DQNOptions<T>
Options & Configuration · AiDotNet.Models.Options
Configuration options for Deep Q-Network (DQN) agents.
Inheritance: ModelOptions → DQNOptions<T>
Constructors
| Constructor | Summary |
|---|---|
DQNOptions() |
Creates options with the default values. |
DQNOptions(DQNOptions<T> other) |
Initializes a new instance as a copy of other, with every setting copied. |
Properties
| Property | Summary |
|---|---|
Func<IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>>> OptimizerFactory { get; set; } |
Gets or sets a factory for the optimizer each of this agent’s networks trains with. |
ILossFunction<T> LossFunction { get; set; } |
Loss function used to fit Q-values to their TD targets. |
List<int> HiddenLayers { get; set; } |
Hidden layer sizes of the Q-network. |
T DiscountFactor { get; set; } |
Discount factor gamma applied to future rewards in the Q-learning target. |
T LearningRate { get; set; } |
Learning rate of the optimizer. |
double EpsilonDecay { get; set; } |
Rate at which the exploration rate decays from EpsilonStart toward EpsilonEnd. |
double EpsilonEnd { get; set; } |
Final, minimum exploration rate for epsilon-greedy action selection. |
double EpsilonStart { get; set; } |
Initial epsilon of the epsilon-greedy exploration schedule. |
int ActionSize { get; set; } |
Number of discrete actions, or the dimension of the action vector for a continuous action space. |
int BatchSize { get; set; } |
Number of samples in each training batch. |
int ReplayBufferSize { get; set; } |
Maximum number of transitions kept in the replay buffer. |
int StateSize { get; set; } |
Number of features in one observation of the environment state. |
int TargetUpdateFrequency { get; set; } |
Number of training steps between target-network updates. |
int WarmupSteps { get; set; } |
Number of environment steps collected before learning starts. |