Grid2Game is a python package that allows you to interact with a powergrid (using grid2op action types).
It allows to:
- "play" the game as if you were operating a powergrid
- inspect the state of powergrid visually
- inspect the past state of the grid (with the productions, loads as well as maximum flows on the powerlines)
- design some basic collaboration between an agent (called "assistant" in this package) and a human, by having the assistant proposing actions to the human
- save the current experiment in a standardized manner (using the grid2op runner) that can be read back with grid2op (for example if you want to perform some kind of supervized learning or an agent) or with grid2viz (for manual a posteriori examination)
quickstart play_the_game help_your_agent imitation_learning
This is a work in progress at the moment
load_assistant save_expe control_scenario navigate_in_scenario control_display timeline grid_displayed available_action temporal_data
genindexmodindexsearch