def calculate_values(grid):
# initialize V(s)
V = {}
states = grid.all_states()
for s in states:
V[s] = 0
# repeat until convergence
# V[s] = max[a]{ sum[s',r] { p(s',r|s,a)[r + gamma*V[s']] } }
i = 0
while i < 4:
# biggest_change is referred to by the mathematical symbol delta in equations
biggest_change = 0
for s in grid.non_terminal_states():
old_v = V[s]
_, new_v = best_action_value(grid, V, s)
V[s] = new_v
biggest_change = max(biggest_change, np.abs(old_v - new_v))
if biggest_change < SMALL_ENOUGH:
break
print("values:")
print_values(V, grid)
i += 1
return V
def main():
grid = standard_grid(obey_prob=1.0, step_cost=None)
print_values(grid.rewards, grid)
V, Policy, Deltas = monte_carlo(grid)
print_values(V, grid)
print_policy(Policy, grid)
plt.plot(Deltas)
plt.show()
def visit(inst, s, solved, values):
# TODO: add your code here.
# Make use of compute_greedy_action_and_value, sample_successor, and
# check_solved.
# Return updated labeling solved and updated value function values.
"""
Run the LRTDP algorithm until it converges.
"""
def lrtdp(inst, values):
solved = { s: False for s in inst.states }
iteration = 1
while not solved[inst.init]:
wait_for_input("Press enter for another iteration of LRTDP...".format(iteration))
solved, values = visit(inst, inst.init, solved, values)
print("Values after iteration {}: ".format(iteration))
print_values(inst, values)
print("Solved after iteration {}: ".format(iteration))
print_solved(inst, solved)
iteration += 1
return values
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument(
'algorithm', choices=['rtdp', 'lrtdp'],
help="Choose the algorithm."
)
args = parser.parse_args()
inst = instance.get_example_instance()
print(inst)
values = { s : heuristic(inst, s) for s in inst.states }
print("")
print("Initial state-values:")
print_values(inst, values)
if args.algorithm == 'rtdp':
values = rtdp(inst, values)
elif args.algorithm == 'lrtdp':
values = lrtdp(inst, values)
else:
sys.exit("Unknown algorithm")
print("")
print("Final values:")
print_values(inst, values)
policy = get_greedy_policy(inst, values)
print("Corresponding final policy:")
print_policy(inst, policy)
def lrtdp(inst, values):
solved = { s: False for s in inst.states }
iteration = 1
while not solved[inst.init]:
wait_for_input("Press enter for another iteration of LRTDP...".format(iteration))
solved, values = visit(inst, inst.init, solved, values)
print("Values after iteration {}: ".format(iteration))
print_values(inst, values)
print("Solved after iteration {}: ".format(iteration))
print_solved(inst, solved)
iteration += 1
return values
def main():
grid = standard_grid(obey_prob=1.0, step_cost=None)
# print rewards
print("rewards:")
print_values(grid.rewards, grid)
V, policy, deltas = monte_carlo(grid)
print("final values:")
print_values(V, grid)
print("final policy:")
print_policy(policy, grid)
plt.plot(deltas)
plt.show()
def test_data_store_variable(variable, name, indent=''):
print('%s - Test %s:' % (indent, name))
print('%s - Name: %s' % (indent, variable.name()))
print('%s - Unit: %s' % (indent, variable.unit()))
print('%s - URI: %s' % (indent, variable.uri()))
values_count = variable.values_count()
test_data_store_variable_index(variable, -1, indent)
test_data_store_variable_index(variable, 0, indent)
test_data_store_variable_index(variable, values_count - 1, indent)
test_data_store_variable_index(variable, values_count, indent)
for run in range(variable.runs_count() + 3):
print('%s - values(%d): ' % (indent, run - 2), end='')
utils.print_values(variable.values(run - 2))
def rtdp(inst, values):
iteration = 1
while True:
wait_for_input("Press enter for another iteration of RTDP...")
old_values = dict(values)
values = perform_trial(inst, values)
print("Values after iteration {}: ".format(iteration))
print_values(inst, values)
change = compute_max_difference(old_values, values)
if change < EPSILON:
print("Converged in iteration {}".format(iteration))
break
iteration += 1
return values
def forward(self, x):
if self.filter_size > 0:
return self.layers(x) #image, conv, batchnorm, relu
else:
y = torch.add(x.unsqueeze(2), self.noise * self.level)
# (10, 3, 1, 32, 32) + (1, 3, 128, 32, 32) --> (10, 3, 128, 32, 32)
if self.debug:
print_values(x, self.noise, y, self.unique_masks)
y = y.view(-1, self.in_channels * self.nmasks, self.input_size,
self.input_size)
y = self.layers(y)
if self.mix_maps:
y = self.mix_layers(y)
return y #image, perturb, (relu?), conv1x1, batchnorm, relu + mix_maps (conv1x1, batchnorm relu)
from neutronclient.v2_0 import client from credentials import get_credentials from utils import print_values credentials = get_credentials() neutron = client.Client(**credentials) ports = neutron.list_ports() print_values(ports, 'ports')
}
return keystone_client.Client(**params)
tests = [
{"user": "******", "password": "******", "tenant": "admin"},
]
for test in tests:
print "Attempting authentication for tenant %s by user %s" % (
test['tenant'], test['user']
),
try:
ks = get_keystone_client(**test)
print "Authorized"
except:
print "Denied"
#
#
#
#Listing All the networks from OpenStack environment.
