Python implementation of monte carlo tree search for 2 players zero-sum game
Project description
mctspy : python implementation of Monte Carlo Tree Search algorithm
Basic python implementation of Monte Carlo Tree Search (MCTS) intended to run on small game trees.
Installation
pip3 install mctspy
Running tic-tac-toe example
to run tic-tac-toe example:
import numpy as np
from mctspy.tree.nodes import TwoPlayersGameMonteCarloTreeSearchNode
from mctspy.tree.search import MonteCarloTreeSearch
from mctspy.games.examples.tictactoe import TicTacToeGameState
state = np.zeros((3,3))
initial_board_state = TicTacToeGameState(state = state, next_to_move=1)
root = TwoPlayersGameMonteCarloTreeSearchNode(state = initial_board_state)
mcts = MonteCarloTreeSearch(root)
best_node = mcts.best_action(10000)
Running MCTS for your own 2 players zero-sum game
If you want to apply MCTS for your own game, its state implementation should derive from
mmctspy.games.common.TwoPlayersGameState
(lookup mctspy.games.examples.tictactoe.TicTacToeGameState
for inspiration)
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
mctspy-0.1.1.tar.gz
(4.1 kB
view hashes)
Built Distribution
mctspy-0.1.1-py3-none-any.whl
(6.3 kB
view hashes)