Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel

Marc Lanctot; Viliam Lisý; Mark H. M. Winands

doi:10.1007/978-3-319-05428-5_3

Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel

Marc Lanctot, Viliam Lisý, Mark H. M. Winands

Networks and Strategic Optimization

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic

Abstract

Monte carlo tree search (mcts) has become a widely popular sampled-based search algorithm for two-player games with perfect information. When actions are chosen simultaneously, players may need to mix between their strategies. In this paper, we discuss the adaptation of mcts to simultaneous move games. We introduce a new algorithm, online outcome sampling (oos), that approaches a nash equilibrium strategy over time. We compare both head-to-head performance and exploitability of several mcts variants in goofspiel. We show that regret matching and oos perform best and that all variants produce less exploitable strategies than uct.

Original language	English
Title of host publication	Computer Games
Subtitle of host publication	Workshop on Computer Games, CGW 2013, Held in Conjunction with the 23rd International Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, August 3, 2013, Revised Selected Papers
Publisher	Springer
Pages	28-43
Number of pages	16
ISBN (Print)	978-3-319-05428-5
DOIs	https://doi.org/10.1007/978-3-319-05428-5_3
Publication status	Published - 2014

Publication series

Series	Communications in Computer and Information Science
Volume	408

Access to Document

10.1007/978-3-319-05428-5_3

http://dx.doi.org/10.1007/978-3-319-05428-5_3

Cite this

Lanctot, M., Lisý, V., & Winands, M. H. M. (2014). Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel. In Computer Games: Workshop on Computer Games, CGW 2013, Held in Conjunction with the 23rd International Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, August 3, 2013, Revised Selected Papers (pp. 28-43). Springer. Communications in Computer and Information Science Vol. 408 https://doi.org/10.1007/978-3-319-05428-5_3

Lanctot, Marc ; Lisý, Viliam ; Winands, Mark H. M. / Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel. Computer Games: Workshop on Computer Games, CGW 2013, Held in Conjunction with the 23rd International Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, August 3, 2013, Revised Selected Papers. Springer, 2014. pp. 28-43 (Communications in Computer and Information Science, Vol. 408).

@inbook{2f8e218b63aa49c5be1af056235fe00d,

title = "Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel",

abstract = "Monte carlo tree search (mcts) has become a widely popular sampled-based search algorithm for two-player games with perfect information. When actions are chosen simultaneously, players may need to mix between their strategies. In this paper, we discuss the adaptation of mcts to simultaneous move games. We introduce a new algorithm, online outcome sampling (oos), that approaches a nash equilibrium strategy over time. We compare both head-to-head performance and exploitability of several mcts variants in goofspiel. We show that regret matching and oos perform best and that all variants produce less exploitable strategies than uct.",

author = "Marc Lanctot and Viliam Lis{\'y} and Winands, {Mark H. M.}",

year = "2014",

doi = "10.1007/978-3-319-05428-5_3",

language = "English",

isbn = "978-3-319-05428-5",

series = "Communications in Computer and Information Science",

publisher = "Springer",

pages = "28--43",

booktitle = "Computer Games",

address = "United States",

}

Lanctot, M, Lisý, V & Winands, MHM 2014, Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel. in Computer Games: Workshop on Computer Games, CGW 2013, Held in Conjunction with the 23rd International Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, August 3, 2013, Revised Selected Papers. Springer, Communications in Computer and Information Science, vol. 408, pp. 28-43. https://doi.org/10.1007/978-3-319-05428-5_3

Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel. / Lanctot, Marc; Lisý, Viliam; Winands, Mark H. M.

Computer Games: Workshop on Computer Games, CGW 2013, Held in Conjunction with the 23rd International Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, August 3, 2013, Revised Selected Papers. Springer, 2014. p. 28-43 (Communications in Computer and Information Science, Vol. 408).

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic

TY - CHAP

T1 - Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel

AU - Lanctot, Marc

AU - Lisý, Viliam

AU - Winands, Mark H. M.

PY - 2014

Y1 - 2014

N2 - Monte carlo tree search (mcts) has become a widely popular sampled-based search algorithm for two-player games with perfect information. When actions are chosen simultaneously, players may need to mix between their strategies. In this paper, we discuss the adaptation of mcts to simultaneous move games. We introduce a new algorithm, online outcome sampling (oos), that approaches a nash equilibrium strategy over time. We compare both head-to-head performance and exploitability of several mcts variants in goofspiel. We show that regret matching and oos perform best and that all variants produce less exploitable strategies than uct.

AB - Monte carlo tree search (mcts) has become a widely popular sampled-based search algorithm for two-player games with perfect information. When actions are chosen simultaneously, players may need to mix between their strategies. In this paper, we discuss the adaptation of mcts to simultaneous move games. We introduce a new algorithm, online outcome sampling (oos), that approaches a nash equilibrium strategy over time. We compare both head-to-head performance and exploitability of several mcts variants in goofspiel. We show that regret matching and oos perform best and that all variants produce less exploitable strategies than uct.

U2 - 10.1007/978-3-319-05428-5_3

DO - 10.1007/978-3-319-05428-5_3

M3 - Chapter

SN - 978-3-319-05428-5

T3 - Communications in Computer and Information Science

SP - 28

EP - 43

BT - Computer Games

PB - Springer

ER -

Lanctot M, Lisý V, Winands MHM. Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel. In Computer Games: Workshop on Computer Games, CGW 2013, Held in Conjunction with the 23rd International Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, August 3, 2013, Revised Selected Papers. Springer. 2014. p. 28-43. (Communications in Computer and Information Science, Vol. 408). doi: 10.1007/978-3-319-05428-5_3