enjeeneer/pearl

Code for "Low Emissions Building Control with Zero-Shot Reinforcement Learning" by Jeen, Abate and Cullen (2022).

pearl's Issues

CO2

Hello, I think this is a very useful program and will help me a lot, but I would like to ask you where the CO2 data here comes from? Is it calculated? You are using energym environment, so I am not sure how your co2 value is calculated, so I would like to ask you.

Request for assistance

May I ask under what configuration this code should run and for how long? Because this code has been running for several days and has been stuck at the following interface. Although Wandb shows it is running, there are no metric records.

Issue with Running PEARL and Inquiring

Hi,

I've been attempting to run PEARL, but I've encountered some bugs, such as
File "main.py", line 51, in <module> action_dict, state_action, last_obs = agent.plan(obs, env,prev_action) TypeError: plan() takes 3 positional arguments but 4 were given
or
File "main.py", line 72, in <module> if (agent.n_steps <= agent.cfg.exploration_steps) and (agent.n_steps > cfg.window): ...... File "C:\Program Files (x86)\anaconda3\envs\pearl\lib\site-packages\omegaconf\dictconfig.py", line 480, in _get_node raise ConfigKeyError(f"Missing key {key!s}") omegaconf.errors.ConfigAttributeError: Missing key exploration_steps full_key: exploration_steps object_type=dict

I'm wondering whether these issues are a result of my own actions and is there currently a smoother-running version available?

By the way, I can not find a usable version of pymodelica. Do you know how to install it?

Thank you!