dart_rl 0.1.0-alpha.1
dart_rl: ^0.1.0-alpha.1 copied to clipboard

Published 2 months ago •

vezz.io• Latest: 0.2.0-alpha.3

[unknown platforms]

2

→

Metadata

A Dart package implementing reinforcement learning algorithms (SARSA, Q-Learning, Expected-SARSA) for both Dart and Flutter applications.

Changelog #

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.1.0-alpha.1 2024-01-XX #

Added #

Initial alpha release of dart_rl package
Q-Learning algorithm implementation (QLearningAgent)
SARSA algorithm implementation (SarsaAgent)
Expected-SARSA algorithm implementation (ExpectedSarsaAgent)
Environment interface for creating custom RL environments
Agent base class with epsilon-greedy exploration strategy
State, Action, and StateAction classes for representing RL components
StepResult class for environment step results
Grid World example environment
Frozen Lake example environment
Comprehensive unit tests
Documentation and README with usage examples

Features #

Support for discrete state and action spaces
Configurable learning rate (α), discount factor (γ), and epsilon (ε)
Epsilon-greedy exploration with decay functionality
Q-table access for inspection and debugging
Training methods for single episodes and multiple episodes
Compatible with both Dart and Flutter applications

0.1.0 2024-01-XX #

Added #

Initial release of dart_rl package
Q-Learning algorithm implementation (QLearningAgent)
SARSA algorithm implementation (SarsaAgent)
Expected-SARSA algorithm implementation (ExpectedSarsaAgent)
Environment interface for creating custom RL environments
Agent base class with epsilon-greedy exploration strategy
State, Action, and StateAction classes for representing RL components
StepResult class for environment step results
Grid World example environment
Frozen Lake example environment
Comprehensive unit tests
Documentation and README with usage examples

Features #

Support for discrete state and action spaces
Configurable learning rate (?), discount factor (?), and epsilon (?)
Epsilon-greedy exploration with decay functionality
Q-table access for inspection and debugging
Training methods for single episodes and multiple episodes
Compatible with both Dart and Flutter applications

← Metadata

Publisher

Weekly Downloads

Metadata

A Dart package implementing reinforcement learning algorithms (SARSA, Q-Learning, Expected-SARSA) for both Dart and Flutter applications.

Repository (GitHub)
View/report issues

License

unknown (license)

Dependencies

collection, equatable

More

Packages that depend on dart_rl