Forward Estimation for Game-Tree Search

No. 1: Agents, AI in Art and Entertainment, Knowledge Representation, and Learning

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 13

Game-Tree Search

It is known that bounds on the minimax values of nodes in a game tree can be used to reduce the computational complexity of minimax search for two-player games. We describe a very simple method to estimate bounds on the minimax values of interior nodes of a game tree, and use the bounds to improve minimax search. The new algorithm, called forward estimation, does not require additional domain knowledge other than a static node evaluation function, and has small constant overhead per node expansion. We also propose a variation of forward estimation, which provides a tradeoff between computational complexity and decision quality. Our experimental results show that forward estimation outperforms alpha-beta pruning on random game trees and the game of Othello.