Skip to main content

run_episode

Function run_episode 

Source
pub fn run_episode(
    agent: &mut dyn PlannerAgent,
    schedule: &PlannerSchedule,
    env_factory: &dyn EnvironmentFactory,
    random_seed: u64,
    compiled: &CompiledPlannerRunSpec,
) -> Result<PlannerRunReport, PlannerAgentError>
Expand description

Run one complete environment episode with a supplied agent and environment factory.

The agent’s learned model state is preserved across calls. Before the fresh environment is used, the agent receives an episode-boundary reset so controller-side randomness and episode-local transient state are aligned with random_seed.