def run(use_gui=True, runs=1):
out_csv = 'outputs/double/sarsa-double'
env = SumoEnvironment(net_file='nets/double/network.net.xml',
single_agent=False,
route_file='nets/double/flow.rou.xml',
out_csv_name=out_csv,
use_gui=use_gui,
num_seconds=86400,
yellow_time=3,
min_green=5,
max_green=60)
fixed_tl = False
agents = {
ts_id: TrueOnlineSarsaLambda(env.observation_spaces(ts_id),
env.action_spaces(ts_id),
alpha=0.000000001,
gamma=0.95,
epsilon=0.05,
lamb=0.1,
fourier_order=7)
for ts_id in env.ts_ids
}
for run in range(1, runs + 1):
obs = env.reset()
done = {'__all__': False}
if fixed_tl:
while not done['__all__']:
_, _, done, _ = env.step(None)
else:
while not done['__all__']:
actions = {
ts_id: agents[ts_id].act(obs[ts_id])
for ts_id in obs.keys()
}
next_obs, r, done, _ = env.step(action=actions)
for ts_id in next_obs.keys():
agents[ts_id].learn(state=obs[ts_id],
action=actions[ts_id],
reward=r[ts_id],
next_state=next_obs[ts_id],
done=done[ts_id])
obs[ts_id] = next_obs[ts_id]
env.save_csv(out_csv, run)
def test_api():
env = SumoEnvironment(single_agent=True,
num_seconds=100000,
net_file='nets/single-intersection/single-intersection.net.xml',
route_file='nets/single-intersection/single-intersection.rou.xml')
env.reset()
check_env(env)
env.close()
prs.add_argument("-mingreen", dest="min_green", type=int, default=5, required=False, help="Minimum green time.\n")
prs.add_argument("-maxgreen", dest="max_green", type=int, default=50, required=False, help="Maximum green time.\n")
prs.add_argument("-gui", action="store_true", default=False, help="Run with visualization on SUMO.\n")
prs.add_argument("-fixed", action="store_true", default=False, help="Run with fixed timing traffic signals.\n")
prs.add_argument("-s", dest="seconds", type=int, default=400000, required=False, help="Number of simulation seconds.\n")
prs.add_argument("-runs", dest="runs", type=int, default=1, help="Number of runs.\n")
args = prs.parse_args()
out_csv = 'outputs/2way-single-intersection/sarsa_lambdavai'
write_route_file('nets/2way-single-intersection/single-intersection-gen.rou.xml', 400000, 100000)
env = SumoEnvironment(net_file='nets/2way-single-intersection/single-intersection.net.xml',
single_agent=True,
route_file=args.route,
out_csv_name=out_csv,
use_gui=args.gui,
num_seconds=args.seconds,
min_green=args.min_green,
max_green=args.max_green,
max_depart_delay=0)
for run in range(1, args.runs+1):
obs = env.reset()
agent = TrueOnlineSarsaLambda(env.observation_space, env.action_space, alpha=args.alpha, gamma=args.gamma, epsilon=args.epsilon, fourier_order=7, lamb=0.9)
done = False
if args.fixed:
while not done:
_, _, done, _ = env.step({})
else:
while not done:
prs.add_argument("-runs",
dest="runs",
type=int,
default=1,
help="Number of runs.\n")
args = prs.parse_args()
experiment_time = str(datetime.now()).split('.')[0]
out_csv = 'outputs/single-intersection/{}_alpha{}_gamma{}_eps{}_decay{}_reward{}'.format(
experiment_time, args.alpha, args.gamma, args.epsilon, args.decay,
args.reward)
env = SumoEnvironment(
net_file='nets/single-intersection/single-intersection.net.xml',
route_file=args.route,
out_csv_name=out_csv,
use_gui=args.gui,
num_seconds=args.seconds,
min_green=args.min_green,
max_green=args.max_green,
max_depart_delay=0)
for run in range(1, args.runs + 1):
initial_states = env.reset()
ql_agents = {
ts: QLAgent(starting_state=env.encode(initial_states[ts], ts),
state_space=env.observation_space,
action_space=env.action_space,
alpha=args.alpha,
gamma=args.gamma,
exploration_strategy=EpsilonGreedy(
initial_epsilon=args.epsilon,
import os
import sys
if 'SUMO_HOME' in os.environ:
tools = os.path.join(os.environ['SUMO_HOME'], 'tools')
sys.path.append(tools)
else:
sys.exit("Please declare the environment variable 'SUMO_HOME'")
from sumo_rl import SumoEnvironment
from sumo_rl.util.gen_route import write_route_file
import traci
from stable_baselines.common.policies import MlpPolicy
from stable_baselines.common.vec_env import SubprocVecEnv
from stable_baselines import A2C
if __name__ == '__main__':
write_route_file('nets/2way-single-intersection/single-intersection-gen.rou.xml', 400000, 100000)
env = SubprocVecEnv([lambda: SumoEnvironment(net_file='nets/2way-single-intersection/single-intersection.net.xml',
route_file='nets/2way-single-intersection/single-intersection-gen.rou.xml',
out_csv_name='outputs/2way-single-intersection/a2c',
single_agent=True,
use_gui=False,
num_seconds=100000,
min_green=5)])
model = A2C(MlpPolicy, env, verbose=1, learning_rate=0.001, lr_schedule='constant')
model.learn(total_timesteps=100000)
import sys
if 'SUMO_HOME' in os.environ:
tools = os.path.join(os.environ['SUMO_HOME'], 'tools')
sys.path.append(tools)
else:
sys.exit("Please declare the environment variable 'SUMO_HOME'")
from sumo_rl import SumoEnvironment
import traci
if __name__ == '__main__':
env = SumoEnvironment(
net_file='nets/2way-single-intersection/single-intersection.net.xml',
route_file=
'nets/2way-single-intersection/single-intersection-vhvh.rou.xml',
out_csv_name='outputs/2way-single-intersection/dqn',
single_agent=True,
use_gui=False,
num_seconds=100000,
max_depart_delay=0)
model = DQN(env=env,
policy="MlpPolicy",
learning_rate=0.01,
learning_starts=0,
train_freq=1,
target_update_interval=100,
exploration_initial_eps=0.05,
exploration_final_eps=0.01,
verbose=1)
model.learn(total_timesteps=100000)
prs.add_argument("-s",
dest="seconds",
type=int,
default=80000,
required=False,
help="Number of simulation seconds.\n")
args = prs.parse_args()
experiment_time = str(datetime.now()).split('.')[0].replace(' ', '_')
scenario = args.network.replace('nets/5x5-Raphael/',
'').replace('.net.xml', '')
out_csv = f'outputs/5x5-Raphael/{scenario}_{experiment_time}_alpha{args.alpha}_gamma{args.gamma}_eps{args.epsilon}_decay{args.decay}'
env = SumoEnvironment(net_file=args.network,
route_file=args.route,
out_csv_name=out_csv,
use_gui=args.gui,
num_seconds=args.seconds,
min_green=args.min_green,
max_green=args.max_green,
max_depart_delay=0)
initial_states = env.reset()
ql_agents = {
ts: QLAgent(starting_state=env.encode(initial_states[ts], ts),
state_space=env.observation_spaces(ts),
action_space=env.action_spaces(ts),
alpha=args.alpha,
gamma=args.gamma,
exploration_strategy=EpsilonGreedy(
initial_epsilon=args.epsilon,
min_epsilon=args.min_epsilon,
decay=args.decay))
import sys
if 'SUMO_HOME' in os.environ:
tools = os.path.join(os.environ['SUMO_HOME'], 'tools')
sys.path.append(tools)
else:
sys.exit("Please declare the environment variable 'SUMO_HOME'")
import numpy as np
from sumo_rl import SumoEnvironment
import traci
from stable_baselines import DQN
env = SumoEnvironment(net_file='nets/DiamondNet/DiamondTLs.net.xml',
single_agent=True,
route_file='nets/DiamondFlowsAcrossNet.rou.xml',
out_csv_name='outputs/DiamondNet/dqn',
use_gui=True,
num_seconds=5400,
yellow_time=4,
min_green=5,
max_green=60,
max_depart_delay=0)
model = DQN(env=env,
policy="MlpPolicy",
learning_rate=1e-3,
buffer_size=50000,
exploration_fraction=0.05,
exploration_final_eps=0.02)
model.learn(total_timesteps=100000)
if 'SUMO_HOME' in os.environ:
tools = os.path.join(os.environ['SUMO_HOME'], 'tools')
