Monte Carlo Tree Search

MCTS is a heuristic search algorithm for some types of decision-making processes, particularly those used in game theory. Traditional artificial-intelligence algorithms are very powerful, yet they require a lot of memory in their execution. In 2006, Kocsis and Szepesvari, with the aim of improving the classic algorithms, introduced the MCTS. Compared to the Minimax, the main difference of the MCTS method is that the expansion of the state tree does not take place completely.

The minimax, in decision theory, is a method to minimize the maximum (minimax) possible loss; alternatively, to maximize the minimum gain (maximin). It was discovered in the theory of games in the case of a zero-sum game with two players, both ...

Get Keras Reinforcement Learning Projects now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.