Trait vrp_core::algorithms::mdp::LearningStrategy[−][src]

pub trait LearningStrategy<S: State> {
    fn value(
        &self, 
        reward_value: f64, 
        old_value: f64, 
        estimates: &ActionEstimates<S>
    ) -> f64;
}

Expand description

A learning strategy for the MDP.

Required methods

fn value(
    &self,
    reward_value: f64,
    old_value: f64,
    estimates: &ActionEstimates<S>
) -> f64

Estimates an action value given received reward, current value, and actions values from the new state.

Implementors

impl<S: State> LearningStrategy<S> for MonteCarlo

impl<S: State> LearningStrategy<S> for QLearning