def value_iteration_demo():
np.random.seed(0)
env = SnakeEnv(10, [3, 6])
agent = TableAgent(env)
pi_algo = ValueIteration()
with timer('Timer ValueIter'):
pi_algo.value_iteration(agent)
print('return_pi={}'.format(eval_game(env, agent)))
def generalized_iteration_demo():
np.random.seed(0)
env = SnakeEnv(10, [3, 6])
agent = TableAgent(env)
pi_algo = GeneralizedPolicyIteration()
with timer('Timer GeneralizedIter'):
pi_algo.generalized_policy_iteration(agent)
print('return_pi={}'.format(eval_game(env, agent)))
def policy_iteration_demo():
np.random.seed(0)
env = SnakeEnv(10, [3, 6])
agent = TableAgent(env)
pi_algo = PolicyIterationWithTimer()
with timer('Timer PolicyIter'):
pi_algo.policy_iteration(agent)
print 'return_pi={}'.format(eval_game(env, agent))