ValueIterationAgent<T>
Models & Types ยท AiDotNet.ReinforcementLearning.Agents.DynamicProgramming
Value Iteration agent for reinforcement learning using dynamic programming.
Inheritance: ReinforcementLearningAgentBase<T> โ ValueIterationAgent<T>
For Beginners
Value Iteration repeatedly updates state values using the Bellman optimality equation until they converge. Unlike Policy Iteration which does full evaluation then improvement, Value Iteration does both in a single sweep - making it simpler but potentially slower per iteration. Think of it like iteratively improving a GPS estimate: each pass gets closer to the true shortest path. Requires a complete model of the environment.
How It Works
Value Iteration combines policy evaluation and improvement in a single update step, converging to the optimal value function.
Example
// Create a Value Iteration agent for dynamic programming
var options = new ValueIterationOptions<double> { StateSize = 4, ActionSize = 2 };
var agent = new ValueIterationAgent<double>(options);
// Select an action using the converged value function
var state = new Vector<double>(new double[] { 0.5, -0.3, 1.0, 0.2 });
var action = agent.SelectAction(state);
Constructors
| Constructor | Summary |
|---|---|
ValueIterationAgent() |
Initializes a new instance with default settings. |
ValueIterationAgent(ValueIterationOptions<T> options) |
Properties
| Property | Summary |
|---|---|
int FeatureCount { get; } |
Methods
| Method | Summary |
|---|---|
Dictionary<string, T> GetMetrics() |
|
ModelMetadata<T> GetModelMetadata() |
|
T Train() |
|
Task TrainAsync() |
|
Task<Vector<T>> PredictAsync(Vector<T> input) |
|
Vector<T> Predict(Vector<T> input) |
|
Vector<T> SelectAction(Vector<T> state, bool training) |
|
byte[] Serialize() |
Auto-generated common model serialization surface. |
void Deserialize(byte[] data) |
Auto-generated common model deserialization surface. |
void LoadModel(string filepath) |
|
void ResetEpisode() |
|
void SaveModel(string filepath) |
|
void StoreExperience(Vector<T> state, Vector<T> action, T reward, Vector<T> nextState, bool done) |