Combining Monte-Carlo Tree Search with Proof-Number Search

Elliot Doe; Mark H. M. Winands; Dennis J. N. J. Soemers; Cameron Browne

doi:10.1109/CoG51982.2022.9893635

Combining Monte-Carlo Tree Search with Proof-Number Search

Elliot Doe^*, Mark H. M. Winands, Dennis J. N. J. Soemers, Cameron Browne

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

30 Downloads (Pure)

Abstract

Proof-Number Search (PNS) and Monte-Carlo Tree Search (MCTS) have been successfully applied for decision making in a range of games. This paper proposes a new approach called PN-MCTS that combines these two tree-search methods by incorporating the concept of proof and disproof numbers into the UCT formula of MCTS. Experimental results demonstrate that PN-MCTS outperforms basic MCTS in several games including Lines of Action, MiniShogi, Knightthrough, and Awari, achieving win rates up to 94.0%.

Original language	English
Title of host publication	2022 IEEE Conference on Games (CoG)
Publisher	IEEE
Pages	206-212
Number of pages	7
ISBN (Print)	978-1-6654-5990-7
DOIs	https://doi.org/10.1109/CoG51982.2022.9893635
Publication status	Published - 24 Aug 2022
Event	2022 IEEE Conference on Games (CoG) - Beijing, China Duration: 21 Aug 2022 → 24 Aug 2022

Conference

Conference	2022 IEEE Conference on Games (CoG)
Period	21/08/22 → 24/08/22

Keywords

Monte Carlo methods
Decision making
Games

Access to Document

10.1109/CoG51982.2022.9893635

Full TextFinal published version, 520 KBLicence: Taverne

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9893635

Cite this

@inproceedings{94fe5082c7594fcb974ceb35e805438e,

title = "Combining Monte-Carlo Tree Search with Proof-Number Search",

abstract = "Proof-Number Search (PNS) and Monte-Carlo Tree Search (MCTS) have been successfully applied for decision making in a range of games. This paper proposes a new approach called PN-MCTS that combines these two tree-search methods by incorporating the concept of proof and disproof numbers into the UCT formula of MCTS. Experimental results demonstrate that PN-MCTS outperforms basic MCTS in several games including Lines of Action, MiniShogi, Knightthrough, and Awari, achieving win rates up to 94.0%.",

keywords = "Monte Carlo methods, Decision making, Games",

author = "Elliot Doe and Winands, {Mark H. M.} and Soemers, {Dennis J. N. J.} and Cameron Browne",

note = "Funding Information: ACKNOWLEDGMENTS This research is partially funded by the European Research Council as part of the Digital Ludeme Project (ERC Consolidator Grant #771292). Publisher Copyright: {\textcopyright} 2022 IEEE.; 2022 IEEE Conference on Games (CoG) ; Conference date: 21-08-2022 Through 24-08-2022",

year = "2022",

month = aug,

day = "24",

doi = "10.1109/CoG51982.2022.9893635",

language = "English",

isbn = "978-1-6654-5990-7",

pages = "206--212",

booktitle = "2022 IEEE Conference on Games (CoG)",

publisher = "IEEE",

address = "United States",

}

TY - GEN

T1 - Combining Monte-Carlo Tree Search with Proof-Number Search

AU - Doe, Elliot

AU - Winands, Mark H. M.

AU - Soemers, Dennis J. N. J.

AU - Browne, Cameron

N1 - Funding Information: ACKNOWLEDGMENTS This research is partially funded by the European Research Council as part of the Digital Ludeme Project (ERC Consolidator Grant #771292). Publisher Copyright: © 2022 IEEE.

PY - 2022/8/24

Y1 - 2022/8/24

N2 - Proof-Number Search (PNS) and Monte-Carlo Tree Search (MCTS) have been successfully applied for decision making in a range of games. This paper proposes a new approach called PN-MCTS that combines these two tree-search methods by incorporating the concept of proof and disproof numbers into the UCT formula of MCTS. Experimental results demonstrate that PN-MCTS outperforms basic MCTS in several games including Lines of Action, MiniShogi, Knightthrough, and Awari, achieving win rates up to 94.0%.

AB - Proof-Number Search (PNS) and Monte-Carlo Tree Search (MCTS) have been successfully applied for decision making in a range of games. This paper proposes a new approach called PN-MCTS that combines these two tree-search methods by incorporating the concept of proof and disproof numbers into the UCT formula of MCTS. Experimental results demonstrate that PN-MCTS outperforms basic MCTS in several games including Lines of Action, MiniShogi, Knightthrough, and Awari, achieving win rates up to 94.0%.

KW - Monte Carlo methods

KW - Decision making

KW - Games

UR - https://ieeexplore.ieee.org/document/9893635/

U2 - 10.1109/CoG51982.2022.9893635

DO - 10.1109/CoG51982.2022.9893635

M3 - Conference article in proceeding

SN - 978-1-6654-5990-7

SP - 206

EP - 212

BT - 2022 IEEE Conference on Games (CoG)

PB - IEEE

T2 - 2022 IEEE Conference on Games (CoG)

Y2 - 21 August 2022 through 24 August 2022

ER -

Combining Monte-Carlo Tree Search with Proof-Number Search

Abstract

Conference

Keywords

Access to Document

Other files and links

Cite this