DynaQPlusOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for the Dyna-Q+ agent, which plans with a learned model and rewards revisiting long-untried actions.

Inheritance: ModelOptions → ReinforcementLearningOptions<T> → DynaQPlusOptions<T>

Constructors

Constructor Summary
DynaQPlusOptions()

Properties

Property Summary
double Kappa { get; set; } Weight of the exploration bonus kappa * sqrt(tau) given to state-action pairs not tried for tau steps (Sutton and Barto, Dyna-Q+).
int ActionSize { get; set; } Number of discrete actions, or the dimension of the action vector for a continuous action space.
int PlanningSteps { get; set; } Number of simulated planning updates from the learned model per real step.
int StateSize { get; set; } Number of features in one observation of the environment state.