Module type Obandit.DecayingEpsilonGreedyParam

module type DecayingEpsilonGreedyParam = sig .. end

Use to instanciate a Bandit from MakeDecayingEpsilonGreedy .


val k : int

The number of actions $ K $ .

val c : float

The $ c$ hyperparameter.

val d : float

The $ d$ hyperparameter, a tight lower bound on $ \max_{i=1,\cdots,K} \Delta_i $.