Open AccessA Parallel Monte-Carlo Tree Search Algorithm
Author(s) -
Tristan Cazenave,
Nicolas Jouandeau
Publication year - 2008
Publication title -
lecture notes in computer science
Language(s) - English
Resource type - Book series
SCImago Journal Rank - 0.249
H-Index - 400
eISSN - 1611-3349
pISSN - 0302-9743
DOI - 10.1007/978-3-540-87608-3_7
Subject(s) - monte carlo tree search , speedup , computer science , monte carlo method , tree (set theory) , algorithm , monte carlo algorithm , game tree , parallel computing , theoretical computer science , mathematics , game theory , statistics , combinatorics , sequential game , mathematical economics
Monte-Carlo Tree Search is a powerful paradigm for the game of Go. In this contribution we present a parallel Master-Slave algorithm for Monte-Carlo Tree Search and test it on a network of computers using various configurations: from 12,500 to 100,000 playouts, from 1 to 64 slaves, and from 1 to 16 computers. On our own architecture we obtain a speedup of 14 for 16 slaves. With a single slave and five seconds per move our algorithm scores 40.5% against GNU Go, with sixteen slaves and five seconds per move it scores 70.5%. At the end we give the potential speedups of our algorithm for various playout times.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom