max_green=args.max_green,
max_depart_delay=0,
time_to_load_vehicles=120,
phases=[
traci.trafficlight.Phase(32, "GGrrrrGGrrrr"),
traci.trafficlight.Phase(2, "yyrrrryyrrrr"),
traci.trafficlight.Phase(32, "rrGrrrrrGrrr"),
traci.trafficlight.Phase(2, "rryrrrrryrrr"),
traci.trafficlight.Phase(32, "rrrGGrrrrGGr"),
traci.trafficlight.Phase(2, "rrryyrrrryyr"),
traci.trafficlight.Phase(32, "rrrrrGrrrrrG"),
traci.trafficlight.Phase(2, "rrrrryrrrrry")
])
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=21)
done = False
if args.fixed:
while not done:
_, _, done, _ = env.step({})
else:
while not done:
action = agent.act(agent.get_features(obs))
next_obs, r, done, _ = env.step(action=action)
agent.learn(state=obs, action=action, reward=r, next_state=next_obs, done=done)
obs = next_obs
traci.trafficlight.Phase(32, "GGrrrrGGrrrr"),
traci.trafficlight.Phase(2, "yyrrrryyrrrr"),
traci.trafficlight.Phase(32, "rrGrrrrrGrrr"),
traci.trafficlight.Phase(2, "rryrrrrryrrr"),
traci.trafficlight.Phase(32, "rrrGGrrrrGGr"),
traci.trafficlight.Phase(2, "rrryyrrrryyr"),
traci.trafficlight.Phase(32, "rrrrrGrrrrrG"),
traci.trafficlight.Phase(2, "rrrrryrrrrry")
])
if args.reward == 'queue':
env._compute_rewards = env._queue_average_reward
else:
env._compute_rewards = env._waiting_time_reward
for run in range(1, args.runs + 1):
initial_states = env.reset()
ql_agents = {
ts: QLAgent(starting_state=env.encode(initial_states[ts]),
state_space=env.observation_space,
action_space=env.action_space,
alpha=args.alpha,
gamma=args.gamma,
exploration_strategy=EpsilonGreedy(
initial_epsilon=args.epsilon,
min_epsilon=args.min_epsilon,
decay=args.decay))
for ts in env.ts_ids
}
done = {'__all__': False}
infos = []