def __init__(
self,
action_space: gym.Space,
observation_space: gym.Space,
max_concurrent: int = 100,
):
"""Initializes an ExternalMultiAgentEnv instance.
Args:
action_space: Action space of the env.
observation_space: Observation space of the env.
max_concurrent: Max number of active episodes to allow at
once. Exceeding this limit raises an error.
"""
ExternalEnv.__init__(self, action_space, observation_space, max_concurrent)
# We require to know all agents' spaces.
if isinstance(self.action_space, dict) or isinstance(
self.observation_space, dict
):
if not (self.action_space.keys() == self.observation_space.keys()):
raise ValueError(
"Agent ids disagree for action space and obs "
"space dict: {} {}".format(
self.action_space.keys(), self.observation_space.keys()
)
)
def __init__(self, number_of_plants, server_port=9920, max_bid=600):
self.SERVER_ADDRESS = "localhost"
# SERVER_PORT = 9920
self.CHECKPOINT_FILE = "last_checkpoint.out"
self.server_port = server_port
self.max_bid = max_bid
self.number_of_plants = number_of_plants
lower_bounds = [-100000] * 7
# lower_bounds.extend([-99999])
upper_bounds = [10000000] * 7
# upper_bounds.extend([99999])
ExternalEnv.__init__(
self,
# MultiDiscrete([16, 10]),
# Discrete(159),
# action_space=Box(shape=37),
# action_space=Box(low=0, high=200, shape=(37,), dtype=np.float),
action_space=Box(low=0,
high=self.max_bid,
shape=(self.number_of_plants, ),
dtype=np.float),
observation_space=Box(np.array(lower_bounds),
np.array(upper_bounds)))
def __init__(self, config):
ExternalEnv.__init__(
self, spaces.Discrete(config["action_size"]),
spaces.Box(
low=-10, high=10,
shape=(config["observation_size"],),
dtype=np.float32))
self.server = config["input"]
def __init__(self, config):
self.config = config
ExternalEnv.__init__(
self,
action_space=spaces.Discrete(config['action_size']),
observation_space=spaces.Box(low=-10,
high=10,
shape=(config['observation_size'], ),
dtype=np.float32))
def __init__(self,
external_env: ExternalEnv,
preprocessor: "Preprocessor" = None):
self.external_env = external_env
self.prep = preprocessor
self.multiagent = issubclass(type(external_env), ExternalMultiAgentEnv)
self.action_space = external_env.action_space
if preprocessor:
self.observation_space = preprocessor.observation_space
else:
self.observation_space = external_env.observation_space
external_env.start()
def __init__(self, action_space, observation_space, max_concurrent=100):
"""Initialize a multi-agent external env.
ExternalMultiAgentEnv subclasses must call this during their __init__.
Arguments:
action_space (gym.Space): Action space of the env.
observation_space (gym.Space): Observation space of the env.
max_concurrent (int): Max number of active episodes to allow at
once. Exceeding this limit raises an error.
"""
ExternalEnv.__init__(self, action_space, observation_space,
max_concurrent)
# we require to know all agents' spaces
if isinstance(self.action_space, dict) or isinstance(
self.observation_space, dict):
if not (self.action_space.keys() == self.observation_space.keys()):
raise ValueError("Agent ids disagree for action space and obs "
"space dict: {} {}".format(
self.action_space.keys(),
self.observation_space.keys()))
def __init__(self, env, off_pol_frac):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.env = env
self.off_pol_frac = off_pol_frac
def __init__(self, action_space, observation_space):
ExternalEnv.__init__(self, action_space, observation_space)
def __init__(self, env, fixed_action):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.env = env
self.fixed_action = fixed_action
def __init__(self, env):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.env = env
def __init__(self, env, off_pol_frac):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.env = env
self.off_pol_frac = off_pol_frac
def __init__(self):
ExternalEnv.__init__(self, spaces.Discrete(64),
spaces.Box(low, high, dtype=np.float32))
def __init__(self, env_creator):
self.env_creator = env_creator
self.env = env_creator()
ExternalEnv.__init__(self, self.env.action_space,
self.env.observation_space)
def __init__(self, env, fixed_action):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.env = env
self.fixed_action = fixed_action
def __init__(self, env, file_name: str):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.csv_file = open(file_name, newline='')
def __init__(self, env, episodes: int):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.env = env
self.episodes = episodes
def __init__(self):
ExternalEnv.__init__(
self, spaces.Discrete(3), #discrete state == [0, 1, 2] or sell, hold, buy
spaces.Box(low=-100, high=100, shape=(10, ), dtype=np.float32)) # we'll take an NP array of 10 positions between -100, and 100 to train
def __init__(self, env_creator):
self.env_creator = env_creator
self.env = env_creator()
ExternalEnv.__init__(self, self.env.action_space,
self.env.observation_space)
def __init__(self):
ExternalEnv.__init__(
self, spaces.Discrete(2),
spaces.Box(low=-10, high=10, shape=(4, ), dtype=np.float32))
def __init__(self):
ExternalEnv.__init__(self, RLHelper.action_space(),
RLHelper.observation_space())
self.port = SERVER_PORT
def __init__(self, env):
ExternalEnv.__init__(self, env.action_space, env.observation_space)
self.env = env