Skip to main content

hydrogym.nek.parallel_env

Dict-based multi-agent wrapper for NekEnv. Converts between array-based and dict-based interfaces.

NekParallelEnv Objects​

class NekParallelEnv()

Multi-agent wrapper for NekEnv with dict-based interface.

This wrapper treats each actuator as a separate agent with its own observation and action. Useful for multi-agent RL experiments or when you need per-actuator control.

Arguments:

  • nek_env - The base NekEnv instance to wrap

observation_space​

def observation_space(agent)

Get observation space for a specific agent.

action_space​

def action_space(agent)

Get action space for a specific agent.

reset​

def reset(seed=None, options=None) -> Tuple[Dict[str, np.ndarray], Dict]

Reset environment and return dict of observations.

step​

def step(
    actions: Dict[str, np.ndarray]
) -> Tuple[Dict[str, np.ndarray], Dict[str, float], Dict[str, bool], Dict[
        str, bool], Dict[str, dict]]

Step environment with dict of actions.

Arguments:

  • actions - Dict mapping agent names to actions

Returns:

  • observations - Dict of observations per agent
  • rewards - Dict of rewards per agent
  • terminated - Dict of terminated flags per agent
  • truncated - Dict of truncated flags per agent
  • infos - Dict of info dicts per agent

render​

def render(mode="human")

Render the environment.

close​

def close()

Close the environment.