DeterministicPolicyOptions<T>

Options & Configuration · AiDotNet.ReinforcementLearning.Policies

Configuration options for deterministic policies: DeterministicPolicy’s #ctor constructor builds its network from them.

Inheritance: ModelOptions → DeterministicPolicyOptions<T>

For Beginners

The policy always gives the same action for the same situation; during training a little noise is added so it tries nearby actions too.

How It Works

The actor of DDPG (Lillicrap et al., 2016) and TD3 (Fujimoto et al., 2018): the network maps a state straight to one action, and exploration comes from noise added to it during training.

Constructors

Constructor Summary
DeterministicPolicyOptions()

Properties

Property Summary
IExplorationStrategy<T> ExplorationStrategy { get; set; } Gets or sets the exploration noise added to actions during training.
ILossFunction<T> LossFunction { get; set; } Gets or sets the loss the policy network is trained with when it is fitted directly.
bool UseTanhSquashing { get; set; } Gets or sets whether actions pass through tanh, bounding them to [-1, 1].
int ActionSize { get; set; } Gets or sets the number of action dimensions (the network’s output width).
int StateSize { get; set; } Gets or sets the number of features in an observation (the network’s input width).
int[] HiddenLayers { get; set; } Gets or sets the width of each hidden layer of the policy network, in order.