def run_experiment(player_policies, experiment_name, n):
games = []
winners = []
X_wins = np.zeros(n)
O_wins = np.zeros(n)
X_win_rate = np.zeros(n)
O_win_rate = np.zeros(n)
for i in range(n):
G, winner = play_game(player_policies)
games.append(G)
winners.append(winner)
# Copy cumulative totals from previous timestep
if i > 0:
X_wins[i] = X_wins[i - 1]
O_wins[i] = O_wins[i - 1]
if winner == 'X':
X_wins[i] += 1
elif winner == 'O':
O_wins[i] += 1
X_win_rate[i] = X_wins[i] / (i + 1)
O_win_rate[i] = O_wins[i] / (i + 1)
from collections import Counter
c = Counter(winners)
print(c.items())
print(X_win_rate)
print(X_wins)
from matplotlib import pyplot as plt
# Plot cumulative win rate over time
plt.plot(X_win_rate)
plt.plot(O_win_rate)
plt.legend(['X', 'O'])
plt.xlabel('simulation')
plt.ylabel('cumulative win rate')
plt.ylim([-0.1, 1.1])
plt.title('Experiment ID: {}'.format(name.replace('_', ' ').upper()))
# Save figure to disk with a unique identifier
plt.savefig('simulation_time_' + str(simulation_time) + '_' +
'{}.png'.format(experiment_name))
plt.cla()
# Save a visualization of the first n levels of
# the MCTS trees used by each player that used MCTS
for policy in player_policies:
if type(policy) is MCTSPolicy:
visualize_mcts_tree(mcts=policy,
depth=0,
filename='{}_{}_{}'.format(
name, policy.player, n))
from gameplay import play_game
from policies import RandomPolicy, MCTSPolicy
import numpy as np
import networkx as nx
player_policies = [MCTSPolicy(), RandomPolicy()]
# For reproducibility
np.random.seed(0)
games = []
for i in range(100):
games.append(play_game(player_policies))
graphs = [game[0] for game in games]
dot_graph_combined = nx.compose_all(graphs)
dot_graph = nx.to_pydot(dot_graph_combined)
dot_graph.set_graph_defaults(fontname='Courier')
dot_graph.write_png('multiple_game_graph.png')
def run_experiment(player_policies, experiment_name):
# For reproducibility
np.random.seed(0)
# Number of games to play per experiment
n = 100
games = []
winners = []
X_wins = np.zeros(n)
O_wins = np.zeros(n)
X_win_rate = np.zeros(n)
O_win_rate = np.zeros(n)
for i in range(n):
G, winner = play_game(player_policies)
games.append(G)
winners.append(winner)
# Copy cumulative totals from previous timestep
if i > 0:
X_wins[i] = X_wins[i - 1]
O_wins[i] = O_wins[i - 1]
if winner == "X":
X_wins[i] += 1
elif winner == "O":
O_wins[i] += 1
X_win_rate[i] = X_wins[i] / (i + 1)
O_win_rate[i] = O_wins[i] / (i + 1)
from collections import Counter
c = Counter(winners)
print(c.items())
print(X_win_rate)
print(X_wins)
from matplotlib import pyplot as plt
# Plot cumulative win count over time
# plt.plot(X_wins)
# plt.plot(O_wins)
# plt.legend(['X wins', 'O wins'])
# Plot cumulative win rate over time
plt.plot(X_win_rate)
plt.plot(O_win_rate)
plt.legend(["X", "O"])
plt.xlabel("simulation")
plt.ylabel("cumulative win rate")
plt.ylim([-0.1, 1.1])
plt.title("Experiment ID: {}".format(name.replace("_", " ").upper()))
# Save figure to disk with a unique identifier
import uuid
plt.savefig("{}.png".format(experiment_name))
# plt.show()
plt.cla()