fig = plt.figure()
ax = fig.gca(projection='3d')
ax.plot_surface(dealer, player, values_ua, cmap=cm.coolwarm,
linewidth=0)
plt.title("State value function (usable ace)")
plt.xlabel("Dealer showing")
plt.ylabel("Player sum")
#%%
if __name__ == "__main__":
num_episodes = 500000
gamma = 1
env = BlackJack()
states = env.state_space()
init_policy_player = policy(states)
policy_player = Monte_carlo_exploring_starts(env, states,
init_policy_player,
num_episodes, gamma)
#%%
VF = first_visit_MC_prediction(env, states, policy_player,
num_episodes, gamma)
#%%
plot_policy(policy_player)
plot_value_function(VF)
