-max? Have you got anything working? I known from > experience > that this ones a cast iron SOB. > > http://arrayboundserror.blogspot.com/search/label/min%20max >
No. Min-max is only good for theory. I'm also not doing alpha-beta which is successful in chess. I'm doing UCT and mostly aim to play the game of "go". UCT uses statistical bounds instead of hard heuristics. It also has less uniform tree exploration.
