Table of Contents

Class TabularQLearningOptions<T>

Namespace
AiDotNet.Models.Options
Assembly
AiDotNet.dll

Configuration options for Tabular Q-Learning agents.

public class TabularQLearningOptions<T> : ReinforcementLearningOptions<T>

Type Parameters

T

The numeric type used for calculations.

Inheritance
TabularQLearningOptions<T>
Inherited Members

Remarks

Tabular Q-Learning maintains a lookup table of Q-values for discrete state-action pairs. No neural networks or function approximation.

For Beginners: This is the simplest form of Q-Learning where we literally maintain a table. Each row is a state, each column is an action, and the cells contain Q-values.

Best for:

  • Small discrete state spaces (e.g., 10x10 grid world)
  • Discrete action spaces
  • Learning exact optimal policies
  • Understanding RL fundamentals

Not suitable for:

  • Continuous states (infinitely many states)
  • Large state spaces (millions of states)
  • High-dimensional observations (images, etc.)

Properties

ActionSize

Size of the action space (number of possible actions).

public int ActionSize { get; init; }

Property Value

int

StateSize

Size of the state space (number of state features).

public int StateSize { get; init; }

Property Value

int
Edit this page