EnvironmentFactory ()
Initialize self. See help(type(self)) for accurate signature.
ControlEnvironment (name=None, value=None, links=None, new_name=True, namespace=None)
Abstract ControlEnvironment
OpenAIGym (env_name=None, render=False, render_mode='rgb_array', video_wrap=False, value=0, name='gym', seed=None, links=None, new_name=True, early_termination=False, namespace=None, **cargs)
A function that creates and runs an environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function.
EnvironmentRuntimeFactory ()
create a factory class that creates objects of type BaseEnvironmentRuntime based on the environment name
CartPoleV1Runtime ()
A runtime environment for the CartPole-v1 environment. Inherits from BaseEnvironmentRuntime.
LunarLanderContinuousV2Runtime ()
A runtime environment for the LunarLanderContinuous-v2 environment. Inherits from BaseEnvironmentRuntime.
BaseEnvironmentRuntime ()
Helper class that provides a standard way to create an ABC using inheritance.
# Test for EnvironmentRuntimeFactory
env_runtime = EnvironmentRuntimeFactory.create('LunarLanderContinuous-v2')
#print(env_runtime)
print(env_runtime.get_env_inputs_names())
env_runtime = EnvironmentRuntimeFactory.create('CartPole-v1')
#print(env_runtime)
print(env_runtime.get_env_inputs_names())['IX', 'IY', 'IVX', 'IVY', 'IA', 'IVA', 'ILC', 'IRC']
['ICP', 'ICV', 'IPA', 'IPV']GymMetaData (env)
Initialize self. See help(type(self)) for accurate signature.
GenericGym (render=False, render_mode='rgb_array', video_wrap=False, value=0, gym_name=None, name='GenericGym', seed=None, links=None, new_name=True, namespace=None, **cargs)
A function that creates an runs the GenericGym environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function.
CartPoleV1 (render=False, render_mode='rgb_array', video_wrap=False, value=0, name='CartPoleV1', seed=None, links=None, new_name=True, namespace=None, **cargs)
A function that creates an runs the CartPole-v1 environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function.
CartPoleDV1 (render=False, render_mode='rgb_array', video_wrap=False, value=0, name='CartPoleD-v1', seed=None, links=None, new_name=True, namespace=None, **cargs)
A function that creates an runs the CartPole-v1 environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function.
Pendulum (render=False, render_mode='rgb_array', video_wrap=False, value=0, name='Pendulum', seed=None, links=None, new_name=True, namespace=None, **cargs)
A function that creates an runs the Pendulum-v1 environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function.
class Pendulum_1(OpenAIGym):
"A function that creates an runs the Pendulum-v1 environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function."
# from obs[0], indices
# 0 cos(theta) - +1 is up, -1 is down, 0 is left and right
# 1 sin(theta) - +1 is left, -1 is right, 0 is up and down
# 2 theta dot - +dot is anti-clockwise, -dot is clockwise
# 3 theta dot - normalised to +/- 1
# 4 theta +1/-1 (added here) 1 is pointing upwards, + is anti-clockwise, - is clockwise
# reward - -(theta^2 + 0.1*theta_dt^2 + 0.001*action^2)
def __init__(self, render=False, render_mode="rgb_array", video_wrap=False, value=0, name="Pendulum_1",
seed=None, links=None, new_name=True, namespace=None,**cargs):
super().__init__('Pendulum-v1', render=render, render_mode=render_mode, video_wrap=video_wrap, value=value, name=name, seed=seed,
links=links, new_name=new_name, namespace=namespace, **cargs)
def process_hierarchy_values(self):
self.hierarchy_values = self.links[0].get_value()
def process_actions(self):
pass
def apply_actions_get_obs(self):
return self.env.step([self.hierarchy_values])
def parse_obs(self):
self.value = self.obs[0]
self.reward = -self.obs[1]
self.done = self.obs[2]
self.info = self.obs[3]
def process_values(self):
vel = self.obs[0][2]/8.0
self.value = np.append(self.value, vel)
x = math.copysign(math.acos(self.obs[0][0]), self.obs[0][1])/math.pi
#theta = 100 - (10 * math.copysign(1-abs(x), x))
theta = 100 - (10 * x)
self.value = np.append(self.value, theta)
class Factory:
def create(self, namespace=None, seed=None, gym_name=None): return Pendulum_1(namespace=namespace, seed=seed, gym_name=gym_name)MountainCarV0 (render=False, render_mode='rgb_array', video_wrap=False, value=0, name='MountainCarV0', seed=None, links=None, new_name=True, namespace=None, **cargs)
A function that creates and runs the MountainCar-v0 environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function.
MountainCarContinuousV0 (render=False, render_mode='rgb_array', video_wrap=False, value=0, name='MountainCarContinuousV0', seed=None, links=None, new_name=True, early_termination=True, namespace=None, **cargs)
A function that creates and runs the MountainCarContinuous-v0 environment from OpenAI Gym. Parameter: The environment name. Flag to display environment. Links: Link to the action function.
