flatland.core.policy module#
- class flatland.core.policy.Policy[source]#
Bases:
ABC,Generic[T_env,T_obs,T_act]Abstract base class for Flatland policies. Used for evaluation.
Loosely corresponding to ray-project/ray, but much simpler.
- act(observation: List[T_obs], **kwargs) T_act[source]#
Get action for agent. Called by act_many() for each agent.
Parameters#
- observation: Any
the agent’s observation
- kwargs
forward compatibility placeholder
Returns#
- Any
the action dict
- act_many(handles: List[int], observations: List[T_obs], **kwargs) Dict[int, T_act][source]#
Get action_dict for all agents. Default implementation calls act() for each handle in the list.
Override if you need to initialize before / cleanup after calling act() for individual agents.
Parameters#
- handles: List[int]
the agents’ handles
- observations: List[Any]
the agents’ observations
- kwargs
forward compatibility placeholder
Returns#
- Dict[int, Any]
the action dict