sys.path.append(tools)
else:
sys.exit("Please declare the environment variable 'SUMO_HOME'")
import numpy as np
from sumo_rl import SumoEnvironment
import traci
from stable_baselines import DQN
env = SumoEnvironment(net_file='nets/big-intersection/big-intersection.net.xml',
single_agent=True,
route_file='nets/big-intersection/routes.rou.xml',
out_csv_name='outputs/big-intersection/dqn',
use_gui=False,
num_seconds=5400,
yellow_time=4,
min_green=5,
max_green=60)
model = DQN(
env=env,
policy="MlpPolicy",
learning_rate=1e-3,
buffer_size=50000,
exploration_fraction=0.05,
exploration_final_eps=0.02
)
model.learn(total_timesteps=100000)
sys.exit("Please declare the environment variable 'SUMO_HOME'")
from sumo_rl import SumoEnvironment
from sumo_rl.agents import QLAgent
from sumo_rl.exploration import EpsilonGreedy
if __name__ == '__main__':
alpha = 0.1
gamma = 0.995
decay = 1
runs = 4
env = SumoEnvironment(net_file='nets/4x4-Lucas/4x4.net.xml',
route_file='nets/4x4-Lucas/4x4c1c2c1c2.rou.xml',
use_gui=False,
num_seconds=80000,
min_green=5,
delta_time=5)
initial_states = env.reset()
ql_agents = {
ts: QLAgent(starting_state=env.encode(initial_states[ts], ts),
state_space=env.observation_space,
action_space=env.action_space,
alpha=alpha,
gamma=gamma,
exploration_strategy=EpsilonGreedy(initial_epsilon=0.05,
min_epsilon=0.005,
decay=decay))
for ts in env.ts_ids
}
import traci
from sumo_rl import SumoEnvironment
from sumo_rl.agents import QLAgent
from sumo_rl.exploration import EpsilonGreedy
if __name__ == '__main__':
alpha = 0.1
gamma = 0.99
decay = 1
runs = 1
env = SumoEnvironment(net_file='nets/4x4-Lucas/4x4.net.xml',
route_file='nets/4x4-Lucas/4x4c1c2c1c2.rou.xml',
use_gui=True,
num_seconds=80000,
max_depart_delay=0)
for run in range(1, runs+1):
initial_states = env.reset()
ql_agents = {ts: QLAgent(starting_state=env.encode(initial_states[ts], ts),
state_space=env.observation_space,
action_space=env.action_space,
alpha=alpha,
gamma=gamma,
exploration_strategy=EpsilonGreedy(initial_epsilon=0.05, min_epsilon=0.005, decay=decay)) for ts in env.ts_ids}
infos = []
done = {'__all__': False}
while not done['__all__']:
actions = {ts: ql_agents[ts].act() for ts in ql_agents.keys()}
from stable_baselines import A2C
if __name__ == '__main__':
write_route_file(
'nets/2way-single-intersection/single-intersection-gen.rou.xml',
400000, 100000)
# multiprocess environment
n_cpu = 1
env = SubprocVecEnv([
lambda: SumoEnvironment(
net_file=
'nets/2way-single-intersection/single-intersection.net.xml',
route_file=
'nets/2way-single-intersection/single-intersection-gen.rou.xml',
out_csv_name='outputs/2way-single-intersection/a2c',
single_agent=True,
use_gui=False,
num_seconds=100000,
min_green=5,
max_depart_delay=0) for _ in range(n_cpu)
])
model = A2C(MlpPolicy,
env,
verbose=1,
learning_rate=0.001,
lr_schedule='constant')
model.learn(total_timesteps=100000)
from ray.rllib.agents.a3c.a3c import A3CTrainer
from ray.rllib.agents.a3c.a3c_tf_policy import A3CTFPolicy
from ray.tune.registry import register_env
from gym import spaces
import numpy as np
from sumo_rl import SumoEnvironment
import traci
if __name__ == '__main__':
ray.init()
register_env(
"4x4grid", lambda _: SumoEnvironment(
net_file='nets/4x4-Lucas/4x4.net.xml',
route_file='nets/4x4-Lucas/4x4c1c2c1c2.rou.xml',
out_csv_name='outputs/4x4grid/a3c',
use_gui=False,
num_seconds=80000,
max_depart_delay=0))
trainer = A3CTrainer(
env="4x4grid",
config={
"multiagent": {
"policies": {
'0':
(A3CTFPolicy, spaces.Box(low=np.zeros(10),
high=np.ones(10)),
spaces.Discrete(2), {})
},
"policy_mapping_fn":
