Struct NStepBuffer 

pub struct NStepBuffer { /* private fields */ }

Circular accumulator for n-step returns.

Implementations

impl NStepBuffer

pub fn new(n: usize, gamma: f32) -> Self

Create an n-step buffer.

• n — number of steps to accumulate.
• gamma — discount factor γ ∈ (0, 1].

Panics

Panics if n == 0.

pub fn n(&self) -> usize

Number of steps n.

pub fn gamma(&self) -> f32

Discount factor γ.

pub fn count(&self) -> usize

Current number of accumulated steps.

pub fn push( &mut self, obs: impl Into<Vec<f32>>, action: impl Into<Vec<f32>>, reward: f32, next_obs: impl Into<Vec<f32>>, done: bool, ) -> Option<NStepTransition>

Push a (obs, action, reward, next_obs, done) step.

Returns Some(NStepTransition) once n steps have been accumulated (or immediately on terminal). Returns None while the buffer is still filling up.

pub fn flush(&mut self) -> Vec<NStepTransition>

Flush any remaining steps in the buffer (for end-of-episode cleanup).

Returns all remaining n-step transitions.

pub fn reset(&mut self)

Clear the buffer (e.g. at episode reset).

pub fn try_get(&self) -> RlResult<NStepTransition>

Attempt to produce an n-step transition from the current state.

Errors

• RlError::NStepIncomplete if fewer than n steps have been accumulated.

Trait Implementations

impl Clone for NStepBuffer

fn clone(&self) -> NStepBuffer

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for NStepBuffer

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.