Topic: Delphi and Monte Carlo Tree Search
I write the chess program. I use a minimax and an alpha-betoj. There it is more and more less clear, it is necessary to cause recursively function of search a signed minus (Negamax). But that is not the question.
Esteemed about Monte Carlo Tree Search and it decided to try to be played with it. Monte Carlo Tree Search it is rough speaking, when we come into the most perspective nodes is more often and we prolong them. But (unlike a minimax) it is necessary to store in Monte-Carlo in storage all tree to do constant Update estimations of nodes and higher nodes.
From here a question:
How it is better to organize the data? Can eat implementations on Delphi the similar?
Thanks in advance.