Accelerated UCT and Its Application to Two-Player Games
نویسندگان
چکیده
Monte-Carlo Tree Search (MCTS) is a very successful approach for improving the performance of game-playing programs. This paper presents the Accelerated UCT algorithm, which overcomes a weakness of MCTS caused by deceptive structures which often appear in game tree search. It consists in using a new backup operator that assigns higher weights to recently visited actions, and lower weights to actions that have not been visited for a long time. Results in Othello, Havannah and Go show that Accelerated UCT is not only more effective than previous approaches but also improves the strength of Fuego, which is one of the best computer Go programs.
منابع مشابه
An Analysis of UCT in Multi-player Games
The UCT algorithm has been exceedingly popular for Go, a two-player game, significantly increasing the playing strength of Go programs in a very short time. This paper provides an analysis of the UCT algorithm in multi-player games, showing that UCT, when run in a multi-player game, is computing a mixed-strategy equilibrium, as opposed to max, which computes a pure-strategy equilibrium. We anal...
متن کاملDelete Relaxation and Traps in General Two-Player Zero-Sum Games
General game playing (GGP) is concerned with constructing players that can handle any game describable in a pre-defined language reasonably well. Nowadays, the most common approach is to make use of simulation based players using UCT. In this paper we consider the alternative, i.e., an Alpha-Beta based player. In planning, delete relaxation heuristics have been very successful for guiding the s...
متن کاملA Doppelkopf Player Based on UCT
We propose doppelkopf, a trick-taking card game with similarities to skat, as a benchmark problem for AI research. While skat has been extensively studied by the AI community in recent years, this is not true for doppelkopf. However, it has a substantially larger state space than skat and a unique key feature which distinguishes it from skat and other card games: players usually do not know wit...
متن کاملUsing Modified UCT Algorithm Basing on Risk Estimation Methods in Imperfect Information Games
Risk dominance and payoff dominance strategy are two complementary parts of the game theory decision strategy. While payoff dominance is still the basic principle in perfect information, two player games, risk dominance has shown its advantages in imperfect information conditions. In this paper, we first review the related work in the area of estimation methods and the influence of risk factors...
متن کاملComparing UCT versus CFR in Simultaneous Games
Simultaneous move games where all the player have to take their actions simultaneously are a class of games in general game playing. In this paper we analyze how UCT performs in this class of games. We argue that UCT does not converge to a Nash equilibrium in general and the situation that it converges to can be exploited. We also analyze CFR (CounterFactual Regret) and show how it can be used ...
متن کامل