WindTurbine (value=0, name='WindTurbine', links=None, new_name=True, namespace=None, seed=None, gym_name=None, **cargs)
A function that creates and runs the YawEnv environment for a wind turbine. Indexes 0 - action, 1 - yaw error, 2 - wind direction, 3 - wind speed (ignore 0).
VelocityModel (mass=50, value=0, name='VelocityModel', links=None, num_links=1, new_name=True, indexes=0, namespace=None, **cargs)
A simple model of a moving object of a particular mass. Parameters: The environment name, mass. Links: Link to the action function.
DummyModel (name='World', value=0, links=None, new_name=True, namespace=None, seed=None, **cargs)
Base class of a PCT function. This class is not used directly by developers, but defines the functionality common to all.
WebotsWrestler (render=False, value=0, name='WebotsWrestler', seed=None, links=None, new_name=True, early_termination=True, namespace=None)
A function that creates and runs a Webots Wrestler robot.
WebotsWrestlerSupervisor (render=False, value=0, name='WebotsWrestlerSupervisor', seed=None, links=None, new_name=True, early_termination=True, namespace=None)
A function that creates and runs a Webots Wrestler robot.
Bridge (render=False, value=0, name='Bridge', seed=None, links=None, new_name=True, early_termination=True, namespace=None)
An environment function with sensors set by external system.
MicroGrid (value=0, name='MicroGrid', links=None, new_name=True, namespace=None, seed=None, **cargs)
A function that creates and runs the microgrid environment for an energy management system.
‘Deep reinforcement learning for energy management in a microgrid with flexible demand.’
Taha Abdelhalim Nakabi, Pekka Toivanen.
https://doi.org/10.1016/j.segan.2020.100413
Inputs - st = [SoCt, BSCt, Cbt, Tt, Gt, Put, Lb,t, t].
0 - ISC - the average SoC (state-of-charge) of the TCLs,
1 - IL - the current load value of the daily consumption pattern.
2 - IPC - the pricing counter,
3 - IBS - the battery SoC,
4 - IEG - the energy generation,
5 - IT - the temparature,
6 - IEP - the electricty prices,
7 - ITS - the time step,
Actions:
0 - TCL action, Atcl,
1 - price action, Ap,
2 - energy deficiency action, Ad,
3 - energy excess action, Ae
ARC (value:float=0, name:str='ARC', links:Optional[List]=None, new_name:bool=True, render:bool=False, seed:int=None, namespace:Optional[str]=None, gym_name=None, **cargs:dict)
A function that creates and runs an ARC environment from a file given the rask code.
env = ARC()
env.add_link(Constant(1))
# env.add_link(Constant(0))
properties = { 'dir': 'C:\\packages\\arc-prize-2024', 'file_prefix':'arc-agi_training_', 'code':'007bbfb7', 'dataset': 'train', 'control_set': ['dims'], 'input_set': ['env']}
# file_name = os.path.join(properties['dir'], properties['file_prefix']) + 'challenges.json'
# challenges_manager = ChallengesDataManager(file_name)
# data = challenges_manager.get_data_for_key(properties['code'])
# properties['data']=data
# env.set_properties(properties)
# env.set_render(True)
# env.reset()
# print(env.env.info)
# for i in range(6):
# state = env()
# env.summary()
# # print()
# print(env.get_config())
# print()
# print(env.output_string())
# print()
# print(state)
# env.close()An example showing how to use an OpenAI Gym function. And how to have another function which accesses one of the values of the gym environment.
render=False
print(render)
acrobot = OpenAIGym("Acrobot-v1", render=render, seed=1)
namespace=acrobot.namespace
acrobot.add_link(Constant(1, namespace=namespace))
#acrobot.get_config()FalseThe IndexedParameter type retrieves a value from a linked function based upon an index.
cos_angle1 = IndexedParameter(0, name="cos_angle1", namespace=namespace)
cos_angle1.add_link(acrobot)
print(cos_angle1.get_config())
#cos_angle1(){'type': 'IndexedParameter', 'name': 'cos_angle1', 'value': 0, 'links': {0: 'gym'}, 'index': 0}pen = Pendulum(render=True, namespace=namespace, seed=1)
pen.add_link(Constant([1], namespace=namespace))
print(pen.get_config())
#pen.run(steps=10, verbose=True)
# why TypeError: size must be two numbers?{'type': 'Pendulum', 'name': 'Pendulum', 'value': 0, 'links': {0: 'constant1'}, 'env_name': 'Pendulum-v1', 'reward': 0, 'done': False, 'info': {}}vm = VelocityModel(name='VelocityModel', namespace=namespace)
vm.add_link(Constant(1, name='mycon'))
print(vm.summary())
config = vm.get_config()
print(config)
assert config == {'type': 'VelocityModel', 'name': 'VelocityModel', 'value': 0, 'links': {0: 'mycon'}, 'mass': 50}
vm.run(steps=10, verbose=True)VelocityModel VelocityModel | 0 | links mycon
None
{'type': 'VelocityModel', 'name': 'VelocityModel', 'value': 0, 'links': {0: 'mycon'}, 'mass': 50}
0.020 0.040 0.060 0.080 0.100 0.120 0.140 0.160 0.180 0.200 0.19999999999999998