Grid2Op is a pythonic, easy to use framework, to be able to develop, train or evaluate performances of "agent" or "controler" that acts on a power grid in different ways.
It is modular and can be use to train reinforcement learning agent or to assess the performance of optimal control algorithm.
It is flexible and allows the power flow to be computed by the algorithm of your choice. It abstracts the modification of a powergrid and use this abstraction to compute the cascading failures resulting from powerlines disconnection for example.
Features
- abstract the computation of the cascading failures
- ability to have the same code running with multiple powerflow calculator
- parallel execution of one agent on multiple independent scenarios (multiprocessing)
- fully customisable: this software has been built to be fully customizable to serve different purposes and not only reinforcement learning, or the L2RPN competition.
.. toctree:: :maxdepth: 2 intro quickstart grid2op makeenv
.. toctree:: :maxdepth: 2 plot
The gym framework in reinforcement learning is widely used. Starting from version 1.2.0 we improved the compatibility with this framework.
Before grid2op 1.2.0 only some classes fully implemented the open AI gym interface:
- the :class:`grid2op.Environment` (with methods such as env.reset, env.step etc.)
- the :class:`grid2op.Agent` (with the agent.act etc.)
- the creation of pre defined environments (with grid2op.make)
Starting from 1.2.0 we implemented some automatic converters that are able to automatically map grid2op representation for the action space and the observation space into open AI gym "spaces". More precisely these are represented as gym.spaces.Dict.
If you are interested by this feature, we recommend you to proceed like this:
import grid2op
from grid2op.Converters import GymActionSpace, GymObservationSpace
from grid2op.Agent import BaseAgent
class MyAgent(BaseAgent):
def __init__(self, action_space, observation_space):
BaseAgent.__init__(self, action_space)
self.gym_obs_space = GymObservationSpace(observation_space)
self.gym_action_space = GymActionSpace(observation_space)
def act(self, obs, reward, done=False):
# convert the observation to gym like one:
gym_obs = self.gym_obs_space.to_gym(obs)
# do whatever you want, as long as you retrieve a gym-like action
gym_action = ...
grid2op_action = self.gym_action_space.from_gym(gym_action)
# NB advanced usage: if action_space is a grid2op.converter (for example coming from IdToAct)
# then what's called "grid2op_action" is in fact an action that can be understood by the converter.
# to convert it back to grid2op action you need to convert it. See the documentation of GymActionSpace
# for such purpose.
return grid2op_action
env = grid2op.make(...)
my_agent = MyAgent(env.action_space, env.observation_space)
# and now do anything you like
# for example
done = False
reward = env.reward_range[0]
obs = env.reset()
while not done:
action = my_agent.act(obs, reward, done)
obs, reward, done, info = env.step(action)We also implemented some "converter" that allow the conversion of some action space into more convenient gym.spaces (this is only available if gym is installed of course). Please check :class:`grid2op.Converter.GymActionSpace` for more information and examples.
.. toctree:: :maxdepth: 2 action agent backend chronics converter environment rules observation opponent parameters reward runner space voltagecontroler
.. toctree:: :maxdepth: 2 exception