Monte-Carlo proof-number search for computer Go

J.T. Saito; G.M.J.B. Chaslot; J.W.H.M. Uiterwijk; H.J. van den Herik

doi:10.1007/978-3-540-75538-8_5

Monte-Carlo proof-number search for computer Go

J.T. Saito^*, G.M.J.B. Chaslot, J.W.H.M. Uiterwijk, H.J. van den Herik

^*Corresponding author for this work

Department of Advanced Computing Sciences

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

Abstract

In the last decade, proof-number search and Monte-Carlo methods have successfully been applied to the combinatorial-games domain. Proof-number search is a reliable algorithm. It requires a well defined goal to prove. This can be seen as a disadvantage. In contrast to proof-number search, Monte-Carlo evaluation is a flexible stochastic evaluation for game-tree search. In order to improve the efficiency of proof-number search, we introduce a new algorithm, Monte-Carlo Proof-Number search. It enhances proof-number search by adding the flexible Monte-Carlo evaluation. We present the new algorithm and evaluate it on a sub-problem of Go, the Life-and-Death problem. The results show a clear improvement in time efficiency and memory usage: the test problems are solved two times faster and four times less nodes are expanded on average. Future work will assess possibilities to extend this method to other enhanced Proof-Number techniques.

Original language	English
Title of host publication	Lecture Notes in Computer Science - Computers and Games
Editors	H. Matsubara, H.J. van den Herik, P. Ciancarini, H.H.L. Dronkers
Pages	50-61
DOIs	https://doi.org/10.1007/978-3-540-75538-8_5
Publication status	Published - 1 Jan 2006
Event	5th International Conference CG, Turin, Italy - Duration: 29 May 2006 → 31 May 2006

Conference

Conference	5th International Conference CG, Turin, Italy
Period	29/05/06 → 31/05/06

Access to Document

10.1007/978-3-540-75538-8_5

Cite this

@inproceedings{5453908a675845bd804871ac51df51a3,

title = "Monte-Carlo proof-number search for computer Go",

abstract = "In the last decade, proof-number search and Monte-Carlo methods have successfully been applied to the combinatorial-games domain. Proof-number search is a reliable algorithm. It requires a well defined goal to prove. This can be seen as a disadvantage. In contrast to proof-number search, Monte-Carlo evaluation is a flexible stochastic evaluation for game-tree search. In order to improve the efficiency of proof-number search, we introduce a new algorithm, Monte-Carlo Proof-Number search. It enhances proof-number search by adding the flexible Monte-Carlo evaluation. We present the new algorithm and evaluate it on a sub-problem of Go, the Life-and-Death problem. The results show a clear improvement in time efficiency and memory usage: the test problems are solved two times faster and four times less nodes are expanded on average. Future work will assess possibilities to extend this method to other enhanced Proof-Number techniques.",

author = "J.T. Saito and G.M.J.B. Chaslot and J.W.H.M. Uiterwijk and {van den Herik}, H.J.",

note = "Lecture Notes in Computer Science - Computers and Games Vol.:4630 ; 5th International Conference CG, Turin, Italy ; Conference date: 29-05-2006 Through 31-05-2006",

year = "2006",

month = jan,

day = "1",

doi = "10.1007/978-3-540-75538-8_5",

language = "English",

pages = "50--61",

editor = "H. Matsubara and Herik, {H.J. van den} and P. Ciancarini and H.H.L. Dronkers",

booktitle = "Lecture Notes in Computer Science - Computers and Games",

}

TY - GEN

T1 - Monte-Carlo proof-number search for computer Go

AU - Saito, J.T.

AU - Chaslot, G.M.J.B.

AU - Uiterwijk, J.W.H.M.

AU - van den Herik, H.J.

N1 - Lecture Notes in Computer Science - Computers and Games Vol.:4630

PY - 2006/1/1

Y1 - 2006/1/1

N2 - In the last decade, proof-number search and Monte-Carlo methods have successfully been applied to the combinatorial-games domain. Proof-number search is a reliable algorithm. It requires a well defined goal to prove. This can be seen as a disadvantage. In contrast to proof-number search, Monte-Carlo evaluation is a flexible stochastic evaluation for game-tree search. In order to improve the efficiency of proof-number search, we introduce a new algorithm, Monte-Carlo Proof-Number search. It enhances proof-number search by adding the flexible Monte-Carlo evaluation. We present the new algorithm and evaluate it on a sub-problem of Go, the Life-and-Death problem. The results show a clear improvement in time efficiency and memory usage: the test problems are solved two times faster and four times less nodes are expanded on average. Future work will assess possibilities to extend this method to other enhanced Proof-Number techniques.

AB - In the last decade, proof-number search and Monte-Carlo methods have successfully been applied to the combinatorial-games domain. Proof-number search is a reliable algorithm. It requires a well defined goal to prove. This can be seen as a disadvantage. In contrast to proof-number search, Monte-Carlo evaluation is a flexible stochastic evaluation for game-tree search. In order to improve the efficiency of proof-number search, we introduce a new algorithm, Monte-Carlo Proof-Number search. It enhances proof-number search by adding the flexible Monte-Carlo evaluation. We present the new algorithm and evaluate it on a sub-problem of Go, the Life-and-Death problem. The results show a clear improvement in time efficiency and memory usage: the test problems are solved two times faster and four times less nodes are expanded on average. Future work will assess possibilities to extend this method to other enhanced Proof-Number techniques.

U2 - 10.1007/978-3-540-75538-8_5

DO - 10.1007/978-3-540-75538-8_5

M3 - Conference article in proceeding

SP - 50

EP - 61

BT - Lecture Notes in Computer Science - Computers and Games

A2 - Matsubara, H.

A2 - Herik, H.J. van den

A2 - Ciancarini, P.

A2 - Dronkers, H.H.L.

T2 - 5th International Conference CG, Turin, Italy

Y2 - 29 May 2006 through 31 May 2006

ER -