Trait renforce::model::Model

pub trait Model<S: Space, A: Space> {
    fn transition(&self,
                  curr: &S::Element,
                  action: &A::Element,
                  next: &S::Element)
                  -> f64;
    fn reward(&self,
              curr: &S::Element,
              action: &A::Element,
              next: &S::Element)
              -> f64;
    fn update(&mut self, transition: Transition<S, A>);
}

Represents a (nondeterministic) model of an environment The model itself is composed of the transition and reward functions

Required Methods

fn transition(&self,
              curr: &S::Element,
              action: &A::Element,
              next: &S::Element)
              -> f64

Returns the probabilty of moving from curr to next when performing action

fn reward(&self,
          curr: &S::Element,
          action: &A::Element,
          next: &S::Element)
          -> f64

Returns the reward received when moving from curr to next when performing action

fn update(&mut self, transition: Transition<S, A>)

Updates the model using information from the given transition

Implementors

• impl<S: FiniteSpace, A: FiniteSpace> Model<S, A> for PlainModel<S, A> where S::Element: Hash + Eq,
        A::Element: Hash + Eq