#
#
print "Listing All Networks."
credentials = get_credentials()
neutron = client.Client(**credentials)
list_network = neutron.list_networks()
print_values(list_network, 'networks')
from neutronclient.v2_0 import client
from credentials import get_credentials_tenant_one
from utils import print_values
credentials = get_credentials_tenant_one("user1", "user1", "user1-project")
neutron = client.Client(**credentials)
netw = neutron.list_networks()
print_values(netw, 'networks')
tests = [
{
"user": "******",
"password": "******",
"tenant": "admin"
},
]
for test in tests:
print "Attempting authentication for tenant %s by user %s" % (
test['tenant'], test['user']),
try:
ks = get_keystone_client(**test)
print "Authorized"
except:
print "Denied"
#
#
#
#Listing All the networks from OpenStack environment.
#
#
print "Listing All Networks."
credentials = get_credentials()
neutron = client.Client(**credentials)
list_network = neutron.list_networks()
print_values(list_network, 'networks')
from neutronclient.v2_0 import client
from credentials import get_credentials
from utils import print_values
try:
credentials = get_credentials()
neutron = client.Client(**credentials)
routers_list = neutron.list_routers(retrieve_all=True)
print_values(routers_list, 'routers')
finally:
print 'Execution Completed'
from neutronclient.v2_0 import client from credentials import get_credentials from utils import print_values credentials = get_credentials() neutron = client.Client(**credentials) ports = neutron.list_ports() print_values(ports, 'ports')
def list_the_ports(): ports = neutron.list_ports() print print_values(ports, 'ports') return
Q[s][a] = np.mean(returns[sa])
biggest_change = max(biggest_change, np.abs(old_q - Q[s][a]))
seen_state_action_pairs.add(sa)
deltas.append(biggest_change)
for s in policy.keys():
a, _ = max_dict(Q[s])
policy[s] = a
V = {}
for s in policy.keys():
V[s] = max_dict(Q[s])[1]
return V, policy, deltas
if __name__ == '__main__':
grid = standard_grid(obey_prob=1.0, step_cost=None)
print("rewards:")
print_values(grid.rewards, grid)
V, policy, deltas = monte_carlo(grid)
print("final values:")
print_values(V, grid)
print("final policy:")
print_policy(policy, grid)
plt.plot(deltas)
plt.show()
for i in range(N):
visited_states = set()
states_and_returns = play_episode(standard_grid(), pi)
for s, g in states_and_returns:
if s not in visited_states:
visited_states.add(s)
if s not in all_returns:
all_returns[s] = []
all_returns[s].append(g)
V[s] = np.mean(all_returns[s])
return V
ALL_POSSIBLE_ACTIONS = ['U', 'D', 'L', 'R']
if __name__ == '__main__':
pi = {
(0, 0): 'R',
(0, 1): 'R',
(0, 2): 'R',
(1, 2): 'U',
(2, 2): 'U',
(2, 1): 'L',
(2, 0): 'U',
(1, 0): 'U',
(2, 3): 'L',
}
grid = standard_grid()
print_policy(pi, grid)
V = first_visit_monte_carlo_prediction(pi, 100)
print_values(V, grid)
#!/usr/bin/env python
from neutronclient.v2_0 import client
from credentials import get_credentials
from utils import print_values
#List the routers created by the create-router code
try:
credentials = get_credentials()
neutron = client.Client(**credentials)
routers_list = neutron.list_routers(retrieve_all=True)
print_values(routers_list, 'routers')
finally:
print("Execution completed")
#List the subnets which queries the neutron.list method
credentials = get_credentials()
neutron = client.Client(**credentials)
subnets = neutron.list_subnets()
print(subnets)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('algorithm',
choices=['rtdp', 'lrtdp'],
help="Choose the algorithm.")
args = parser.parse_args()
inst = instance.get_example_instance()
print(inst)
values = {s: heuristic(inst, s) for s in inst.states}
print("")
print("Initial state-values:")
print_values(inst, values)
if args.algorithm == 'rtdp':
values = rtdp(inst, values)
elif args.algorithm == 'lrtdp':
values = lrtdp(inst, values)
else:
sys.exit("Unknown algorithm")
print("")
print("Final values:")
print_values(inst, values)
policy = get_greedy_policy(inst, values)
print("Corresponding final policy:")
print_policy(inst, policy)
