本文整理汇总了Python中gamestate.GameState.transition_function方法的典型用法代码示例。如果您正苦于以下问题：Python GameState.transition_function方法的具体用法？Python GameState.transition_function怎么用？Python GameState.transition_function使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类gamestate.GameState的用法示例。
在下文中一共展示了GameState.transition_function方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: __init__
# 需要导入模块: from gamestate import GameState [as 别名]
# 或者: from gamestate.GameState import transition_function [as 别名]
def __init__(self, player):
"""
Implementation of Monte Carlo Tree Search
Creates a root of an MCTS tree to keep track of the information
obtained throughout the course of the game in the form of a tree
of MCTS nodes
The data structure of a node consists of:
- the game state which it corresponds to
- w, the number of wins that have occurred at or below it in the tree
- n, the number of plays that have occurred at or below it in the tree
- expanded, whether all the children (legal moves) of the node have
been added to the tree
To access the node attributes, use the following format. For example,
to access the attribute 'n' of the root node:
policy = MCTSPolicy()
current_node = policy.root
policy.tree.node[current_node]['n']
"""
self.digraph = nx.DiGraph()
self.player = player
self.num_simulations = 0
# Constant parameter to weight exploration vs. exploitation for UCT
self.uct_c = np.sqrt(2)
self.node_counter = 0
empty_board = GameState()
self.digraph.add_node(self.node_counter, attr_dict={'w': 0,
'n': 0,
'uct': 0,
'expanded': False,
'state': empty_board})
empty_board_node_id = self.node_counter
self.node_counter += 1
self.last_move = None
if player is 'O':
for successor in [empty_board.transition_function(*move) for move in empty_board.legal_moves()]:
self.digraph.add_node(self.node_counter, attr_dict={'w': 0,
'n': 0,
'uct': 0,
'expanded': False,
'state': successor})
self.digraph.add_edge(empty_board_node_id, self.node_counter)
self.node_counter += 1