microgrid generated with m_env.load(...) yields not usable env for RLLIB

[Pierrick-Pochelu]

This bug is not blocking for me because we can use" generate_microgrid" instead and everything works well. I do not see the difference between both way to generate microgrid.

from pymgrid import MicrogridGenerator as mg
from pymgrid.Environments.pymgrid_cspla import MicroGridEnv
from ray.rllib.agents import dqn
m_env = mg.MicrogridGenerator(nb_microgrid=25)
pymgrid25 = m_env.load('pymgrid25')
microgrids = pymgrid25.microgrids
mgi=microgrids[1]
mgi.set_horizon(24*30)
default_config={}
default_config["env_config"] = {"microgrid": mgi}
trainer=dqn.DQNTrainer(env=MicroGridEnv,config=default_config) # AttributeError: 'NoneType' object has no attribute 'copy'

[GonzagueHenri]

I don't have an error when I run this code

[Pierrick-Pochelu]

My version of Pymgrid is up to date. "mgi.parameters" lead to a DataFrame error, DQN cannot handle it (for example copy it).

We can use m.generate_microgrid() instead and does not yield error. Does "generate_microgrid" and "load('pymgrid25')" building ways are equivalent ?

[GonzagueHenri]

I am not sure about the error, I cannot reproduce it. But if you don't change the random seed the 25 first microgrids generated by pymgrid should be the same as in pymgrid25.

[GonzagueHenri]

Closing this issue for lack of activity