Packages

package policy

Ordering
  1. Alphabetic
Visibility
  1. Public
  2. All

Type Members

  1. case class EpsilonGreedy[A, R, T](config: Config[R, T], actionValues: Map[A, T])(implicit evidence$1: Semigroup[T], evidence$2: Ordering[T]) extends Policy[A, R, EpsilonGreedy[A, R, T]] with Product with Serializable

    This is a version that accumulates the reward using a monoid.

  2. case class Gradient[A, R, T](config: Config[R, T], actionValues: Map[A, Item[T]])(implicit evidence$1: Equiv[A], evidence$2: ToReal[R], evidence$3: ToReal[T]) extends Policy[A, R, Gradient[A, R, T]] with Product with Serializable

    This thing needs to track its average reward internally...

    This thing needs to track its average reward internally... then, if we have the gradient baseline set, use that thing to generate the notes.

    T is the "average" type.

  3. case class Random[A, R]() extends Policy[A, R, Random[A, R]] with Product with Serializable

    Random policy.

  4. case class UCB[A, R, T](config: Config[R, T], actionValues: Map[A, Choice[T]], time: Time) extends Policy[A, R, UCB[A, R, T]] with Product with Serializable

Value Members

  1. object EpsilonGreedy extends Serializable
  2. object Gradient extends Serializable
  3. object Random extends Serializable
  4. object UCB extends Serializable

Ungrouped